Article

Object-level change detection in spectral imagery

Authors:
To read the full-text of this research, you can request a copy directly from the author.

Abstract

Multitemporal monitoring of sites using spectral imagery is addressed. A comprehensive architecture is presented for the detection of significant changes in scene composition described at the object level of spatial scale. An object-level scene description is obtained by applying a statistical spectral anomaly detector followed by a competitive region growth object extractor. The competitive region growth algorithm is derived as the solution to an approximate maximum likelihood image segmentation problem. Gaussian spectral clustering is used to model the scene background. A digital site model is constructed that contains image segmentation maps and extracted object features. Object-level change detection (OLCD) is accomplished by comparing objects extracted from a new image to objects recorded in the site model. A restricted implementation of the architecture is described and tested on long-wave infrared hyperspectral imagery. It is demonstrated that spectral OLCD can eliminate false alarms based on their multitemporal persistence. Incorporating multiple images in the site model is observed to improve OLCD performance

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the author.

... For example, an instantaneous understanding of geometric changes within the environment and keeping environmental maps up to date are crucial for ensuring navigational accuracy in self-driving vehicles and indoor service robots. Change detection [1,2] is a technology that automatically detects environmental changes in data from sensors and has been studied using satellite [3][4][5][6][7][8], aerial [9][10][11] and street view images [12][13][14][15][16][17][18][19][20][21]. Among these, change detection using street view images can capture detailed changes that cannot be captured in satellite or aerial images; therefore, it is expected to be employed in various applications, such as the updating of maps for self-driving vehicles. ...
... Although there has been limited research on object-level change detection using street view data, several studies have been conducted using satellite images [7,[22][23][24]. In objectlevel change detection, the key is how to extract object regions and over recent years, methods that use deep learning have been proposed for this task [22][23][24]. ...
... Another important issue is the detection of changes at the object level. For example, Hazel et al. [7] proposed a method for detecting changes in spectral imagery at the object level by combining multiple steps, such as object region/feature extraction, object association and change detection. This method, which performed object extraction and mapping sequentially, is analogous to our proposed method. ...
Article
Full-text available
We developed a robust object-level change detection method that could capture distinct scene changes in an image pair with viewpoint differences. To achieve this, we designed a network that could detect object-level changes in an image pair. In contrast to previous studies, we considered the change detection task as a graph matching problem for two object graphs that were extracted from each image. By virtue of this, the proposed network more robustly detected object-level changes with viewpoint differences than existing pixel-level approaches. In addition, the network did not require pixel-level change annotations, which have been required in previous studies. Specifically, the proposed network extracted the objects in each image using an object detection module and then constructed correspondences between the objects using an object matching module. Finally, the network detected objects that appeared or disappeared in a scene using the correspondences that were obtained between the objects. To verify the effectiveness of the proposed network, we created a synthetic dataset of images that contained object-level changes. In experiments on the created dataset, the proposed method improved the F1 score of conventional methods by more than 40%. Our synthetic dataset will be available publicly online.
... This presents higher requirements and challenges for remote sensing image change detection technology. At present, many research scholars at home and abroad use optical remote sensing image, according to different changing detection target, a lot of methods and effective models are proposed (Su et al.,2011;Hazel G et al.,2001;Dai et al.,2012;Liang et al.,2013;Zhong et al.,2005;Wang et al.,2013;Sun et al.,2010). These methods can be roughly divided into 3 categories: pixel-level change detection, feature-level change detection, object-level change detection (Hazel G et al., 2001). ...
... At present, many research scholars at home and abroad use optical remote sensing image, according to different changing detection target, a lot of methods and effective models are proposed (Su et al.,2011;Hazel G et al.,2001;Dai et al.,2012;Liang et al.,2013;Zhong et al.,2005;Wang et al.,2013;Sun et al.,2010). These methods can be roughly divided into 3 categories: pixel-level change detection, feature-level change detection, object-level change detection (Hazel G et al., 2001). Traditional change detection methods mostly belong to the pixel level, the results of the change detection are generally broken, prone to the "salt and pepper phenomenon". ...
... Object-based change detection methods in common use can be divided into the following two categories. One is to carry out the object-oriented classification, and then change detection (Su et al., 2011;Hazel G et al., 2001;Liang et al., 2013). The other category is the object change vector analysis (Wang et al., 2013;Sun et al., 2010;Wu et al., 2013), which is a direct comparison method. ...
Article
Full-text available
In the process of object-oriented change detection, the determination of the optimal segmentation scale is directly related to the subsequent change information extraction and analysis. Aiming at this problem, this paper presents a novel object-level change detection method based on multi-scale segmentation and fusion. First of all, the fine to coarse segmentation is used to obtain initial objects of different sizes; then, according to the features of the objects, Change Vector Analysis is used to obtain the change detection results of various scales. Furthermore, in order to improve the accuracy of change detection, this paper introduces fuzzy fusion and two kinds of decision level fusion methods to get the results of multi-scale fusion. Based on these methods, experiments are done with SPOT5 multi-spectral remote sensing imagery. Compared with pixel-level change detection methods, the overall accuracy of our method has been improved by nearly 10%, and the experimental results prove the feasibility and effectiveness of the fusion strategies.
... This presents higher requirements and challenges for remote sensing image change detection technology. At present, many research scholars at home and abroad use optical remote sensing image, according to different changing detection target, a lot of methods and effective models are proposed (Su et al.,2011;Hazel G et al.,2001;Dai et al.,2012;Liang et al.,2013;Zhong et al.,2005;Wang et al.,2013;Sun et al.,2010). These methods can be roughly divided into 3 categories: pixel-level change detection, feature-level change detection, object-level change detection (Hazel G et al., 2001). ...
