No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Towards automatic building extraction from high-resolution digital elevation models
Abstract
This paper deals with an approach for extracting the 3D shape of buildings from high-resolution Digital Elevation Models (DEMs), having a grid resolution between 0.5 and 5 m. The steps of the proposed procedure increasingly use explicit domain knowledge, specifically geometric constraints in the form of parametric and prismatic building models. A new MDL-based approach generating a polygonal ground plan from segment boundaries is given. The used knowledge is object-related making adaption to data of different density and resolution simple and transparent.
... The dense 3D point clouds are employed for building detection and extraction either directly or after being processed. Digital surface models (DSM), for example, which are derived by applying interpolations to irregular dense 3D point clouds, have proved to be more efficient than the direct use of dense 3D point clouds [10]. Symmetric neighborhoods around 3D structured data points can be defined using the DSM of LiDAR data and sharp linear features (e.g. ...
... Symmetric neighborhoods around 3D structured data points can be defined using the DSM of LiDAR data and sharp linear features (e.g. building boundaries) can be extracted easily by image-processing algorithms, such as morphological operators and gradient-based edge detectors [10,11,12]. ...
... buildings and trees). Weidner and Förstner [10] proposed a method based on morphological opening to approximate the DTM. This method subtracts the resulting DTM from a DSM to obtain a normalized DSM (nDSM) followed by a sequential height thresholding. ...
Automatic three-dimensional (3D) building model reconstruction using remote sensing data is crucial in applications which require large-scale and frequent building model updates, such as disaster monitoring and urban management, to avoid huge manual efforts and costs. Recent advances in the availability of very high-resolution satellite data together with efficient data acquisition and large area coverage have led to an upward trend in their applications for 3D building model reconstructions. In this dissertation, a novel multistage hybrid automatic 3D building model reconstruction approach is proposed which reconstructs building models in level of details 2 (LOD2) based on digital surface model (DSM) data generated from the very high-resolution stereo imagery of the WorldView-2 satellite. This approach uses DSM data in combination with orthorectified panchromatic (PAN) and pan-sharpened data of multispectral satellite imagery to overcome the drawbacks of DSM data, such as blurred building boundaries. In the first stage, the rough building boundaries in the DSM-based building masks are refined by classifying the geometrical features of the corresponding PAN images. The refined boundaries are then simplified in the second stage through a parametrization procedure which represents the boundaries by a set of line segments. The main orientations of buildings are then determined, and the line segments are regularized accordingly. The regularized line segments are then connected to each other based on a rule-based method to form polygonal building boundaries. In the third stage, a novel technique is proposed to decompose the building polygons into a number of rectangles under the assumption that buildings are usually composed of rectangular structures. In the fourth stage, a roof model library is defined, which includes flat, gable, half-hip, hip, pyramid and mansard roofs. These primitive roof types are then assigned to the rectangles based on a deep learning-based classification method. In the fifth stage, a novel approach is developed to reconstruct watertight parameterized 3D building models based on the results of the previous stages and normalized DSM (nDSM) of satellite imagery. In the final stage, a novel approach is proposed to optimize building parameters based on an exhaustive search, so that the two-dimensional (2D) distance between the 3D building models and the building boundaries (obtained from building masks and PAN image) as well as the 3D normal distance between the 3D building models and the 3D point clouds (obtained from nDSM) are minimized. In the final stage, a new approach is proposed to optimize building parameters based on an exhaustive search, so that the 2D distance between the two-dimensional (2D) and the building boundaries obtained from building masks and PAN images and normal distance between 3D building models and 3D point clouds derived from nDSM are minimized. Different parts of the building blocks are then merged through a newly proposed intersection and merging process. All corresponding experiments were conducted on four areas of the city of Munich including 208 buildings and the results were evaluated qualitatively and quantitatively. According to the results, the proposed approach can accurately reconstruct 3D models of buildings, even the complex ones with several inner yards and multiple orientations. Furthermore, the proposed approach provides a high level of automation by the limited number of primitive roof model types required and by performing automatic parameter initializations. In addition, the proposed boundary refinement method improves the DSM-based building masks specified by 8% in area accuracy. Furthermore, the ridge line directions and roof types were detected accurately for most of the buildings. The combination of the first three stages improved the accuracy of the building boundaries by 70% in comparison to using line segments extracted from building masks without refinement. Moreover, the proposed optimization approach can achieve in most cases the best combinations of 2D and 3D geometrical parameters of roof models. Finally, the intersection and merging process could successfully merge different parts of the complex building models.
... The suggested method is based on a subset of cartographic regularisation algorithms, which removes erroneous vertices while keeping important building shape information. The work of Weidner and Forstner (1995) inspired our method. The generated vectors from BSP are referred as nosy model boundaries. ...
... The generated vectors from BSP are referred as nosy model boundaries. Based on the Minimum Description Length also known as MDL, the proposed approach gradually corrects them [8][9][10][11][12]. ...
3D city models enable us to gain a better grasp of how various city components interact with one another. Advances in geosciences now allow for the automatic creation of high-quality, realistic 3D city models. It is not limited to visualization and navigation, however, also for shadow and solar potential analysis. Solar radiation is an example of a 3D GIS tool that is in high demand. The calculation of solar radiation that reaches 3D objects can be simple, but the shadow effect of nearby buildings is a considerably more challenging issue because some facades or roofs are only partially shadowed. The present study is analyzed into two approaches. The first approach is considered as Visualization (client-side) approach to visualize the 3D city models on the website using NodeJS and CesiumJS. The second approach is considered as Analyzation (Server-side) approach to analyze the solar potential using python for faster processing and deeming the future development aspects.
... For instance, Lafarge et al. (2010) reconstructed building models with a set of predefined libraries where geometric regularity is explicitly imposed. Also, Weidner and Förstner (1995) adopted the MDL concept, which depends on the mutual fit of the data and model and on the complexity of the model, to regularize noisy building boundary. Furthermore, Jung et al. (2017) extended the MDL-based regularization method to implicitly derive the shape regularity of 3D building rooftops from given noisy information of building boundary in a progressive manner. ...
... "Model complexity", which fluctuates based on the regularity (similar or repetitive patterns) hidden in observations, is measured to estimate this bias. Weidner and Förstner (1995) posed building outline delineation as the model selection problem using MDL. Jung et al. (2017) extended it to 3D rooftop models comprising multiple planes by implicitly generating model hypotheses. ...
In recent years, many mega-cities have provided 3D photorealistic virtual models, a digital replica of the geometrical structures of cities, for more effective decision support in public safety, urban planning, and engineering applications. Most research attempts at reconstructing geometric models of cities treat such urban systems as if they are in a static environment. However, cities are dynamic systems that continuously change over time. Accordingly, their virtual representations need to be regularly updated in a timely manner to allow for accurate analysis. The concept of progressive city modelling is to continuously reconstruct city models by accommodating changes recognized in the spatio-temporal domain, while preserving unchanged structures. This paper proposes a novel fusion method to progressively refine building rooftop models over time by integrating multi-sensor data. The proposed method integrates the line modelling cues of existing rooftop models produced by airborne laser scanning data with the new ones extracted from optical imagery. This modelling cue integration process is developed to progressively rectify geometric errors based on Hypothesize and Test optimization using Minimum Description Length. A stochastic method, Markov Chain Monte Carlo, coupled with simulated annealing, is employed to generate model hypotheses and perform a global optimization for finding the best solution. This fusion method is designed to offset the limitations of respective sensors and thus rectify various modelling errors (shape deformation, boundary displacement, and orientation errors) that are often involved in rooftop building models. The performance evaluation tested over the ISPRS show the proposed modelling method can achieve the improvements of 1.8%, 0.54°, 0.33 m, and 0.007 for the quality, orientation difference, Hausdorff distance, and turning function distance, respectively, compared with initial building models. In addition, the proposed methods show the highest performance in the quality measure among the state of the art methods, while demonstrates competitive performance in the completeness and correctness measures.
... With the rapid development of remote sensing technology, high-resolution (HR) and very-high-resolution (VHR) remote sensing images have been popularized in practical applications. In particular, many scholars have focused on the study of building extraction (BE) and building change detection algorithms in urban scenes [1][2][3]. As one of the most valuable Earth surface targets, effective BE can be used for dynamic assessment and monitoring of urban development, urban disaster assessment and monitoring, and urban management and construction [4][5][6]. ...
Buildings are one of the most important goals of human transformation of the Earth’s surface. Therefore, building extraction (BE), such as in urban resource management and planning, is a task that is meaningful to actual production and life. Computational intelligence techniques based on convolutional neural networks (CNNs) and Transformers have begun to be of interest in BE, and have made some progress. However, the BE methods based on CNNs are limited by the difficulty in capturing global long-range relationships, while Transformer-based methods are often not detailed enough for pixel-level annotation tasks because they focus on global information. To conquer the limitations, a multi-scale Transformer (MSTrans) is proposed for BE from high-resolution remote sensing images. In the proposed MSTrans, we develop a plug-and-play multi-scale Transformer (MST) module based on atrous spatial pyramid pooling (ASPP). The MST module can effectively capture tokens of different scales through the Transformer encoder and Transformer decoder. This can enhance multi-scale feature extraction of buildings, thereby improving the BE performance. Experiments on three real and challenging BE datasets verify the effectiveness of the proposed MSTrans. While the proposed approach may not achieve the highest Precision and Recall accuracies compared with the seven benchmark methods, it improves the overall metrics F1 and mIoU by 0.4% and 1.67%, respectively.
... S EMANTIC segmentation in remote sensing imagery is to classify geographic spatial data at the pixel level, thereby enhancing the understanding and analysis of the observed landscape [1]- [3]. Semantic segmentation extensively applied in various fields of remote sensing, encompassing tasks such as land use and land cover (LULC) [4]- [6], building extraction [7]- [9], impervious surface mapping [10]- [13], landslide mapping [14], [15], and others. We counted the proportion of these tasks in semantic segmentation, as shown in the Fig. 1. ...