... At present, many research scholars at home and abroad use optical remote sensing image, according to different changing detection target, a lot of methods and effective models are proposed (Su et al.,2011;Hazel G et al.,2001;Dai et al.,2012;Liang et al.,2013;Zhong et al.,2005;Wang et al.,2013;Sun et al.,2010). These methods can be roughly divided into 3 categories: pixel-level change detection, feature-level change detection, object-level change detection (Hazel G et al., 2001). Traditional change detection methods mostly belong to the pixel level, the results of the change detection are generally broken, prone to the "salt and pepper phenomenon". ...
... Object-based change detection methods in common use can be divided into the following two categories. One is to carry out the object-oriented classification, and then change detection (Su et al., 2011;Hazel G et al., 2001;Liang et al., 2013). The other category is the object change vector analysis (Wang et al., 2013;Sun et al., 2010;Wu et al., 2013), which is a direct comparison method. ...
Article
In the process of object-oriented change detection, the determination of the optimal segmentation scale is directly related to the subsequent change information extraction and analysis. Aiming at this problem, this paper presents a novel object-level change detection method based on multi-scale segmentation and fusion. First of all, the fine to coarse segmentation is used to obtain initial objects of different sizes; then, according to the features of the objects, Change Vector Analysis is used to obtain the change detection results of various scales. Furthermore, in order to improve the accuracy of change detection, this paper introduces fuzzy fusion and two kinds of decision level fusion methods to get the results of multi-scale fusion. Based on these methods, experiments are done with SPOT5 multi-spectral remote sensing imagery. Compared with pixel-level change detection methods, the overall accuracy of our method has been improved by nearly 10%, and the experimental results prove the feasibility and effectiveness of the fusion strategies.
... Change detection (CD) is an important research topic that leverages quantitative analysis of multi-temporal remotely sensed images to determine the process of land cover change, especially in the monitoring of building land, urban development and disaster assessment (Hazel 2001;Hussain et al. 2013). Along with the rapid development of remotely sensed image acquisition means and the gradual shortening of the acquisition cycle, the scope of its applications is becoming increasingly widespread and the application demand is expanding. ...
... The automatic CD technology based on pixel spectral statistics is not able to meet the requirement of the extraction of change information and becomes the main obstacle for the widespread application of high resolution remotely sensed images. The emergence of object-oriented technology for highresolution remote sensing image analysis provides a new way of thinking, and the basic unit of CD has also transformed from pixel to object (Hazel 2001). Since the object-based change detection (OBCD) approach has more advantages than the pixel-based change detection (PBCD) approach, it has received extensive attention and been developed in recent years (Wang, Zhao and Zhu 2007;Emary et al. 2010; Wang and Xu et al. 2013;Hao and Shi et al. 2016;Xiao and Zhang et al. 2016;Xiao and Yuan et al. 2017). ...
Article
Full-text available
Studies based on object-based image analysis (OBIA) representing the paradigm shift in change detection (CD) have achieved remarkable progress in the last decade. Their aim has been developing more intelligent interpretation analysis methods in the future. The prediction effect and performance stability of random forest (RF), as a new kind of machine learning algorithm, are better than many single predictors and integrated forecasting method. In this paper, we present a novel CD approach for high-resolution remote sensing images, which incorporates visual saliency and RF. First, highly homogeneous and compact image super-pixels are generated using super-pixel segmentation, and the optimal segmentation result is obtained through image superimposition and principal component analysis (PCA). Second, saliency detection is used to guide the search of interest regions in the initial difference image obtained via the improved robust change vector analysis (RCVA) algorithm. The salient regions within the difference image that correspond to the binarized saliency map are extracted, and the regions are subject to the fuzzy c-means (FCM) clustering to obtain the pixel-level pre-classification result, which can be used as a prerequisite for superpixel-based analysis. Third, on the basis of the optimal segmentation and pixel-level pre-classification results, different super-pixel change possibilities are calculated. Furthermore, the changed and unchanged super-pixels that serve as the training samples are automatically selected. The spectral features and Gabor features of each super-pixel are extracted. Finally, superpixel-based CD is implemented by applying RF based on these samples. Experimental results on Ziyuan 3 (ZY3) multi-spectral images show that the proposed method outperforms the compared methods in the accuracy of CD, and also confirm the feasibility and effectiveness of the proposed approach.
... The classified objects are compared for a detailed change analysis. Objects are compared based on both the geometry and the class membership (Chant and Kelly, 2009;Hazel, 2001;Jiang and Narayanan, 2003). A theoretical framework of OBCD based on post-classification comparison was provided by Blaschke (2005) for the comparison of multi-temporal map objects to detect and identify changes. ...
Article
Full-text available
The appetite for up-to-date information about earth's surface is ever increasing, as such information provides a base for a large number of applications, including local, regional and global resources monitoring, land-cover and land-use change monitoring, and environmental studies. The data from remote sensing satellites provide opportunities to acquire information about land at varying resolutions and has been widely used for change detection studies. A large number of change detection methodologies and techniques, utilizing remotely sensed data, have been developed, and newer techniques are still emerging. This paper begins with a discussion of the traditionally pixel-based and (mostly) statistics-oriented change detection techniques which focus mainly on the spectral values and mostly ignore the spatial context. This is succeeded by a review of object-based change detection techniques. Finally there is a brief discussion of spatial data mining techniques in image processing and change detection from remote sensing data. The merits and issues of different techniques are compared. The importance of the exponential increase in the image data volume and multiple sensors and associated challenges on the development of change detection techniques are highlighted. With the wide use of very-high-resolution (VHR) remotely sensed images, object-based methods and data mining techniques may have more potential in change detection.
... Un vantaggio di questo metodo però è quello di minimizzare gli errori dovuti a problemi di normalizzazione radiometrica e atmosferica delle due immagini (Singh, 1989). per effettuare un dettagliata analisi dei cambiamenti, basata sia sulla geometria sia sulla appartanenza alle varie classi (Hazel, 2001). ...
Thesis
Full-text available
Italy is characterized by a high hydrogeological risk; in case of landslides event, occurred as a result of heavy rainfall, for the institutions responsible of the emergency management as the Civil Protection, it is very important to have a rapid mapping in order to operate in the best way. The classical methods suitable to detect and map the landslides using remote sensing data are not always quickly applicable and sometimes require operators' experience. In order to find a method that can detect the most landslide areas in the shortest possible time, it have been carried out, using the software of digital's image analysis ENVI, some data processing on an image taken by the QuickBird satellite: the study area is part of the Province of Messina, near Giampilieri and Briga. In this area, in October 2009, as a result of intense precipitation, there were hundreds of landslides, debris flows mainly; this type of landslides leave distinctive marks on the ground and they are easily detectable by remote sensing data, especially in areas mainly vegetated. At the beginning, it were performed some elaborations in order to improve the contrast between vegetated areas and soils, such as the calculation of the normalized vegetation index (NDVI), the ratio between R/IR bands (Ratio Vegetation Index), and the principal component analysis (PCA). Subsequently, it were applied textural filters based on the co-occurrence matrix (Gray Level Co-Occurrence Matrix). The application of filter "Mean", represented by the values of the local media in the editing window, on the image obtained by the ratio R / IR, result in a second image characterized by very few shades of gray, in which are recognizable possible landslides areas, even if the same gray tones are applied to population centers; the contour instead is completely black. This image can be easily transformed into a vector image. The application of other filters on the images did not give satisfactory results. Only the filter "correlation", applied to the PC4, is useful, because, containing minor information but keeping visible the major population centres, it was helpful in the distinction between the latter of these and the landslide areas. By transforming these two images in two vector files and overlaid its on a GIS software (ArcGIS), it was possible to manually remove the vectors corresponding to towns from the image obtained by the application of the filter "mean." It was also decided to carry out a manual digitization of the image relative to NDVI and PC2, in order to compare the results. Using an inventory map of the same area, created by stereoscopic analysis and field surveys (Ardizzone, et al., 2012) it was carried out a quantitative analysis, based on the index of positioning error E proposed by Carrara, et al. (1992), and also it was calculated the percentage of the area of the landslide correctly recognized. The values of E, refers to the automatically digitized image, do not differ from those found in the literature, although the percentage of the recognized landslide area is far away from the values reached by classical methods. The images created by manual vectorization has a lower value of E, so they are more accurate in terms of positioning, but the percentage of recognized landslide areas is less. In conclusion, the automatically method tested could be useful in the early hours following a fast landslide event, as debris flows, to evaluate in a short time the possible residual risks for the population and emergency management; the resulting map can then be improved by performing manual scans, waiting a more precise mapping derived from methods already well experienced in bibliography.
... I. GİRİŞ Aynı hedef bölgeden farklı zamanlarda elde edilen görüntüler arasındaki farklılıkların belirlenmesi işlemi değişim algılama olarak adlandırılmaktadır. Değişim algılama işlemlerinin, uzaktan algılama [1], medikal tanı ve tedavi [2], kentsel yapılanma [3], askeri uygulamalar [4] gibi birçok alanda kullanımı mevcuttur. Değişim etkisi, bir nesnenin hareketi, sahneye eklenmesi, çıkarılması veya farklılaşması (mevsimsel değişimler vb..) durumlarında oluşmaktadır ve ilgili piksellerin spektral değerlerinde farklılaşmaya neden olmaktadır. ...
Conference Paper
Full-text available
Change detection aims to detect the differences occurring between images of the same scene taken at different times. Hyperspectral images, due to their high spectral resolution, provide more reliable results for change detection applications with respect to other imaging systems. In this paper, Multi-band Census Transform (MCT) is proposed for change detection for hyperspectral images. Experimental results show that the proposed method improves change detection performance.
... In these circumstances, the inclusion of additional classification features -facilitated by units of analysis other than the pixelmay be used to improve change detection results. For example, kernel based texture (He et al., 2011), multi-temporal image-object texture (Desclée et al., 2006), image-object shape comparison (Boldt et al., 2012), local image correlation from kernel (Im & Jensen, 2005) and multi-temporal image-objects (Im et al., 2008) and lastly, context modelled with kernels (Volpi et al., 2013) and image-object comparison (Hazel, 2001). To summarise, if the target of interest is associated with a measurable spectral signature then the separation may be 'trivial' (Blaschke et al., 2014, p 182), opening up all available units of analysis. ...
... Detection accuracy over 90% has been achieved with SPOT-HRV images in that way. Geoffrey (Hazel, 2001) has used a competitive region growing method based on Maximum Likelihood (ML) for image segmentation, and then image background is modeled with Gaussian model and objects are extracted; Next, the segmentation results and extracted objects are organized together with a digital node model; at last, object-level change detection is completed by comparing extracted objects in the digital node models of the new and old images. AI-Khudhalry (Ai-Khudhalry, 2005) has achieved good results in detecting structural damage change information in IKONOS images by using object-oriented image segmentation and classification technology. ...