With the advent of the era of high-resolution remote sensing, semantic segmentation methods for solving pixel-level classification have been widely studied. Deep learning has significantly advanced deep feature extraction methods, becoming widely employed in remote sensing image analysis. Deep feature fusion methods is able to effectively combine features from different sources. Optical and synthetic aperture radar (SAR) images stand out as primary data sources in remote sensing, offering complementary and consistent information. Fusion of deep semantic features of optical and SAR images can alleviate the limitations of single-source images in application and improve semantic segmentation accuracy. Therefore, this article reviews the research on deep fusion of optical and SAR images in semantic segmentation tasks from four aspects. Firstly, we provide a summary of challenges and research methods pertinent to semantic segmentation of remote sensing images. Then the challenges and urgent needs of deep feature fusion of optical and SAR images are analyzed, and current research is summarized from the perspective of structural design by studying various feature fusion strategies. Additionally, the compilation and in-depth analysis of open-source optical and SAR datasets suitable for semantic segmentation are undertaken, serving as fundamental resources for future research endeavors. Finally, the article identifies the major challenges summarized from the literature review in this field, outlining expectations and potential future directions for researchers.
... -знания о геометрической форме и радиометрических свойствах объекта [10]; ...
The classification of methods for land surface image segmentation is presented in the paper. Such approaches as template matching, machine learning and deep neural networks, as well as application of knowledge about analyzed objects are considered. Peculiarities of vegetation indices application for satellite images data segmentation are considered. Advantages and disadvantages are noted. The results obtained by the authors of the methods that have appeared over the last 10 years are systematized, which will allow those interested to get oriented faster and form ideas for further research.
... But it suffers from low detection accuracy. [62] proposed a simple method for extracting the three-dimensional shape of buildings from highresolution digital elevation models (DEMs) with a grid resolution of 0.5 to 5 m. The knowledge used relates to an object that is consistent, simple, and transparent with data of varying density and resolution. ...
Today, the presentation of a three-dimensional model of real-world features is very important and widely used and has attracted the attention of researchers in various fields, including surveying and spatial information systems, and those interested in the three-dimensional reconstruction of buildings. The building is the key part of the information in a three-dimensional city model, so extracting and modeling buildings from remote sensing data is an important step in building a digital model of a city. LiDAR technology due to its ability to map in all three modes of one-dimensional, two-dimensional, and three-dimensional is a suitable solution to provide hyperspectral and comprehensive images of the building in an urban environment. In this review article, a comprehensive review of the methods used in identifying buildings from the past to the present and appropriate solutions for the future is discussed.
... For example, details of building structure including age, construction method and material, and height, are key information for modeling disaster risk in seismically active regions [4,[15][16][17]. The automatic mapping of 2D urban growth using satellite data and image classification techniques is now widespread, e.g., [18,19]; however, retrieving the building-level 3D structure of a city generally requires expensive aerial imagery or light detection and ranging (LiDAR) surveys [20,21] or high-resolution digital elevation models (DEMs) [22]. Recently, approaches that can derive aggregated building height maps over 30-90 m grids have been developed, though they are best suited to buildings less than ~20-30 m tall [23,24]. ...
Earth observation (EO) data can provide large scale, high-resolution, and transferable methodologies to quantify the sprawl and vertical development of cities and are required to inform disaster risk reduction strategies for current and future populations. We synthesize the evolution of Bishkek, Kyrgyzstan, which experiences high seismic hazard, and derive new datasets relevant for seismic risk modeling. First, the urban sprawl of Bishkek (1979–2021) was quantified using built-up area land cover classifications. Second, a change detection methodology was applied to a declassified KeyHole Hexagon (KH-9) and Sentinel-2 satellite image to detect areas of redevelopment within Bishkek. Finally, vertical development was quantified using multi-temporal high-resolution stereo and tri-stereo satellite imagery, which were used in a deep learning workflow to extract buildings footprints and assign building heights. Our results revealed urban growth of 139 km2 (92%) and redevelopment of ~26% (59 km2) of the city (1979–2021). The trends of urban growth were not reflected in all the open access global settlement footprint products that were evaluated. Building polygons that were extracted using a deep learning workflow applied to high-resolution tri-stereo (Pleiades) satellite imagery were most accurate (F1 score = 0.70) compared to stereo (WorldView-2) imagery (F1 score = 0.61). Similarly, building heights extracted using a Pleiades-derived digital elevation model were most comparable to independent measurements obtained using ICESat-2 altimetry data and field-measurements (normalized absolute median deviation < 1 m). Across different areas of the city, our analysis suggested rates of building growth in the region of 2000–10,700 buildings per year, which when combined with a trend of urban growth towards active faults highlights the importance of up-to-date building stock exposure data in areas of seismic hazard. Deep learning methodologies applied to high-resolution imagery are a valuable monitoring tool for building stock, especially where country-level or open-source datasets are lacking or incomplete.
... With the vegetated grid cells located, the vegetation height could be calculated. As discussed in Weidner and Förstner [42], the DSM differs from the DEM because it comprehends not only the ground level but the level of the objects above; therefore, by normalizing the DSM with the DEM, the height of the aboveground objects can be determined. Thus, the vegetation height was obtained by this method only at the vegetated grid cells. ...
Hydraulic models of rivers are essential for vulnerability assessment in disaster management. This study simulates the 2019 Typhoon Hagibis at the Nanakita River using a dynamic roughness model. The model estimates the roughness of the river on a pixel level from the relationship between the Manning roughness coefficient and the degree of submergence of vegetation. This degree is defined as the ratio of water depth to plant height. After validating the model, the effect of vegetation on the water level in different seasons from April 2020 to March 2021 was assessed. The vegetation area and height were obtained on a pixel level using unmanned aerial vehicle photogrammetry. The dynamic roughness model showed that the water level profile increased by 7.03% on average. The seasonal effect of vegetation was observed, revealing a strong correlation between variations in the vegetation conditions and water level profile. This approach may help mitigate flood damage by indicating the factors that can increase the risk of flooding.
... Light detection and ranging (LiDAR) points are usually used as the data basis for most building footprint recognition approaches. However, Weidner and Förstner (1995) have shown that this can be also achieved with a digital terrain model. In this terrain model, the heights of the buildings are incorporated and used to detect the jumps in their values which allows for inferring the according boundaries. ...
Building footprints are a prerequisite for many tasks such as urban mapping and planning. Such structures are mostly derived using airborne laser scanning which reveals rather roof structures than the underlying hidden footprint boundary. This paper introduces an approach to extract a 2D building boundary from a 3D point cloud stemming from either terrestrial scanning or via close-range sensing using a mobile platform, e.g. drone. To this end, a pipeline of methods including non-parametric kernel density estimation (KDE) of an underlying probability density function, a solution of the Travelling Salesperson Problem (TSP), outlier elimination and line segmentation are presented to extract the underlying building footprint. KDE turns out to be suitable to automatically determine a horizontal cut in the point cloud. An ordering of the resulting points in this cut using a shortest possible tour based on TSP allows for the application of existing line segmentation algorithms, otherwise dedicated to indoor segmentation. Outliers in the resulting segments are removed using Density-Based Spatial Clustering of Applications with Noise (DBSCAN). The segments are then generalized leading to the final footprint geometry. We applied our approach on real-world examples and achieved an IoU between 0.930 and 0.998 assessed by ground truth footprints from both authoritative and volunteered geographic information (VGI) data.
... To ameliorate adverse social problems, building density regulations (such as those for building heights or floor area ratios) are common practices in urban planning and management worldwide [6]. Various remote sensing products and classification methods have been used to extract building coverage areas [7,8] and building heights [9,10]; the nDSM [11] (the difference between a DSM and a digital terrain models (DTM)) is widely used in height estimation [12]. Ji and Tang [13] proposed three methods for gross floor area estimation from monocular optical imagery using the NoS R-CNN model. ...
Dynamic monitoring of building environments is essential for observing rural land changes and socio-economic development, especially in agricultural countries, such as China. Rapid and accurate building extraction and floor area estimation at the village level are vital for the overall planning of rural development and intensive land use and the “beautiful countryside” construction policy in China. Traditional in situ field surveys are an effective way to collect building information but are time-consuming and labor-intensive. Moreover, rural buildings are usually covered by vegetation and trees, leading to incomplete boundaries. This paper proposes a comprehensive method to perform village-level homestead area estimation by combining unmanned aerial vehicle (UAV) photogrammetry and deep learning technology. First, to tackle the problem of complex surface feature scenes in remote sensing images, we proposed a novel Efficient Deep-wise Spatial Attention Network (EDSANet), which uses dual attention extraction and attention feature refinement to aggregate multi-level semantics and enhance the accuracy of building extraction, especially for high-spatial-resolution imagery. Qualitative and quantitative experiments were conducted with the newly built dataset (named the rural Weinan building dataset) with different deep learning networks to examine the performance of the EDSANet model in the task of rural building extraction. Then, the number of floors of each building was estimated using the normalized digital surface model (nDSM) generated from UAV oblique photogrammetry. The floor area of the entire village was rapidly calculated by multiplying the area of each building in the village by the number of floors. The case study was conducted in Helan village, Shannxi province, China. The results show that the overall accuracy of the building extraction from UAV images with the EDSANet model was 0.939 and that the precision reached 0.949. The buildings in Helan village primarily have two stories, and their total floor area is 3.1 × 105 m2. The field survey results verified that the accuracy of the nDSM model was 0.94; the RMSE was 0.243. The proposed workflow and experimental results highlight the potential of UAV oblique photogrammetry and deep learning for rapid and efficient village-level building extraction and floor area estimation in China, as well as worldwide.
... The possible reason could be due to variation of extraction methods due to the nature and complexity of data used. Most of the current developed techniques use Digital Surface Model (DSM) whereby [5,6] developed Informal Settlement Modeler (ISM) using shadow and linear feature data derived from low cost small format digital imagery to extract buildings in complex buildings. The limitation of the ISM includes insufficient ground sampling data and matching errors caused by poor image quality, occlusion and shadows which leads to poor definition of building outlines. ...
... There is an expansive amount of literature on how to detect buildings from satellite imagery using traditional approaches. [1] [2] developed feature-based approaches to characterize and detect buildings. [3] presented a region-based technique for building detection. ...