Article
Full-text available
In accordance with the characteristics of change detection based on high-resolution remote-sensing images, this paper has put forward an object-level change detection method that is based on multi-feature integration and can take into account the properties of different types of object. This method classifies the most essential change information in applications into artificial objects related change information, water-related change information and vegetation-related change information. Direct association of object types and radiation, texture and geometric features is established by analyzing the characteristics of the three types of objects. During the application of object-level change detection method, first, feature vectors of objects are constructed by controlling the weight of radiation, texture and geometric features in different ways; then feature vectors of objects in multi-temporal images are analyzed with the method of object change vector analysis to obtain the change information of object types that are sensitive to a certain feature. In order to verify the validity of this method, this paper uses the high-resolution remote-sensing images from the Internet captured before and after the Japanese earthquake on March 11, 2011 to conduct some change detection experiments based on multifeature integration. Damage information is extracted and by controlling the weight of features, building damage, damage caused by submergence of seawater and vegetation damage are detected respectively. Experiments show that the method and processing put forward in this paper, flexible, practical and adaptable, are effective in such applications as the extraction of information about damage caused by earthquake and tsunami, and investigation of land use change.
... For example, when looking for a white automobile, you can remove detections that are not on roads or parking lots. This information can be used to build site models that lead to improved spectral object level change detection (SOLCD) studies [44]. ...
... From previous discussions it emerges that the techniques developed for medium resolution data are often not effective on VHR images and that, in order to overcome the limitation of these techniques, it is important to develop novel methodologies able to integrate the spectral information with the spatial one and model the multiscale properties of the scene. In the literature some methods exist capable to exploit the above-mentioned concepts [13], [24], [25], [26], [27], [28]. Usually, the change detection problem is faced after applying supervised classification to VHR images [29], [30]. ...
... The traditional change detection methods were proposed on the basis of stable intensity images such as optical images (Singh 1989;Rignot and van Zyl 1993;Gamba et al. 2006;Geoffrey 2001;Kasetkasem and Varshney 2002;Li 2010;Chaabouni-Chouayakh et al. 2013;Akiwowo and Eftekhari 2013). However, because of the unique imaging principle of syntheticaperture radar (SAR) images, the radar cross section (RCS) of a scene target characterized by the image has certain randomness and exhibits stochastic distribution. ...
Article
Full-text available
Many methodologies of change detection have been discussed in the literature, but most of them are tested on only optical images or traditional synthetic-aperture radar (SAR) images. Few studies have investigated multipolarimetric SAR image change detection. In this study, we presented a type of multipolarimetric SAR image change detection approach based on nonsubsampled contourlet transform and multiscale feature-level fusion techniques. In this approach, Instead of denoising an image in advance, the nonsubsampled contourlet transform multiscale decomposition was used to reduce the effect of speckle noise by processing only the low-frequency sub-band coefficients of the decomposed image, and the multiscale feature-level fusion technique was employed to integrate the rich information obtained from various polarization images. Because SAR image information is dependent on scale, a multiscale multipolarimetric feature-level fusion strategy is introduced into the change detection to improve change detection precision; this feature-level fusion can not only achieve complementation of information with different polarizations and on different scales, but also has better robustness against noise. Compared with PCA methods, the proposed method constructs better differential images, resulting in higher change detection precision.
... Multitemporal images allow one to follow the evolution in time of a given region of interest by means of change detection techniques [52,33,7], and therefore represent valuable tools for natural resource management. For many advanced applications the basic data-unit of interest becomes a set of multispectral or hyperspectral images acquired at different times. ...
Article
The common framework of this thesis is the three-dimensional (3D) transform approach to the compression of visual data, as video sequences and multispectral (MS) images. Moreover, SAR images compression and lowcomplexity video coding are considered. In particular, the work focuses on 3D wavelet transform (WT), and its variations, such as motion-compensated WT or shape-adaptive WT. This approach can appear natural, as both video sequences and MS images are three-dimensional data. Nevertheless, in the video compression field, 3D-transform approaches have just begun to be competitive with hybrid schemes based on discrete cosine transform (DCT), while, as far as MS images are concerned, the scientific literature misses a comprehensive approach to the compression problem. The 3D WT approach investigated in this thesis has drawn a huge attention by researchers in the data compression field because they hoped it could reply the excellent performances its two-dimensional version achieved in still image coding. Moreover, the WT approach provides a full support for scalability, which seems to be one of the most important topics in the field of multimedia delivery research. A scalable representation of some information is made up of several subsets of data, each of which is an efficient representation of the original information. By taking all the subsets, one has the “maximum quality” version of the original data. By taking only some subsets, one can adjust several reproduction parameters (i.e. reduce resolution or quality) and save the rate corresponding to discarded layers. Such an approach is mandatory for efficient multimedia delivery on heterogeneous networks.
... Hiperspektral görüntülerde değişim tespiti başta uzaktan algılama [1] olmak üzere sivil arama-kurtarma [2] ve askeri uygulamalar [3] gibi pek çok alanda kullanılmaktadır. ...
Conference Paper
Full-text available
In this paper, a hybrid change detection approach is proposed to increase change detection accuracy in hyperspectral images and to provide robustness. Experimental results show that the proposed method improves change detection performance.
... This is the case of modifications of agricultural landscapes, either by increasing size or modifying shapes of crop plots. These changes can be approached by using texture measurements, as well as segmentation and object-oriented classification techniques (Bontemps et al., 2008;Bruzzone and Fernández-Prieto, 2000;Hazel, 2001). See Section 9.6 for a description of OBIA and segmentation. ...
... The merging operation is based on a homogeneity criteria or a combination of size and homogeneity. Examples of applications include change detection (Hazel, 2001), identification of agricultural areas (Evans et al., 2002) and identification of areas with different land uses (Chen et al., 2003b). Region growing algorithms include: region merging, region splitting, and combination of both. ...
Thesis
Full-text available
available at: https://run.unl.pt/handle/10362/19737 Currently, the Portuguese municipalities are required to produce homologated cartography, under the Territorial Management Instruments framework. The Municipal Master Plan (PDM) has to be revised every 10 years, as well as the topographic and thematic maps that describe the municipal territory. However, this period is inadequate for representing counties where urban pressure is high, and where the changes in the land use are very dynamic. Consequently, emerges the need for a more efficient mapping process, allowing obtaining recent geographic information more often. Several countries, including Portugal, continue to use aerial photography for large-scale mapping. Although this data enables highly accurate maps, its acquisition and visual interpretation are very costly and time consuming. Very-High Resolution (VHR) satellite imagery can be an alternative data source, without replacing the aerial images, for producing large-scale thematic cartography. The focus of the thesis is the demand for updated geographic information in the land planning process. To better understand the value and usefulness of this information, a survey of all Portuguese municipalities was carried out. This step was essential for assessing the relevance and usefulness of the introduction of VHR satellite imagery in the chain of procedures for updating land information. The proposed methodology is based on the use of VHR satellite imagery, and other digital data, in a Geographic Information Systems (GIS) environment. Different algorithms for feature extraction that take into account the variation in texture, color and shape of objects in the image, were tested. The trials aimed for automatic extraction of features of municipal interest, based on aerial and satellite high-resolution (orthophotos, QuickBird and IKONOS imagery) as well as elevation data (altimetric information and LiDAR data). To evaluate the potential of geographic information extracted from VHR images, two areas of application were identified: mapping and analytical purposes. Four case studies that reflect different uses of geographic data at the municipal level, with different accuracy requirements, were considered. The first case study presents a methodology for periodic updating of large-scale maps based on orthophotos, in the area of Alta de Lisboa. This is a situation where the positional and geometric accuracy of the extracted information are more demanding, since technical mapping standards must be complied. In the second case study, an alarm system that indicates the location of potential changes in building areas, using a QuickBird image and LiDAR data, was developed for the area of Bairro da Madre de Deus. The goal of the system is to assist the updating of largeviii scale mapping, providing a layer that can be used by the municipal technicians as the basis for manual editing. In the third case study, the analysis of the most suitable rooftops for installing solar systems, using LiDAR data, was performed in the area of Avenidas Novas. A set of urban environment indicators obtained from VHR imagery is presented. The concept is demonstrated for the entire city of Lisbon, through IKONOS imagery processing. In this analytical application, the positional quality issue of extraction is less relevant.
... Although this is a human -computer interaction process, the outcome cannot be satisfactory. In order to overcome the effects of the image registration, some researchers have already put forward various algorithms, such as (1) object-based algorithm [7][8][9], (2) match-based algorithm [10], namely the approach of the image registration and the change detection at the same time. The object-based approach [9] is mainly dependent on the accuracy of the target detection, and the match-based approach is more complicated and difficult than the others. ...
... Even though image segmentation has been heavily studied in image processing and computer vision fields, and despite the early efforts that use spatial information for classification of remotely sensed imagery [3] , segmentation algorithms have only recently started receiving emphasis in remote sensing image analysis. Examples of image segmentation in the remote sensing literature include region growing [4] and Markov random field models [5] for segmentation of natural scenes, hierarchical segmentation for image mining [6] , region growing for object level change detection [7] , and boundary delineation of agricultural fields [8] . First beginning with the availability of very high resolution imagery and their characteristics this method has become popular as a common variant of data interpretation. ...
Article
Full-text available
The methods of segment-based image analysis are becoming more and more important for remote sensing as a result of the progresses in spatial resolution of satellite image. An approach to segmentation of IKONOS panchromatic image based on frequency domain filtering and marker-controlled watershed transform is presented in the paper. Primarily the texture and edge features are extracted from the response of log Gabor filtering. The texture features are obtained from the amplitude response, and phase congruency is introduced as a new method to detect invariant edge features. Then an approach to combining texture with edge features is presented and used to implement the marker-controlled watershed segmentation. Combination of different frequency texture features is used to mark different complicated images. Finally empirical discrepancy is calculated to evaluate the segmentation results. It shows that the precision of right segmentation is up to 80~85%. The approach presented in the paper basically satisfies the demand of feature recognition and extraction of high-resolution remotely sensed imagery.
... If the spatial neighbourhood relationship definition is not accurate enough, it will easily lead to the smooth transition of edge details, resulting in the omission phenomenon. The emergence of object-oriented technology provides a new paradigm that the basic unit of CD has transformed from pixel to object (Hazel 2001). Since the object-based change detection (OBCD) approach has great advantages over the PBCD approach, it has received extensive attention and been developed in recent years ( Wang, Zhao, and Zhu 2009;Emary, Mostafa, and Onsi 2010;Wang et al. 2013;Hao et al. 2016;Xiao et al. 2016Xiao et al. , 2017. ...
Article
Full-text available
This article presents a novel change detection (CD) approach for high-resolution remote-sensing images, which incorporates visual saliency and random forest (RF). First, highly homogeneous and compact image super-pixels are generated using super-pixel segmentation, and the optimal segmentation result is obtained through image superimposition and principal component analysis. Second, saliency detection is used to guide the search of interest regions in the initial difference image obtained via the improved robust change vector analysis algorithm. The salient regions within the difference image that correspond to the binarized saliency map are extracted, and the regions are subject to the fuzzy c-means (FCM) clustering to obtain the pixel-level pre-classification result, which can be used as a prerequisite for super-pixel-based analysis. Third, on the basis of the optimal segmentation and pixel-level pre-classification results, different super-pixel change possibilities are calculated. Furthermore, the changed and unchanged super-pixels that serve as the training samples are automatically selected. The spectral features and Gabor features of each super-pixel are extracted. Finally, super-pixel-based CD is implemented by applying RF based on these samples. Experimental results on Quickbird, Ziyuan 3 (ZY3), and Gaofen 2 (GF2) multi-spectral images show that the proposed method outperforms the compared methods in the accuracy of CD, and also confirm the feasibility and effectiveness of the proposed approach.