... When the DTM and DSM are generated where DTM is the elevation model of the earth surface that does not include the objects above the ground surface (Podobnikar, 2005) and DSM includes the objects with their heights above the ground surface and topography (Li et al., 2004). The nDSM is generated by subtracting the DTM from DSM to separate the man-made objects with different heights over the terrain surface (Weidner & Förstner, 1995) (Figure 4). ...
Automatic buildings footprint extraction is of great importance to city planning, urban growth management, and landscape visualization. Although traditional building footprint extraction from two-dimensional images is relatively easy, but are often both time-consuming and costly. Automated building footprint extraction from imagery has been studied extensively based on image segmentation using the pixel value, while the other dimension of segmentation, such as height value, have not been fully explored to extract the building footprints that have been used in this paper. This approach uses the Digital Terrain model (DTM) and Digital Surface Model (DSM) generated from the stereo imagery using the panchromatic bands with high spatial resolution to calculate the Normalized Digital Surface Model (nDSM) to separate the features which are above the ground surface. After extraction of Elevation (Height) of each building feature, the image segmentation has been performed to separate the building features using the threshold value. When the building features are separated, the Canny Edge Detection algorithm is used to delineate the actual building boundary. After the extraction of building boundary, it is vectorized. After simplification of the vector data, the building footprints are extracted. The vector data is compared to the digitized data sets, which show that the approach can be consistent and precise as the building segmentation approach has achieved greater accuracy because of incorporation of height value. There is no human error involved in the whole process.
... When the DTM and DSM are generated where DTM is the elevation model of the earth surface that does not include the objects above the ground surface (Podobnikar, 2005) and DSM includes the objects with their heights above the ground surface and topography (Li et al., 2004). The nDSM is generated by subtracting the DTM from DSM to separate the man-made objects with different heights over the terrain surface (Weidner & Förstner, 1995) (Figure 4). ...
Automatic buildings footprint extraction is of great importance to city planning, urban growth management, and landscape visualization. Although traditional building footprint extraction from two-dimensional images is relatively easy, but are often both time-consuming and costly. Automated building footprint extraction from imagery has been studied extensively based on image segmentation using the pixel value, while the other dimension of segmentation, such as height value, have not been fully explored to extract the building footprints that have been used in this paper. This approach uses the Digital Terrain model (DTM) and Digital Surface Model (DSM) generated from the stereo imagery using the panchromatic bands with high spatial resolution to calculate the Normalized Digital Surface Model (nDSM) to separate the features which are above the ground surface. After extraction of Elevation (Height) of each building feature, the image segmentation has been performed to separate the building features using the threshold value. When the building features are separated, the Canny Edge Detection algorithm is used to delineate the actual building boundary. After the extraction of building boundary, it is vectorized. After simplification of the vector data, the building footprints are extracted. The vector data is compared to the digitized data sets, which show that the approach can be consistent and precise as the building segmentation approach has achieved greater accuracy because of incorporation of height value. There is no human error involved in the whole process.
... Besides, indoor reconstruction is also related to the task of automatically reconstructing building models from outdoor mapping data such as point clouds acquired by terrestrial or airborne laser scanning systems [618,471,213,323,601,233,645]. Here however, the typical characteristics of indoor building environments as discussed in Sec. ...
Augmented reality (AR) is generally well-suited for the interactive visualization
of all kinds of virtual, three-dimensional data directly within the physical environment
surrounding the user. Beyond that, AR holds the potential of not only
visualizing arbitrary virtual objects anywhere but to visualize geospatial data directly
in-situ in the location that the data refer to. Thus it can be used to enrich a part
of the real world surrounding the user with information about this environment and
the physical objects within it. In the scope of this work, this usage mode is defined and discussed under the term of ’fused reality’.
An appropriate scenario to demonstrate and elaborate on the potential of fused
reality is its application in the context of digital building models, where building
specific information, e.g. about the course of pipelines and cables within the walls,
can be visualized directly in the respective location. In order to realize the envisioned
concept of indoor fused reality, some principal requirements must be fulfilled. Among
these is the need for an appropriate digital model of a building environment at hand
which is to be enriched with virtual content. While building projects are nowadays
oftentimes designed and executed with the help of building information modeling
techniques, appropriate digital representations of older stock buildings are usually
hard to come by. If a corresponding model of a given building environment is
available, the respective AR device needs to be able to determine its current position
and orientation with respect to the model in order to realize a correct registration
of the physical building environment and the virtual content from the model. In
this work, different aspects about how to fulfill these requirements are investigated
and discussed.
First, different ways to map indoor building environments are discussed in order
to acquire raw data for constructing building models. In this context, an investigation
is presented about whether a state-of-the-art AR device can be deployed to this task
as well. In order to generate building models based on this indoor mapping data, a
novel, fully-automated, voxel-based indoor reconstruction method is presented and
evaluated on four datasets with corresponding ground truth data that were acquired
to this aim. Furthermore, different possibilities to localize mobile AR devices within
indoor environments are discussed and the evaluation of a straight-forward, markerbased
approach is presented. Finally, a novel method for aligning indoor mapping
data with the coordinate axes is presented and evaluated.
... The process of buildings extraction consists of three main steps: DTM creation, building detection and building reconstruction [55,56]. After obtaining the segments representing the buildings roofs and facades, information defined as attribute "SegmentID", an individual point cloud was exported for each corresponding segment using the "LAStools" software. ...
3D modelling of urban areas is an attractive and active research topic, as 3D digital models of cities are becoming increasingly common for urban management as a consequence of the constantly growing number of people living in cities. Viewed as a digital representation of the Earth’s surface, an urban area modeled in 3D includes objects such as buildings, trees, vegetation and other anthropogenic structures, highlighting the buildings as the most prominent category. A city’s 3D model can be created based on different data sources, especially LiDAR or photogrammetric point clouds. This paper’s aim is to provide an end-to-end pipeline for 3D building modeling based on oblique UAS images only, the result being a parametrized 3D model with the Open Geospatial Consortium (OGC) CityGML standard, Level of Detail 2 (LOD2). For this purpose, a flight over an urban area of about 20.6 ha has been taken with a low-cost UAS, i.e., a DJI Phantom 4 Pro Professional (P4P), at 100 m height. The resulting UAS point cloud with the best scenario, i.e., 45 Ground Control Points (GCP), has been processed as follows: filtering to extract the ground points using two algorithms, CSF and terrain-mark; classification, using two methods, based on attributes only and a random forest machine learning algorithm; segmentation using local homogeneity implemented into Opals software; plane creation based on a region-growing algorithm; and plane editing and 3D model reconstruction based on piece-wise intersection of planar faces. The classification performed with ~35% training data and 31 attributes showed that the Visible-band difference vegetation index (VDVI) is a key attribute and 77% of the data was classified using only five attributes. The global accuracy for each modeled building through the workflow proposed in this study was around 0.15 m, so it can be concluded that the proposed pipeline is reliable.
... According to former literatures, it has been reported three extraction of DSM and DTM (i.e., two products of DEM can be done using three methods (Krauß et al., 2011), namely (1) the classical morphological approach (Weidner & Förstner, 1995), (2) geodesic dilation (Arefi et al., 2009), and (3) steep edge detection (Krauß & Reinartz, 2007). The latest DTM proposed here is extracted from the development of a combination of the classical morphological approach and geodesic dilation methods. ...
The latest Digital Terrain Model (DTM) is seen as an upgradable DTM that is fitted to the latest combination of DTM master and its displacement. The latest DTM can be used to overcome the problem of static DTM weaknesses in displaying the latest topographic changes. DTM masters are obtained from InSAR and Digital Surface Model (DSM) ALOS PALSAR conversions. Meanwhile, the displacement is obtained from Sentinel-1 images, which can be updated every 6–12 days or at least every month. ALOS PALSAR data were the images acquired in 2008 and 2017, while Sentinel-1 data used were images acquired in 2018 and 2020. This study aims to reveal the importance of an upgradable DTM so called latest DTM which is combination of DTM master and its displacement in order to show the latest condition of study area. The case study is the dynamics analyze of the Semangko fault specifically in the Sianok and Sumani segments situated in Indonesia. The vertical accuracy assessment was done to evaluate the DSM to DTM conversion with a tolerance of 1.96σ. The result obtained is the latest DTM. It is derived from the integration of the DTM master with displacement. The latest DTM can be used to detect the dynamics of Semangko fault. The study area has vertical deformation at a value of –50 cm to 30 cm. The Semangko fault area is dominated by –25 to 5 cm deformation. In general, this region has decreased. The decline in this region ranges from 7.5 cm to 10 cm per year. The latest DTM vertical accuracy is 2.158 m (95% confidence level) with a scale of 1: 10,000 to 1: 20,000.
... UTOMATIC building reconstruction [1][2][3][4][5] is an important research topic in computer vision, photogrammetry, remote sensing, and other fields. With the continuous development of airborne light detection and ranging (LiDAR), fast and accurate acquisition of 3D point cloud data over large-scale urban areas has now become possible, which greatly facilitates 3D building reconstruction. ...
Automatic roof segmentation from airborne light detection and ranging (LiDAR) point cloud data is a key technology for building reconstruction and digital city modeling. In this paper, we develop a novel region expansion based L0 gradient minimization algorithm for processing unordered point cloud data, and a two-stage global optimization method consisting of the L0 gradient minimization and graph cut for roof plane segmentation. Firstly, we extract the LiDAR points of buildings from the original point cloud data with a deep learning based method and separate the points of the different buildings using Euclidean clustering to improve the processing efficiency. Secondly, region expansion based L0 gradient minimization is proposed, which is specially designed for roof plane segmentation from unordered point clouds. To fundamentally avoid the need for empirical parameter tuning in L0 gradient minimization, we propose a multi-stage coarse-to-fine segmentation process, which further improves the effect of the roof plane segmentation. Finally, graph cut is utilized to solve the jagged boundary and over-segmentation problems existing in the segmented roof planes and produce the segmentation results. We conducted comparative experiments on the Vaihingen and Hangzhou datasets. The experimental results show that the proposed approach significantly outperforms the current state-of-the-art approaches at least 6.7% and 8.9% in roof plane quality index in the Vaihingen and Hangzhou datasets, while showing superior robustness to different kinds of data.