... Spatial-context information can be modeled by applying: fixed-shape neighborhood systems for texture information extraction Smits and Annoni (1999), Healey and Slater (1997), Li and Leung (2002), Markov Random Fields Bruzzone and Prieto (2000b), Moser et al. (2011) and morphological filters Dalla Mura et al. (2008, 2010. More advanced methods perform a context-sensitive analysis by considering adaptive neighborhoods modeled by multitemporal parcels Bovolo (2009), Prieto (2000a, 2002) and object properties Lu et al. (2011), Huo et al. (2010, Molinier et al. (2007), Carlotto (2005), Hazel (2001). They better capture the spatial correlation information present in the scene and become particularly promising for VHR images showing complex objects (e.g., buildings and other man-made structures). ...
Chapter
Full-text available
This chapter aims to present a general mathematical framework for the representation and analysis of multispectral images. It introduces two statistical models for the description of the distribution of spectral difference-vectors, and provides from them change detection methods based on image difference. The chapter presents an overview of the change detection problem in multispectral imagery and the methods proposed in the literature to address it, with emphasis on the statistical models associated with the difference image and their challenges. It also introduces the standard two-class unchange/change model for binary change detection, as derived from the hypothesis of the Gaussian distribution of natural classes in the difference image. Experiments on different image pairs from different sensors confirmed that the improved fitting of the magnitude histogram corresponds to nearly optimal change detection accuracy.
... The second group is the object-based method, emphasizing, first, creating image objects and then using them for further analysis (Hussain et al. 2013). Objectbased change detection is accomplished by comparing objects extracted from a new image with objects recorded in the site model (Hazel 2001). This classification involves the segmentation of an image into homogeneous objects followed by the analysis and classification of these objects (Whiteside et al. 2011). ...
Article
Full-text available
Pests and diseases can cause a variety of reactions in plants. In recent years, the boxwood dieback has become one of the essential concerns of practitioners and natural resources managers in Iran. To control the boxwood dieback spread, the early detection and disease distribution maps are required. The boxwood dieback causes a range of changes in colour, shape and leaf size with respect to photosynthesis and transpiration. Through remote sensing techniques, e.g. satellite image processing data, the variation of thermal and visual characteristics of the plant could be used to measure and illustrate the symptoms of the disease. In this study, five common vegetation indices like difference vegetation index (DVI), normalized difference vegetation index (NDVI), soil adjusted vegetation index (SAVI), simple ratio (SR), and plant health index (PHI) were extracted and calculated from Landsat 8 satellite image data from six regions in the Gilan province, located in the northern part of Iran out of 150 maps over the time period 2014‒2018. It turned out that among the aforementioned indices, based upon the results of the models, SR and NDVI indices were more useful for the disease spread, respectively. Our disease progression model fitting criteria showed that this technique could probably be used to assess the extent of the affected areas and also the disease progression in the investigated regions in future.
... Another operation aims at detecting change of stacked MultiTemp-HSIs [146]. Object-level change detection (OLCD) [147] and scene-wide change detection methods [148] have also been presented. Partial change detection algorithms such as multivariate alteration detection (MAD) were also applicable for radiometric normalization, which is often an important pre-processing procedure for change detection [149]. ...
Article
Since the advent of hyperspectral remote sensing in the 1980s, it has made important achievements in aerospace and aviation field and been applied in many fields. Conventional hyperspectral imaging spectrometer extends the number of spectral bands to dozens or hundreds, and provides spatial distribution of the reflected solar radiation from the scene of observation at the same time. Nowadays, with the fast development of new technology in the fields of information and photoelectricity sensing, and the popularity of unmanned aerial vehicle, hyperspectral remote sensing imaging presents the new trends of multimodality and acquires integration information while keeping high or very-high spectral resolution, especially, high temporal even real time sensing and stereo sensing. Therefore, three important modes of hyperspectral imaging come into existence: (1) multitemporal hyperspectral imaging, which refers to the observation of same region at different dates; (2) hyperspectral video imaging, which captures full frame spectral images in real-time; (3) hyperspectral stereo imaging, which obtains the full dimension information (including 2D image, elevation, and spectra) of observed scene. Along this perspective, firstly, the current researches on hyperspectral remote sensing and image processing are briefly reviewed, and then, comprehensive descriptions of the aforementioned three main hyperspectral imaging modes are carried out from the following four aspects: fundamental principle of new mode of hyperspectral imaging, corresponding scientific data acquisition, data processing and application, and potential challenges in data representation, feature learning and interpretation. Through the analysis of development trend of hyperspectral imaging and current research situation, we hope to provide a direction for future research on multimodal hyperspectral remote sensing.
... This approach is easy to implement and use in practice, and it has been widely investigated in OBCD research [13,[15][16][17]. Unlike the FOCD method, COCD can determine the change type of image objects; however, its CD accuracy is highly dependent on the accuracy of classification, which is nontrivial in practical scenarios [18][19][20]. HCD makes full use of the classification and feature extraction techniques for object detection. HCD has the advantages of the first two methods and can achieve relatively high CD accuracy. ...
Article
Full-text available
Change detection (CD) remains an important issue in remote sensing applications, especially for high spatial resolution (HSR) images, but it has yet to be fully resolved. This work proposes a novel object-based change detection (OBCD) method for HSR images that is based on region–line primitive association analysis and evidence fusion. In the proposed method, bitemporal images are separately segmented, and the segmentation results are overlapped to obtain the temporal region primitives (TRPs). The temporal line primitives (TLPs) are obtained by straight line detection on bitemporal images. In the initial CD stage, Dempster–Shafer evidence theory fuses the multiple items of evidence of the TRPs’ spectrum, edge, and gradient changes, and obtains the initial changed areas. In the refining CD stage, the association between the TRPs and their contacting TLPs in the unchanged areas is established on the basis of the region–line primitive association framework, and the TRPs’ main line directions (MLDs) are calculated. Some changed TRPs omitted in the initial CD stage are recovered by their MLD changes, thereby refining the initial CD results. Different from common OBCD methods, the proposed method considers the change evidence of TRPs’ internal and boundary information simultaneously via information complementation between TRPs and TLPs. The proposed method can significantly reduce missed alarms while maintaining a low level of false alarms in OBCD, thereby improving total accuracy. In our experiments, our method is superior to common CD methods, including change vector analysis (CVA), PCA-k-means, and iterative reweighted multivariate alteration detection (IRMAD), in terms of overall accuracy, missed alarms, and Kappa coefficient.
... Firstly, Landsat images of the study area in 1990, 2001, 2010, and 2015 were collected, then radiometric and atmospheric corrections had to be completed. Then the normalized difference vegetation index (NDVI), normalized difference water index (NDWI), and land surface temperature (LST) were estimated (Hazel 2001). Thermal band and spectral band digital number (DN) values were used to estimate the land surface temperature (Quere and Maupin 1997). ...
Book
Full-text available
The aim of this book is to provide information to scientists and local government to help them better understand the particularities of the local climate. Climate change is one of the biggest challenges to society. It can lead to serious impacts on production, life and environment on a global scale. Higher temperatures and sea level rise will cause flooding and water salinity problems which bring about negative effects on agriculture and high risks to industry and socio-economic systems in the future. Climate change leads to many changes in global development and security, especially energy, water, food, society, job, diplomacy, culture, economy and trade. The Intergovernmental Panel on Climate Change (IPCC) defines climate change as: “Any change in climate over time, whether due to natural variability or as a result of human activity.” Global climate change has emerged as a key issue in both political and economic arenas. It is an increasingly questioned phenomenon, and progressive national governments around the world have started taking action to respond to these environmental concerns.
... Firstly, Landsat images of the study area in 1990, 2001, 2010, and 2015 were collected, then radiometric and atmospheric corrections had to be completed. Then the normalized difference vegetation index (NDVI), normalized difference water index (NDWI), and land surface temperature (LST) were estimated (Hazel 2001). Thermal band and spectral band digital number (DN) values were used to estimate the land surface temperature (Quere and Maupin 1997). ...
Chapter
Full-text available
Climate change is one of the greatest threats to human security and sustainability. This chapter illustrates five key areas of its effects in relation to climate change in Bangladesh. These are (i) changes in temperature; (ii) intensity of tropical cyclones; (iii) storm surge heights; (iv) sea level rise; and (v) social vulnerability. In 2008, the Ministry of Environment and Forests revealed that Bangladesh and its adjoining areas had warmed by 0.5 °C over the preceding 100 years. The rise in temperature is generally observed in the monsoon season (June–August). An analysis of the relationship between tropical cyclones and sea surface temperatures (SSTs) for the period between 1901 and 1998 showed that despite increases in SSTs, the frequency of tropical cyclones had decreased since 1981 in the Bay of Bengal. Under the Intergovernmental Panel on Climate Change (IPCC) scenario, it is projected that tropical cyclone activity in the future is likely to decrease in the Bay of Bengal. It is predicted that rises in the mean sea level (MSL) and increases in the tropical cyclone wind speed will increase the depth of inundation along the Bangladeshi coast by more than 3 m and increase exposed areas by 69% in size. Increases in SST of up to 2 °C will increase the height of storm surges by 23% and increase areas of inundation by up to 1.26 times the present levels of inundation. Analyzing 22 years of data (1977–1988), the South Asian Meteorological Research Council (2003) showed that the relative sea levels in the Bay of Bengal have risen by 4.0 mm/year and 7.8-mm/year along the western and eastern coasts, respectively. Climate change and its associated impacts may include, but are not limited to, declines in livelihood diversity, migration, and disease. The government of Bangladesh and local residents have adopted various strategies in response to extreme events related to climate change. This review identifies further areas of research in relation to understanding of the distinctive impacts of climate change and developing synergy between institution- and community-led adaptation strategies.
... Recently, with the development of remote sensing technology, optical remote sensing image change detection has been widely used in the fields of environmental monitoring, crop measurement, urban research, ecosystem monitoring, natural disaster assessment, battlefield target strike effect evaluation and military reconnaissance [1][2][3][4][5]. There are many algorithms for remote sensing image change detection [6][7][8][9][10][11]. Image change detection is mainly divided into pixel level change detection, feature level change detection and target level change detection [12,13]. Since most of the images acquired by remote sensing satellites are optical remote sensing images, such images are easily affected by severe weather, especially by clouds and fog. ...
Article
Full-text available
The detection of changes in optical remote sensing images under the interference of thin clouds is studied for the first time in this paper. First, the optical remote sensing image is subjected to thin cloud removal processing, and then the processed remote sensing image is subjected to image change detection. Based on the analysis of the characteristics of thin cloud images, a method for removing thin clouds based on wavelet coefficient substitution is proposed in this paper. Based on the change in the wavelet coefficient, the high- and low-frequency parts of the remote sensing image are replaced separately, and the low-frequency clouds are suppressed while maintaining the high-frequency detail of the image, which achieves good results. Then, an unsupervised change detection algorithm based on a combined difference graph and fuzzy c-means clustering algorithm (FCM) clustering is applied. First, the image is transformed into a logarithmic domain, and the image is denoised using Frost filtering. Then, the mean ratio method and the difference method are used to obtain two graph difference maps, and the combined difference graph method is used to obtain the final difference image. The experimental results show that the algorithm can effectively solve the problem of image change detection under thin cloud interference.