... Existing state-of-the-art DSM to DTM filtering approaches are slope-based filtering [19], morphological filtering [20], or directional scanning e.g. Multi-directional and Slope Dependent (MSD) [16], or even deep learning methods e.g. ...
... Knowledge-based methods regard the object detection problem as a hypothesis testing problem [22], such as using the geometric knowledge of the object to detect buildings in the image [23][24][25][26]. In this type of method, the establishment of prior knowledge and rules is essential. ...
Onboard real-time object detection in remote sensing images is a crucial but challenging task in this computation-constrained scenario. This task not only requires the algorithm to yield excellent performance but also requests limited time and space complexity of the algorithm. However, previous convolutional neural networks (CNN) based object detectors for remote sensing images suffer from heavy computational cost, which hinders them from being deployed on satellites. Moreover, an onboard detector is desired to detect objects at vastly different scales. To address these issues, we proposed a lightweight one-stage multi-scale feature fusion detector called MSF-SNET for onboard real-time object detection of remote sensing images. Using lightweight SNET as the backbone network reduces the number of parameters and computational complexity. To strengthen the detection performance of small objects, three low-level features are extracted from the three stages of SNET respectively. In the detection part, another three convolutional layers are designed to further extract deep features with rich semantic information for large-scale object detection. To improve detection accuracy, the deep features and low-level features are fused to enhance the feature representation. Extensive experiments and comprehensive evaluations on the openly available NWPU VHR-10 dataset and DIOR dataset are conducted to evaluate the proposed method. Compared with other state-of-art detectors, the proposed detection framework has fewer parameters and calculations, while maintaining consistent accuracy.
... There is a world-wide established need and interest for information of built-up areas for various applications such as urban planning, urban climate studies, urban environment science investigation, city resilience and urban risk management [1]. In early 2000, the Shuttle Radar Topography Mission opened the era of Global Digital Elevation Models (DEM) processed automatically from remotely sensed data, and made available in the open scientific domain at a decametric spatial resolution [2,3]. ...
The estimation of the vertical components of built-up areas from free Digital Elevation Model (DEM) global data filtered by multi-scale convolutional, morphological and textural transforms are generalized at the spatial resolution of 250 meters using linear least-squares regression techniques. Six test cases were selected: Hong Kong, London, New York, San Francisco, Sao Paulo, and Toronto. Five global DEM and two DEM composites are evaluated in terms of 60 combinations of linear, morphological and textural filtering and different generalization techniques. Four generalized vertical components estimates of built-up areas are introduced: the Average Gross Building Height (AGBH), the Average Net Building Height (ANBH), the Standard Deviation of Gross Building Height (SGBH), and the Standard Deviation of Net Building Height (SNBH). The study shows that the best estimation of the net GVC of built-up areas given by the ANBH and SNBH, always contains a greater error than their corresponding gross GVC estimation given by the AGBH and SGBH, both in terms of mean and standard deviation. Among the sources evaluated in this study, the best DEM source for estimating the GVC of built-up areas with univariate linear regression techniques is a composite of the 1-arcsec Shuttle Radar Topography Mission (SRTM30) and the Advanced Land Observing Satellite (ALOS) World 3D–30 m (AW3D30) using the union operator (CMP_SRTM30-AW3D30_U). A multivariate linear model was developed using 16 satellite features extracted from the CMP_SRTM30-AW3D30_U enriched by other land cover sources, to estimate the gross GVC. A RMSE of 2.40 m and 3.25 m was obtained for the AGBH and the SGBH, respectively. A similar multivariate linear model was developed to estimate the net GVC. A RMSE of 6.63 m and 4.38 m was obtained for the ANBH and the SNBH, respectively. The main limiting factors on the use of the available global DEMs for estimating the GVC of built-up areas are two. First, the horizontal resolution of these sources (circa 30 and 90 meters) corresponds to a sampling size that is larger than the expected average horizontal size of built-up structures as detected from nadir-angle Earth Observation (EO) data, producing more reliable estimates for gross vertical components than for net vertical component of built-up areas. Second, post-production processing targeting Digital Terrain Model specifications may purposely filter out the information on the vertical component of built-up areas that are contained in the global DEMs. Under the limitations of the study presented here, these results show a potential for using global DEM sources in order to derive statistically generalized parameters describing the vertical characteristics of built-up areas, at the scale of 250x250 meters. However, estimates need to be evaluated in terms of the specific requirements of target applications such as spatial population modelling, urban morphology, climate studies and so on.
... Among many built-up indices like the Normalized Difference Built-up Index (NDBI), Enhanced Built-up and Bareness Index (EBBI), Urban Index (UI), Index based Built-up Index (IBI) that are used to assess built-up density, the EBBI was found to be more accurate (As-syakur, Adnyana, Arthana, & Nuarsa, 2012;Franceschetti & Iodice, 2016;Quartulli & Datcu, 2004;Vu, Yamazaki, & Matsuoka, 2009;Yang, Yin, Song, Liu, & Xu, 2014). Morphology-based methods (Chao, Yihua, Huajie, Bo, & Jinwen, 2016;Sohn & Dowman, 2007;Weidner & Förstner, 1995) like S-shaped function which included the parameters explicitly describing various characteristics of urban forms were fitted well. It also assessed built-up density for the tested samples and can be applied for non-monocentric cities and used in their planning (Jiao, 2015). ...
Urban growth is a complex spatio-temporal phenomenon that includes built-up activities taking place both horizontally and vertically. The built-up land density in a city is a function of land desirability and suitability of a location in terms of the quality of the available public services, access to infrastructure, neighborhood type, vibrancy of socio-economic and cultural characteristics. The simulation of built-up land density may help planning and development authorities in making better land developmental decisions, making appropriate provisions for services, long term land use planning, and allocation of natural resources. In the recent past, various efforts have been made to assess built-up density in terms of the densification of built-up activities at the city scale in terms of floor space indices, building density, residential density (number of housing units per hectare) and urban volume. However, only limited research was reported on the simulation and prediction of spatially distributed built-up land density. In the present study, an algorithm is developed to simulate the spatial distribution of built-up land density as a function of a set of selected urbanization explanatory variables. Its application has been demonstrated using a newly developed version of the SLEUTH model (SLEUTH-Density) to simulating the built-up land density for Ajmer city in India. Development of SLEUTH-Density included designing a density algorithm, writing the programming code, integrating the code with the existing SLEUTH model, and testing the algorithm. The model results were validated indirectly using few built-up land density indices and directly through field verification, which were found to be in good agreement with the simulated built-up land density from SLEUTH-Density.
... Among many built-up indices like the Normalized Difference Built-up Index (NDBI), Enhanced Built-up and Bareness Index (EBBI), Urban Index (UI), Index based Built-up Index (IBI) that are used to assess built-up density, the EBBI was found to be more accurate (As-syakur, Adnyana, Arthana, & Nuarsa, 2012;Franceschetti & Iodice, 2016;Quartulli & Datcu, 2004;Vu, Yamazaki, & Matsuoka, 2009;Yang, Yin, Song, Liu, & Xu, 2014). Morphology-based methods (Chao, Yihua, Huajie, Bo, & Jinwen, 2016;Sohn & Dowman, 2007;Weidner & Förstner, 1995) like S-shaped function which included the parameters explicitly describing various characteristics of urban forms were fitted well. It also assessed built-up density for the tested samples and can be applied for non-monocentric cities and used in their planning (Jiao, 2015). ...
... Among many built-up indices like the Normalized Difference Built-up Index (NDBI), Enhanced Built-up and Bareness Index (EBBI), Urban Index (UI), Index based Built-up Index (IBI) that are used to assess built-up density, the EBBI was found to be more accurate (As-syakur, Adnyana, Arthana, & Nuarsa, 2012;Franceschetti & Iodice, 2016;Quartulli & Datcu, 2004;Vu, Yamazaki, & Matsuoka, 2009;Yang, Yin, Song, Liu, & Xu, 2014). Morphology-based methods (Chao, Yihua, Huajie, Bo, & Jinwen, 2016;Sohn & Dowman, 2007;Weidner & Förstner, 1995) like S-shaped function which included the parameters explicitly describing various characteristics of urban forms were fitted well. It also assessed built-up density for the tested samples and can be applied for non-monocentric cities and used in their planning (Jiao, 2015). ...
Urban growth is a complex spatio-temporal phenomenon that includes built-up activities taking place both horizontally and vertically. The built-up land density in a city is a function of land desirability and suitability of a location in terms of the quality of the available public services, access to infrastructure, neighborhood type, vibrancy of socio-economic and cultural characteristics. The simulation of built-up land density may help planning and development authorities in making better land developmental decisions, making appropriate provisions for services, long term land use planning, and allocation of natural resources. In the recent past, various efforts have been made to assess built-up density in terms of the densification of built-up activities at the city scale in terms of floor space indices, building density, residential density (number of housing units per hectare) and urban volume. However, only limited research was reported on the simulation and prediction of spatially distributed built-up land density. In the present study, an algorithm is developed to simulate the spatial distribution of built-up land density as a function of a set of selected urbanization explanatory variables. Its application has been demonstrated using a newly developed version of the SLEUTH model (SLEUTH-Density) to simulating the built-up land density for Ajmer city in India. Development of SLEUTH-Density included designing a density algorithm, writing the programming code, integrating the code with the existing SLEUTH model, and testing the algorithm. The model results were validated indirectly using few built-up land density indices and directly through field verification, which were found to be in good agreement with the simulated built-up land density from SLEUTH-Density.
... The 3D parameter is one of the crucial parameters in building height extraction (Rottensteiner et al. 2014). There are various methods to calculate height from DEM data, like shadow-based height analysis (Weidner and Förstner 1995;Irvin and McKeown 1989;Dare 2005). The normalized digital surface model (n-DSM) is the difference between DSM and DTM, and it is the widely used method in height estimations. ...
A high-resolution dataset, such as an unmanned aerial vehicle (UAV) data provides new insight of information extraction for remote sensing applications. Object-based image analysis (OBIA) is emerging as an effective tool in the field of aerial image processing and remote sensing applications. The study primarily demonstrates how UAV data can be utilized for the extraction of urban land spatial information and aims explicitly at the extraction of vacant urban parcels within city premises. The study is initiated with object-based urban feature extraction using Multiresolution segmentation (MRS). Further, classification is performed by defining a set of rules to extract vacant urban parcel spatial information. Digital elevation and normalized surface models (DEM and n-DSM) are utilized for refining the segmentation results. The attribution and reclassification of objects are performed based on DEM and n-DSM values. Moreover, the challenges for removing the obligations in delineating the vacant parcel boundaries are addressed by utilizing the excess vegetation index (EVI). The applicability of the approach is examined by three accuracy indexes, which are completeness, correctness, and quality. Overall high accuracy is obtained for extracted urban land parcels in terms of accuracy indexes. The proposed algorithm can be effectively utilized for numerous applications such as building floor extraction, gathering information for vacant urban parcels within city premises, delineation of building footprints, damaged building estimations, and many more.
... The majority of traditional methods develop rulesets or thresholds to combine spatial and spectral features. Early methods were mainly developed for medium spatial resolution images and digital surface models (DSMs), including minimum description length based on morphological operations, 10 structural, textural, and spectral descriptors for a neuro-fuzzy strategy, 11 data fusionbased approaches, 12 snake models or active contours, 13,14 shadow information and graph cuts, 15 fuzzy object-based analysis, 16 the purposive fast independent component analysis (FastICA) algorithm, 17 and fusion of supervised and unsupervised methods. 18 These approaches use initial spatial or spectral information of images or DSMs such as the gray-level co-occurrence matrix for building footprint extraction. ...
Building footprint segmentation from satellite and aerial images is an essential and challenging step for high-resolution building map generation. In urban management applications, such as building monitoring, infrastructure development, smart three-dimensional cities, and building change detection, building footprints are required to generate precise multiscale building maps. An efficient deep learning-based segmentation approach is proposed for multiscale building footprint extraction, and the results are presented for the most important challenges in photogrammetry and remote sensing, including shadows and occluded areas, vegetation covers, complex roofs, dense building areas, oblique images, and the generalization capability in different locations. The proposed method includes new dilated convolutional blocks containing kernels with different sizes to learn spectral-spatial relationships in multiscale satellite and aerial images with a high level of abstraction. The quantitative assessments of multiscale images from different locations with different spatial resolutions and spectral details show that the average F1 score and the average intersection over union for extracted footprints are about 86% and 76%, respectively. Compared with the state-of-the-art approaches, the proposed method has outstanding generalization capability and provides better performance for building footprint segmentation from multisensor single images. © 2020 Society of Photo-Optical Instrumentation Engineers (SPIE).
... Knowledge-based methods treat the object detection problem as a hypothesis testing process by using preestablished knowledge and rules. Two kinds of well-known knowledge are geometric knowledge [32], [33], [34], [35] and context knowledge [36], [32], [37]. OBIA-based methods start with segmenting images into homogeneous regions that represent a relatively homogeneous group of pixels and then perform region classification using region-level features from hand-crafted feature engineering. ...
In this paper, we deal with the problem of object detection on remote sensing images. Previous methods have developed numerous deep CNN-based methods for object detection on remote sensing images and the report remarkable achievements in regard to both detection performance and efficiency. However, current CNN-based methods mostly require a large number of annotated samples to train deep neural networks and tend to have limited generalization abilities for unseen object categories. In this paper, we introduce a few-shot learning-based method for object detection on remote sensing images where only a few annotated samples are provided for the unseen categories. More specifically, our model contains three main components: a meta feature extractor that learns to extract feature representations from input images, a reweighting module that learn to adaptively assign different weights for each feature representation from the support images, and a bounding box prediction module that carries out object detection on the reweighted feature maps. We build our few-shot object detection model upon YOLOv3 architecture and develop a multi-scale object detection framework. Experiments on two benchmark datasets demonstrate that with only a few annotated samples our model can still achieve a satisfying detection performance on remote sensing images and the performance of our model is significantly better than the well-established baseline models.
... Although many building extraction approaches have been developed [3] , but it is hard to distinguish buildings from backgrounds using spectral-based classification such as support vector machine (SVM) and random forest [4] due to different materials of building roofs. However, traditional building extraction methods based on pixel level predictions, tend to produce inconsistent extraction results with "salt-and-pepper" noise in high-resolution imagery [38] . ...
Building extraction is one of the important tasks for urbanization monitoring, city planning, and urban change detection. It is not an easy task due to spectral heterogeneity and structural diversity of the complex backgrounds. In this paper, an object-based multi-modal convolution neural networks (OMM-CNN) is proposed for building extraction using panchromatic and multispectral imagery. Specifically, a multi-modal deep CNN (the multispectral CNN and the panchromatic CNN) architecture is designed which can mine multiscale spectral-spatial contextual information. In order to fully explore the spatial-spectral information embedded in panchromatic and multispectral images, the complex convolution and complex self-adaption pooling layer are developed. Furthermore, to improve the building extraction accuracy and efficiency, a simple linear iterative clustering (SLIC) algorithm is used to segment the panchromatic and multispectral remote sensing imagery simultaneous. Results demonstrated that the proposed method can extract different types of buildings, and the result is more accurate and effective than that of the recent building extraction methods.
... Applied on each measurement of the DSM it is possible to derive a DTM with this operation (e.g. Harlick et al. 1987;Weidner & Förstner 1995), which is created by a dilation executed on the result of an erosion (Soille 2004, 66ff.): ...
The urban areas around the world are growing and constantly changing. One of the developments that can be observed alongside, especially in the Global North, is the formation of several densification patterns, so-called (sub-)centers, in addition to the traditional center. This phenomenon, known under the scientific term polycentricity, has so far mainly been analysed based on socioeconomic data on employees, population or human activity data, to name a few. These approaches, however, are reaching their limits. Due to data heterogeneity, limited data availability and diverse methodological strategies to identify center structures, comparisons between cities or regions are difficult. Since remote sensing enables to compute the urban morphology on large scale, it has the potential to support the social sciences. Against this background, within this thesis I used remotely sensed data to calculate urban mass concentrations, defined as volume per processing unit, in four German cities and four cities in the USA. The outcome is a morphological 3D characterization of the urban structure. In the scope of this thesis the approach used to generate these urban mass concentrations is applied for the first time to cities in the USA. Originating from the results, in a next step, high urban mass concentrations interpreted as (sub-)centers are identified by applying a threshold method. Based on the identified (sub-)centers a polycentricity analysis is finally carried out to compare the distribution of (sub-)centers within Germany and the USA as well as between Germany and the USA. Concluding, the thesis investigates the phenomenon of polycentricity from a morphological perspective in order to contribute to the understanding of the changes happening in today’s urban environments of the Global North. At the same time, it is a contribution to the further development of polycentricity analyses.
... In theory, the object height may be calculated from the UAV imagery through photogrammetry by subtracting the DTM from the DSM of the plant [37]. It is difficult to extract a DTM from a UAVderived DSM because part of the terrain surface is sheltered by plant canopy [38]. ...
Carbon sink trading is an important aspect of carbon trading in China, and can have important significance in offsetting carbon emissions and improving ecological compensation. The use of unmanned aerial vehicles (UAVs) offers new opportunities for shrub carbon sink and accounts as a substitute for time-consuming and expensive plot investigations to estimate the carbon sink by using the aboveground carbon stock monitored by UAV. However, the UAV-based estimation of the aboveground carbon stock of densely planted shrubs still faces certain challenges. The specific objectives of this research are as follows: (1) to test the statistical relationship between the aboveground carbon stock and volume of a densely planted shrub belt, and (2) to develop a model to estimate aboveground carbon stock by monitoring the volume of the densely planted shrub belt using a UAV. The study showed that (ⅰ) the aboveground carbon stock would increase with the increase in the volume of the shrub belt, (ⅱ) an estimation model of the aboveground carbon stock of the densely planted shrub belt was developed (R2=0.89, P<0.01), and (iii) the validation assessment to estimate aboveground carbon stock by using the UAV-based estimation model produced a coefficient of determination of R2 = 0.74 and an overall root mean square error of 18.79 kg CO2e. Good prediction ability of the model was determined using leave-one-out cross-validation (LOOCV). This output information is valuable for the design of operations in the framework of precise carbon-sink accounting of shrubs. In addition, a method using an UAV was developed and validated for the quick estimation of aboveground carbon stock for densely planted shrubs, thereby providing a potential alternative to time-consuming and expensive plot investigations of aboveground carbon-stock accounting, which is necessary for shrub projects in the carbon trading market in China.
... V. Masson, et al. Urban Climate 31 (2020) 100536 vertical objects not related to the terrain topography (e.g., buildings), a normalized DSM (nDSM) is usually calculated by subtracting the modelled terrain height (digital terrain model) from the DSM (Weidner and Förstner 1995;Reinartz et al. 2017). The traditional technique for DSM/nDSM generation and subsequent analysis of the urban morphology is photogrammetric processing of optical stereo imagery from sensors mounted on airplanes (Fradkin et al. 1999;Hirschmüller et al. 2005) or from high and very-high resolution satellite data (Toutin 2006;Eckert and Hollands 2010;Sirmacek et al. 2012;Aguilar et al. 2014). ...
Cities are particularly vulnerable to meteorological hazards because of the concentration of population, goods, capital stock and infrastructure. Urban climate services require multidisciplinary and multi-sectorial approaches and new paradigms in urban climate modelling. This paper classifies the required urban input data for both mesoscale state-of-the-art Urban Canopy Models (UCMs) and microscale Obstacle Resolving Models (ORM) into five categories and reviews the ways in which they can be obtained. The first two categories are (1) land cover, and (2) building morphology. These govern the main interactions between the city and the urban climate and the Urban Heat Island. Interdependence between morphological parameters and UCM geometric hypotheses are discussed. Building height, plan and wall area densities are recommended as the main input variables for UCMs, whereas ORMs require 3D building data. Recently, three other categories of urban data became relevant for finer urban studies and adaptation to climate change: (3) building design and architecture, (4) building use, anthropogenic heat and socioeconomic data, and (5) urban vegetation data. Several methods for acquiring spatial information are reviewed, including remote sensing, geographic information system (GIS) processing from administrative cadasters, expert knowledge and crowdsourcing. Data availability, data harmonization, costs/efficiency trade-offs and future challenges are then discussed.
... V. Masson, et al. Urban Climate 31 (2020) 100536 vertical objects not related to the terrain topography (e.g., buildings), a normalized DSM (nDSM) is usually calculated by subtracting the modelled terrain height (digital terrain model) from the DSM (Weidner and Förstner 1995;Reinartz et al. 2017). The traditional technique for DSM/nDSM generation and subsequent analysis of the urban morphology is photogrammetric processing of optical stereo imagery from sensors mounted on airplanes (Fradkin et al. 1999;Hirschmüller et al. 2005) or from high and very-high resolution satellite data (Toutin 2006;Eckert and Hollands 2010;Sirmacek et al. 2012;Aguilar et al. 2014). ...
Cities are particularly vulnerable to meteorological hazards because of the concentration of population, goods, capital stock and infrastructure. Urban climate services require multi-disciplinary and multi-sectorial approaches and new paradigms in urban climate modelling. This paper classifies the required urban input data for both mesoscale state-of-the-art Urban Canopy Models (UCMs) and microscale Obstacle Resolving Models (ORM) into five categories and reviews the ways in which they can be obtained. The first two categories are (1) land cover, and (2) building morphology. These govern the main interactions between the city and the urban climate and the Urban Heat Island. Interdependence between morphological parameters and UCM geometric hypotheses are discussed. Building height, plan and wall area densities are recommended as the main input variables for UCMs, whereas ORMs require 3D building data. Recently, three other categories of urban data became relevant for finer urban studies and adaptation to climate change: (3) building design and architecture, (4) building use, anthropogenic heat and socio-economic data, and (5) urban vegetation data. Several methods for acquiring spatial information are reviewed, including remote sensing, geographic information system (GIS) processing from administrative cadasters, expert knowledge and crowdsourcing. Data availability, data harmonization, costs/efficiency trade-offs and future challenges are then discussed.
... Moreover, an increasing number of methods for the detection of building information are based on high-resolution digital surface models (DSMs) generated from satellite images (Lafarge, Descombes, Zerubia, & Pierrot-Deseilligny, 2010;Merciol & J o u r n a l P r e -p r o o f Lefèvre, 2015;Davydova, Cui, & Reinartz, 2016). However, the accuracy and the universality of the applicability of satellite image-based methods have been limited by the cost or accessibility of high-spatial-resolution remotely sensed data (Weidner & Förstner, 1995). Moreover, the interpretation of satellite (e.g., SAR and LiDAR) images is also complicated. ...
Available and accessible three-dimensional (3D) urban morphology data have become essential for extensive academic research on built-up environments and urban climates. A rapid and consistent methodology for extracting urban morphology information is urgently needed for sustainable urban development in global cities, particularly given future trends of rapid urbanization. However, there is still a lack of generally applicable methods that use open-source data in this context. In this study, we developed a simple and highly efficient method for acquiring 3D urban morphology information using open-source data. Building footprints were acquired from the Maps Static application programming interface. Building heights were extracted from an open digital surface model, i.e., the ALOS World 3D model with a resolution of 30 m (AW3D30). Thereafter, urban morphological parameters, including the sky view factor, building coverage ratio, building volume density, and frontal area density, were calculated based on the retrieved building footprints and building heights. The proposed method was applied to extract the 3D urban morphology of Hong Kong, a city with a complex urban environment and a highly mixed geographical context. The results show a usable accuracy and wide applicability for the newly proposed method.
... On the other hand, there are different methods based on finding the ground points to interpolate the remaining surface. These methods mostly depend on rulebased techniques that uses slope-based methods, morphological approaches (Weidner & Förstner, 1995), multi-directional scanlines (Perko, Raggam, Gutjahr, & Schardt, 2015), the cloth simulation (Zhang, et al., 2016) and the sparsity driven algorithms (Nar, Yilmaz, & Camps-Valls, 2018). Even there is a big progress achieved in automated DTM extraction methods, there are still some misclassifications in features such as cliffs, hills or small rocks. ...
A Digital Terrain Model (DTM) is a representation of the bare-earth with elevations at regularly spaced intervals. This data is captured via aerial imagery or airborne laser scanning. Prior to use, all the above-ground natural (trees, bushes, etc.) and man-made (houses, cars, etc.) structures needed to be identified and removed so that surface of the earth can be interpolated from the remaining points. Elevation data that includes above-ground objects is called as Digital Surface Model (DSM). DTM is mostly generated by cleaning the objects from DSM with the help of a human operator. Automating this workflow is an opportunity for reducing manual work and it is aimed to solve this problem by using conditional adversarial networks. In theory, having enough raw and cleaned (DSM & DTM) data pairs will be a good input for a machine learning system that translates this raw (DSM) data to cleaned one (DTM). Recent progress in topics like 'Image-to-Image Translation with Conditional Adversarial Networks' makes a solution possible for this problem. In this study, a specific conditional adversarial network implementation "pix2pix" is adapted to this domain. Data for "elevations at regularly spaced intervals" is similar to an image data, both can be represented as two dimensional arrays (or in other words matrices). Every elevation point map to an exact image pixel and even with a 1-millimeter precision in z-axis, any real-world elevation value can be safely stored in a data cell that holds 24-bit RGB pixel data. This makes total pixel count of image equals to total count of elevation points in elevation data. Thus, elevation data for large areas results in sub-optimal input for "pix2pix" and requires a tiling. Consequently, the challenge becomes "finding most appropriate image representation of elevation data to feed into pix2pix" training cycle. This involves iterating over "elevation-to-pixel-value-mapping functions" and dividing elevation data into sub regions for better performing images in pix2pix.
... Since the nDSM excludes the influence of topography, it represents the height of all overlying objects on the terrain, such as buildings, trees and greenhouses. In this way, several researchers have proposed to incorporate this 3D information as a raster layer to improve the overall accuracy classification and extraction of man-made features on built-up areas (Aguilar et al., 2014b;Luethje et al., 2017;Weidner and Förstner, 1995;Zhang et al., 2015). Recently nDSM have been also used to derive the 3D properties of urban buildings, which represent the three-dimensional nature of living spaces and are needed in population estimation or urban planning (Tomas et al., 2016). ...
Agriculture under Plastic Covered Greenhouses (PCG) has represented a step forward in the evolution from traditional to industrial farming. However, PCG-based agricultural model has been also criticized for its associated environmental impact such as plastic waste, visual impact, soil pollution, biodiversity degradation and local runoff alteration. In this sense, timely and effective PCG mapping is the only way to help policy-makers in the definition of plans dealing with the trade-off between farmers’ profit and environmental impact for the remaining inhabitants.
This work proposes a methodological pipeline for producing high added value 3D geospatial products (Digital Surface Models (DSM) and Digital Terrain Models (DTM)) from VHR satellite imagery over PCG areas. The 3D information layer provided through the devised approach could be very valuable as a complement to the traditional 2D spectral information offered by VHR satellite imagery to improve PCG mapping over large areas.
This methodological approach has been tested in Almeria (Southern Spain) from a WorldView-2 VHR satellite stereo-pair. Once grid spacing format DSM and DTM were built, their vertical accuracy was assessed by means of lidar data provided by the Spanish Government (PNOA Programme).
Regarding DSM completeness results, the image matching method based on hierarchical semi-global matching yielded much better scores (98.87%) than the traditional image matching method based on area-based matching and cross-correlation threshold (86.65%) when they were tested on the study area with the highest concentration of PCG (around 85.65% of PCG land cover). However, both image matching methods yielded similar vertical accuracy results in relation to the finally interpolated DSM, with mean errors ranging from 0.01 to 0.35 m and random errors (standard deviation) between 0.56 and 0.82 m. The DTM error figures also showed no significant differences between both image matching methods, although being highly dependent on DSM-to-DTM filtering error, in turn closely related to greenhouse density and terrain complexity.
... Most of the construction methods are to extract the roof profiles and building heights in different periods through high-resolution remote sensing images. Some literatures have studied the methods of extracting building height and floor area ratio information using remote sensing images (Weidner and Förstner, 1995;Seresht and Azizi, 2000;McIntosh and Krupnik, 2002), but the extraction accuracy and efficiency need to be improved. Based on the improvement of the above extraction method, this paper proposes the research method of multi-dimensional expansion of urban space (MEUS), and then conducts multi-angle study on MEUS. ...
The study of multi-dimensional expansion of urban space (MEUS) addresses the laws of urban spatial expansion from all directions and angles. Using Nanjing as an example, this paper constructs multi-temporal, urban three-dimensional models based on RS and GIS technology and then conducts qualitative and quantitative analysis of MEUS using plot ratio change maps and MEUS quantitative index for built-up areas. Based especially on the concept of volume growth contribution rate, this paper analyzes the characteristics of MEUS in different stages. The results show that in 2000–2004, planar expansion played the main role, the internal potential development (IPD) intensity of the urban built-up areas was relatively large, and the volume growth contribution rate was low; in 2004–2008, planar expansion accelerated, and IPD slowed down; in 2008–2012, planar expansion slowed while IPD intensity increased; the contribution rates of volume growth of urban IPD for the three periods were 22.21%, 24.51% and 73.38%, respectively. This study expands the research perspective of urban spatial expansion, and the adopted methods are instructive and meaningful for MEUS research. In addition, the results of this study will deepen the understanding of MEUS laws and help improve scientific decision-making for urban planning and urban land use management.
... Different methods of targets' and objects' detection from remote-sensing images have been studied by many researchers. These methods were categorized by Cheng and Han in 2016 into four general categories that included template matching (Weber and Lefèvre 2012), knowledgebased methods (Weidner and Förstner 1995), object-based image analysis (OBIA) methods (Hay et al. 2005), and machine learning methods (Sun et al. 2012). With the development of machine learning techniques, approaches of geospatial target detection were considered as a classification issue (Lei et al. 2012;Zhang et al. 2015). ...
Geospatial target detection from visible remote-sensing images is considered as one of the most important issues in the analysis of aerial and satellite imagery. Development of remote-sensing techniques and enhancing resolution of images provide an opportunity to advance automatic analysis. The proposed methods of geospatial target detection have faced to a variety of challenges. Recently, local features are extensively used which play a very effective role in dealing with these issues. High intensity variations between targets and backgrounds in different images are the most critical challenges. Most proposed local feature descriptors are not able to deal with this amount of intensity variations and are accompanied with errors when facing them. In this paper, PIIFD descriptor has been applied to cope with intense intensity variations, as this descriptor is symmetrical against contrast. The proposed framework to automatically detect geospatial targets includes a supervised approach based on local features extraction and description and consists of three main steps including training, image searching, and geospatial target detection. In the training step, local features are extracted by UR-SIFT algorithm that properly matches with remote-sensing images and are described by the PIIFD descriptor. Due to the large dimensions of the extracted features, the SABOVW model has been used for quantization purpose. This model uses soft assignment of features to codebook, and the presentation provided by this model is used to train SVM classifier. In the second step, for the sake of computational efficiency, the salient regions of the image are detected by combining the saliency models, which reduces the image space to search the geospatial targets. In the third step, the salient regions are scanned by the sliding window approach and a descriptor of each position is generated. Eventually, the process of detection of geospatial targets will be carried out by applying the trained SVM model to each window. In order to evaluate the efficiency of the PIIFD descriptor, its performance is compared with descriptors such as SIFT, DAISY, LSS, and LBP. The results showed better performance of the PIIFD descriptor in detection of geospatial targets.
Understanding urban dynamics and promoting sustainable development requires comprehensive insights about buildings. While geospatial artificial intelligence has advanced the extraction of such details from Earth observational data, existing methods often suffer from computational inefficiencies and inconsistencies when compiling unified building-related datasets for practical applications. To bridge this gap, we introduce the Multi-task Building Refiner (MT-BR), an adaptable neural network tailored for simultaneous extraction of spatial and attributional building details from high-resolution satellite imagery, exemplified by building rooftops, urban functional types, and roof architectural types. Notably, MT-BR can be fine-tuned to incorporate additional building details, extending its applicability. For large-scale applications, we devise a novel spatial sampling scheme that strategically selects limited but representative image samples. This process optimizes both the spatial distribution of samples and the urban environmental characteristics they contain, thus enhancing extraction effectiveness while curtailing data preparation expenditures. We further enhance MT-BR's predictive performance and generalization capabilities through the integration of advanced augmentation techniques. Our quantitative results highlight the efficacy of the proposed methods. Specifically, networks trained with datasets curated via our sampling method demonstrate improved predictive accuracy relative to those using alternative sampling approaches, with no alterations to network architecture. Moreover, MT-BR consistently outperforms other state-of-the-art methods in extracting building details across various metrics. The real-world practicality is also demonstrated in an application across Shanghai, generating a unified dataset that encompasses both the spatial and attributional details of buildings.
O uso de LASER scanner aerotransportado tem se constituído numa opção interessante para a obtenção de modelos digitais do terreno devido principalmente a sua rapidez e alta densidade de pontos medidos. Para a obtenção de um modelo digital do terreno são aplicados métodos de filtragem de pontos para eliminar os dados correspondentes a objetos acima do solo, como prédios e árvores. Embora amplamente aceitos, estes métodos podem também modificar a superfície do terreno modelado. Neste trabalho são avaliados e discutidos os resultados obtidos através de três métodos de filtragem, verificando-se regiões de maior interesse para as quais os resultados devem ser confiáveis, e também regiões mais suscetíveis a erros. Assim, é possível apontar quais situações onde se recomenda a utilização desses métodos.
Unmanned aerial vehicles (UAV) have emerged as flexible, swift, and economical imaging systems that have proven their feasibility in urban infrastructure mapping. However, data derived from such systems are not utilized thoroughly. In this study, an object-oriented, multiresolution segmentation-based workflow is explored to automatically extract the urban features such as buildings and roads that can revolutionize the pace of existing mapping methods. This paper contemplates the automatic object-oriented-based feature extraction process on 5-cm true color ortho-rectified images and digital surface model (DSM). The data was generated using the JOUAV/CW-10 model and a Sony Camera with a 40-megapixel resolution. The segmentation procedure was implemented, defining various parameters such as scale, shape, and compactness. Here, the optimum scale, shape, and compactness parameters chosen for buildings and road segmentation are 100, 0.6, and 0.8, and 50, 0.5, and 0.9, respectively. The object-based image analysis (OBIA) results were compared to manually digitized features to assess the accuracy of the automated process. The fractal border error accuracy is calculated for urban features such as roads and buildings. The OBIA results indicated that completeness, correctness, and quality of building features were 98.2%, 97.6%, and 95.9%, respectively. Similarly, the road features’ average completeness, correctness, and quality were 85.8%, 73.8%, and 68.2%, respectively, which is on the lower side due to obscuring of roads by the avenue trees. The methodology yielded promising results for urban feature extraction with substantial accuracy and can be implemented in other areas with little fine-tuning of feature extraction parameters.
The Digital Terrain Model (DTM) is an important dataset for various applications such as hydrology, environmental sciences, urban planning, etc. Majority of the satellite-based medium resolution (30 m − 90 m) Digital Surface Model (DSM) products are freely available and widely used around the world. The generation of a DTM from these products is a tedious process. In this paper, a simple algorithm is developed to classify the ground and non-ground pixels in the DSM. The algorithm is tested on ALOS – World 3 D 30 m (AW3D30) DSM for small patches in the cities of Chennai, Mumbai and Cairo. Based on general assumption of the minimum elevation difference that exists between the ground and non-ground objects a slope threshold is used to initially classify the DSM. The elevations of the classified pixels are then used iteratively to assign counter values to the pixels. These counter values are finally used to classify the data and the ground pixels are interpolated using Inverse Distance Weighted (IDW) technique to generate the DTM. The proposed algorithm is quantitatively assessed and has a Type I error of 2.33%, Type II error of 1.67%, Total error of 2% and Kappa coefficient value of 96%. Additionally, R² and RMSE values of 0.9589 and ± 0.429 m were observed when compared against a reference DTM.
In this article, we deal with the problem of object detection on remote sensing images. Previous researchers have developed numerous deep convolutional neural network (CNN)-based methods for object detection on remote sensing images, and they have reported remarkable achievements in detection performance and efficiency. However, current CNN-based methods often require a large number of annotated samples to train deep neural networks and tend to have limited generalization abilities for unseen object categories. In this article, we introduce a metalearning-based method for few-shot object detection on remote sensing images where only a few annotated samples are needed for the unseen object categories. More specifically, our model contains three main components: a metafeature extractor that learns to extract metafeature maps from input images, a feature reweighting module that learns class-specific reweighting vectors from the support images and use them to recalibrate the metafeature maps, and a bounding box prediction module that carries out object detection on the reweighted feature maps. We build our few-shot object detection model upon the YOLOv3 architecture and develop a multiscale object detection framework. Experiments on two benchmark data sets demonstrate that with only a few annotated samples, our model can still achieve a satisfying detection performance on remote sensing images, and the performance of our model is significantly better than the well-established baseline models.
URBAN ENVIRONMENT 3D STUDIES BY AUTOMATED FEATURE EXTRACTION FROM LiDAR POINT CLOUDS // Visnyk of V.N. Karazin Kharkiv National University. Series Geology, Geography, Ecology, 2020, Vol. 20 (2), pp. 156-181. ISSN 2411-3913
Research problem introduction. Both a number of necessities that require the novel technological introductions in urban studies and the challengers corresponding to these introductions have been outlined with the emphasis on the urban remote sensing tools. The research goal of this text is to outline the authors’ original contribution to the algorithmic content of the automated feature extraction upon the urban environment modeling, as well as to represent the original web-software for urban studies. AFE methods in the building detection, extraction and 3D reconstruction within the LiDAR pipeline: a thematic overview. The overall AFE algorithmic approach has been summarized proceeding from an extensional literature review due to the feature extraction from raw lidar data. A sample of the composite model of an urban feature extracted, the overall AFE algorithmic flowchart, and few MSL processed results have been visualized. Feature detection, classification, segmentation and reconstruction have been presented as constituents of the united LiDAR pipeline. EOS LiDAR Tool (ELiT) and our key original algorithmic approaches to the AFE issues. The websoftware has been developed on the base of the outlined multifunctional research approach. This software has several basic functionalities within the distributed information system: building extraction, building extraction in rural areas, change detection, and digital elevation model generation. Two basic algorithmic approaches implemented in the software have been explained in details: High Polyhedral Modeling provided by the Building Extraction tool, and Low Polyhedral Modeling provided by the Building Extraction Rural Area tool. The extensive usage of the Voronoi diagram for cluster adjacency on the finalizing modeling stage has been provided as our original update of the existing LPM methodology: its applying for the roof cluster adjacency determination and for separation of coplanar clusters, applying limited diagram for avoiding side effects of adjacency determination, its applying for the awning / overhand identification. ELiT Geoportal. The EGP has been depicted as a type of web portal used to find, access, and process LiDAR geospatial both primary, and derivative information, as well as to provide the associated geographic services (display, editing, analysis, etc.) via the Internet. The key characteristics of our Geoportal have been listed as well as some illustrations provided for the uploaded projects. Conclusion and future works. The automated feature extraction from lidar data technique has been presented with the authors’ updates as a highly promising solution for the multicomponent simulation of urban environment, that can be used for different applications for cities. The use-cases for the EGP have been outlined as hot issues: Population estimation with building geometries; Energy demand for heating and cooling; Visibility analysis in urban environment. Keywords: LiDAR, lidar data, urban environment, AFE, building model, web-GIS-application, geoportal.
China’s rural population has declined markedly with the acceleration of urbanization and industrialization, but the area under rural homesteads has continued to expand. Proper rural land use and management require large-scale, efficient, and low-cost rural residential surveys; however, such surveys are time-consuming and difficult to accomplish. Unmanned aerial vehicle (UAV) technology coupled with a deep learning architecture and 3D modelling can provide a potential alternative to traditional surveys for gathering rural homestead information. In this study, a method to estimate the village-level homestead area, a 3D-based building height model (BHM), and the number of building floors based on UAV imagery and the U-net algorithm was developed, and the respective estimation accuracies were found to be 0.92, 0.99, and 0.89. This method is rapid and inexpensive compared to the traditional time-consuming and costly household surveys, and, thus, it is of great significance to the ongoing use and management of rural homestead information, especially with regards to the confirmation of homestead property rights in China. Further, the proposed combination of UAV imagery and U-net technology may have a broader application in rural household surveys, as it can provide more information for decision-makers to grasp the current state of the rural socio-economic environment.
The location and delineation of road junctions produce valuable inputs for road network understanding and topology reconstruction; the results also serve as basic information for road construction planning, emergency rescue, and large-scale data registration. This study addresses fully automatic road junction detection and delineation from airborne lidar (or LiDAR – light detection and ranging) data. For this purpose, a higher-order tensor voting-based approach is developed, which can locate road junctions by identifying multidirectional features in encoded higher-order tensor models. This approach comprises three main steps. First, a roughness-enhanced Gabor filter is implemented to process both the lidar-derived intensity image and the normalized digital surface model to extract homogeneous elongated structural keypoints. Second, these keypoints are fed into a higher-order tensor voting algorithm to recognize road junction candidates. We modified the method for encoding higher-order tensor models to enable the algorithm to deal with both antipodal symmetric and nonsymmetric junctions. After a higher-order tensor decomposition, the number of road branches and their directions are obtained; road junction candidates were then located. Third, by matching the results with a geometric template, road junction delineation is performed to retrieve the junction center position and the road branch directions and widths. We tested our method in two study areas and assessed the quality of the results by comparing them with ground-truth data. The results proved the validity of this approach in both urban and semiurban scenarios. We also compared our method with other junction detection algorithms and analyzed the parameter influence of the approach to essential parameters.
One of the most fundamental developments in the history of photogrammetry has been the transition from analytical to digital photogrammetry. This was realized in the early 1990s through softcopy-based systems or Digital Photogrammetric Workstations (DPWs). Today, on the one hand, initial applications of digital photogrammetry in performing routine and operational procedures, such as aerial triangulation and map revision, as well as in generating geospatial datasets, including digital elevation models (DEMs) and digital orthophotos, have been essentially standardized. On the other hand, system development in automated feature extraction for diverse geospatial features have been continually improved and refined.
In dieser Arbeit wird die Laser-Profilmessung vom Flugzeug aus als ein neues Verfahren zur topographischen Geländeaufnahme behandelt. Die Motivation, sich mit dieser Methode zu beschäftigen, entsprang ursprünglich der Suche nach modernen Alternativen zur topographischen Geländeaufnahme in Waldgebieten. Der Einsatz dieser Methode vom Flugzeug aus verspricht eine Durchdringung der Vegetationsschichten, um Messungen direkt am Waldboden vornehmen zu können. Gegenüber Mikrowellenverfahren kann eine hohe Genauigkeit des Meßergebnisses erwartet werden.
In this paper we present an algorithm for parameterfree information-preserving surface restoration. The algorithm is designed for 2.5D and 3D surfaces. The basic idea is to extract noise and signal properties of the data simultaneously by variance-component estimation and use this information for filtering. The variance-component estimation delivers information on how to weigh the influence of the data dependent term and the stabilizing term in regularization techniques, and therefore no parameter which controls this relation has to be set by the user.
Knowledge about the imaging geometry and acquisition parameters
provides useful geometric constraints for the analysis and extraction of
man-made features in aerial imagery, particularly in oblique views. We
discuss the identification of horizontal and vertical lines in the
scene, using image orientation information and vanishing point
calculations, and the calculation of their dimensions. The vertical and
horizontal attributions are used to constrain the set of possible
building hypotheses, and vertical lines are extracted at corners to
estimate structure height and permit the generation of three-dimensional
building models from monocular views. Results of these techniques are
presented for nadir and oblique imagery and evaluated against manually
generated 3D ground truth building models
Working Group III/3 of ISPRS, entitled `Semantic Modelling and Image Understanding,' has organized a test on image understanding. The basic goal of this test is the integration of different information sources in the image interpretation process. In addition to aerial images all sorts of additional information can be used to derive a more reliable result, e.g. color images, stereo, GIS-planimetry. The test deals with man-made objects, for instance houses, streets, or fields of land, which are to be detected and reconstructed from the image data and further information.
This paper addresses the issue of detecting buildings in aerial images by fusing range data and iconic image information. The range and image data are assumed to be captured by an integrated sensor system from a surveying aircraft. Fusion takes place on a symbolic level of image description. The segmented range data mainly are used for the detection of buildings whereas the image primitives extracted by a grouping algorithm are used for the verification and localization of building hypotheses. Terrain heights can be acquired by an airborne laser scanner. The height data are used to detect local maxima in height which give evidence for objects rising from the terrain surface like buildings do. Because the height data is relatively sparse it is impossible to detect the border lines of the buildings from these data. Therefore candidate regions are examined further by a grouping process. Buildings very often show parallelogram structures in aerial images. The grouping process detects these structures by combining image primitives applying relations like colinearity, parallelism, symmetry, and neighborhood.
For the purposes of object or defect identification required in industrial vision applications, the operations of mathematical morphology are more useful than the convolution operations employed in signal processing because the morphological operators relate directly to shape. The tutorial provided in this paper reviews both binary morphology and gray scale morphology, covering the operations of dilation, erosion, opening, and closing and their relations. Examples are given for each morphological concept and explanations are given for many of their interrelationships.
Inverse problems, such as the reconstruction problems that arise in early vision, tend to be mathematically ill-posed. Through regularization, they may be reformulated as well-posed variational principles whose solutions are computable. Standard regularization theory employs quadratic stabilizing functionals that impose global smoothness constraints on possible solutions. Discontinuities present serious difficulties to standard regularization, however, since their reconstruction requires a precise spatial control over the smoothing properties of stabilizers. This paper proposes a general class of controlled-continuity stabilizers which provide the necessary control over smoothness. These nonquadratic stabilizing functionals comprise multiple generalized spline kernels combined with (noncontinuous) continuity control functions. In the context of computational vision, they may be thought of as controlled-continuity constraints. These generic constraints are applicable to visual reconstruction problems that involve both continuous regions and discontinuities, for which global smoothness constraints fail.
As a step toward automating the ability to locate generic objects in an image, we propose an approach based on model-driven correction of an initial low-level scene partition. To accomplish this, we define generic data structures for geometric shapes, along with robust rules for parsing the image geometry and performing a shape-motivated resegmentation. We successfully apply the system to the task of locating and outlining complex rectilinear cultural objects in aerial imagery.
The paper discusses the modeling necessary for recovering man made objects --- in this case buildings --- in complex scenes from digital imagery. The approach addresses all levels of image analysis for deriving semantically meaningful descriptions of the scene from the image, via the geometrical/physical model of the objects and their counterparts in the image. The central link between raster image and scene are network-like organized aspects of parts of the objects. This is achieved by generically modelling the objects using parametrized volume primitives together with the application specific constraints, which seems to be adequate for many types of buildings. The paper sketches the various interrelationships between the different models and their use for feature extraction, hypothesis generation and verification.
The results of terrain mapping experiments utilizing the National Aeronautics and Space Administration (NASA) Airborne Oceanographic Lidar (AOL) over forested areas are presented. The flight tests were conducted as part of a joint NASA/U.S. Army Corps of Engineers (CE) investigation aimed at evaluating the potential of an airborne laser ranging system to provide cross-sectional topographic data on flood plains that are difficult and expensive to survey using conventional techniques. The data described in this paper were obtained in the Wolf River Basin located near Memphis, TN. Results from surveys conducted under winter 'leaves off' and summer 'leaves on' conditions, aspects of day and night operation, and data obtained from decidous and coniferous tree types are compared. Data processing techniques are reviewed. Conclusions relative to accuracy and present limitations of the AOL, and airborne lidar systems in general, to terrain mapping over forested areas are discussed.
The paper describes a new approach to image segmentation. It
accepts the inherent deficiencies occuring when extracting low-level
features and when dealing with the complexity of real scenes. Image
segmentation therefore is understood as deriving a rich symbolic
description useful for tasks such as stereo or object recognition in
outdoor scenes. The approach is based on a polymorphic scheme for
simultaneously extracting points, lines and segments in a topologically
consistent manner, together with their mutual relations derived from the
feature adjacency graph (FAG) thereby performing several grouping steps
which gradually use more and more specific domain knowledge to achieve
an optimal image description. The heart of the approach is (1) a
detailed analysis of the FAG and (2) a robust estimation for validating
the found geometric hypotheses. The analysis of the FAG, derived from
the exoskeleton of the features, allows to detect inconsistencies of the
extracted features with the ideal image model, a cell-complex. The FAG
is used for finding hypotheses about incidence relations and geometric
hypotheses, such as collinearity or parallelity, also between
non-neighbored points and lines. The M-type robust estimation is used
for simultaneously eliminating wrong hypotheses on geometric
relationships. It uses a new argument for the weighting function
We describe a system for detection and description of buildings in
aerial scenes. This is a difficult task as the aerial images contain a
variety of objects. Low-level segmentation processes give highly
fragmented segments due to a number of reasons. We use a perceptual
grouping approach to collect these fragments and discard those that come
from other sources. We use shape properties of the buildings for this.
We use shadows to help form and verify the hypotheses generated by the
grouping process. This latter step also provides 3-D descriptions of the
buildings. Our system has been tested on a number of examples and is
able to work with overhead or oblique views
E×perimental accuracy anal-ysis of automatically measured digital terrain models
- Kxzystek
- D Wild
Kxzystek, E and Wild, D., 1992. E×perimental accuracy anal-ysis of automatically measured digital terrain models, h: W. F6rstner and S. Winter (Editors), Robust Computer "Vision. Wichmann, Karlsruhe, pp. 372-390.
GPS application to laser profiling and laser scanning for digital terrain models
- Krabill
Krabill, W.G., 1989. GPS application to laser profiling and laser scanning for digital terrain models. In: Proc. 42nd Photogrammetric Week, Stuttgart, pp. 329-340.
Image analysis techniques for digital photogrammetry
- Förstner