... Automated CD technology based on pixel spectral statistics cannot satisfactorily extract change information, and this problem is the main obstacle to widespread application of high-resolution remotely sensed images. The emergence of object-oriented technology for high-resolution remote sensing image analyses provides a new paradigm in that the basic unit of CD is changed from pixel to object [11]. Because the object-based change detection (OBCD) approach offers great advantages over the PBCD approach, it has received extensive attention and has been developed in recent years [12][13][14][15][16][17]. ...
Article
Full-text available
In the process of object-based change detection (OBCD), scale is a significant factor related to extraction and analyses of subsequent change data. To address this problem, this paper describes an object-based approach to urban area change detection (CD) using rotation forest (RoF) and coarse-to-fine uncertainty analyses of multi-temporal high-resolution remote sensing images. First, highly homogeneous objects with consistent spatial positions are identified through vector-raster integration and multi-scale fine segmentation. The multi-temporal images are stacked and segmented under the constraints of a historical land use vector map using a series of optimal segmentation scales, ranging from coarse to fine. Second, neighborhood correlation image analyses are performed to highlight pixels with high probabilities of being changed or unchanged, which can be used as a prerequisite for object-based analyses. Third, based on the coarse-to-fine segmentation and pixel-based pre-classification results, change possibilities are calculated for various objects. Furthermore, changed and unchanged objects identified at different scales are automatically selected to serve as training samples. The spectral and texture features of each object are extracted. Finally, uncertain objects are classified using the RoF classifier. Multi-scale classification results are combined using a majority voting rule to generate the final CD results. In experiments using two pairs of real high-resolution remote sensing datasets, our proposed approach outperformed existing methods in terms of CD accuracy, verifying its feasibility and effectiveness.
... The first group is a post-classification-based procedure. In that method, classification approaches are applied to date 1 and date 2, and then the two classified maps are compared on a pixel-by-pixel basis in order to produce the final change map (Ahlqvist, 2008;Hazel, 2001;Serra, Pons, & Sauri, 2003;Singh, 1989;Song, Woodcock, Seto, Lenney, & Macomber, 2001;Walter, 2004). Zhou, Troy and Grove (2008) used an objectbased land-cover classification for extracting changes. ...
Article
Full-text available
The presence of phenomena such as earthquakes, floods and artificial human activities causes changes on the Earth’s surface. Change detection (CD) is an essential tool for the monitoring and managing of resources on local and global scales. Hyperspectral imagery can provide more detailed results for detecting changes in land-cover types. The main objective of this paper is to present a new, supervised CD method by combining similarity-based and distance-based methods to increase the efficiency of already existing CD approaches. The proposed method applies in two phases and uses three different algorithms, including image differencing, modified Z-score analysis and spectral angle mapper. The efficiency of the presented method is evaluated using Hyperion multi-temporal hyperspectral imagery. The receiver-operating characteristic curve index is used for assessing and comparing the results. The results clearly demonstrate the superiority of the proposed method for the detection and production of more accurate change maps. Furthermore, the proposed method is also able to detect changes with an accuracy of more than 96%, a false alarm rate lower than 0.03 and an area under the curve of about 0.986 in overall comparison to other conventional CD techniques. In addition, this method achieved an optimal threshold value with more rapid convergence.
Article
This paper proposes a geographical object-based method for change detection with high resolution images based on the changing areas distributed as a clustered type. This algorithm utilizes the Mean-Shift segmentation algorithm to extract a geographic object, and then uses the gray information of the geographic object with the EM algorithm to automatically extract changed and unchanged areas. This method considers spatial neighborhood information which can avoid the isolation and discrete disconnected areas in change results when using a pixel-based method. This method also reduces intervention when determining the change threshold value. Groups of three different spatial resolution images ( QuickBird, SPOT, TM images) are used to verify this proposed geographic object-based change detection algorithm and compared the accuracy and precision with a pixel-base method. Our results show that the accuracy with object-based change detection method on QuickBird, SPOT and TM images was 91.1%, 87.3% and 84.3%, while for the pixel-based method are 86.41%, 82.48% and 81.02% respectively. These results illustrate that the object-based change detection method has higher change detection accuracy than the pixel based approach. Moreover, the object-based method has better accuracy for high spatial resolution than in middle or low resolution images.
Article
A generalized likelihood ratio test (GLRT) statistic for spectral change detection based on the linear chromodynamics model is extended to accommodate unknown residual misregistration between imagery described by a prior probability density function for the spatial misregistration. Using a normal prior distribution leads to a fourth-order polynomial that can be numerically minimized over the unknown misregistration parameters. A more computationally efficient closed-form solution is developed based on a quadratic approximation and provides comparable results to the numerical minimization for the investigated test cases while running 30 times faster. The results applying the method to hyperspectral imagery indicate up to an order of magnitude reduction in false alarms at the same detection rate relative to baseline change detection methods for synthetically misregistered test data particularly in image regions containing edges and fine spatial features. Sensitivity to model parameters is assessed, and the method is compared with a previously published misregistration compensation approach yielding comparable results. Although the GLRT approach appears to exhibit comparable change detection performance, it offers the possibility of tailoring the algorithm to a priori knowledge of expected misregistration errors or to compensate structured misregistration as would occur due to parallax errors due to perspective variations (e.g., image parallax).
Article
Invasive species are the second greatest threat to biological diversity after habitat loss and fragmentation. Remote sensing is being used to map and model invasive species spread with increasing frequency every year. This study illustrates how remote sensing techniques can be used to map invasive species, examine the effects of environmental factors and management on their distribution, and look at the community response to spread or reduction in invasive species cover. Water hyacinth, a floating aquatic macrophyte, boasts an intimidating reputation as an aggressive global invasive. Water hyacinth was mapped in the Delta by pooling together several remote sensing techniques such as spectral indices, linear spectral unmixing, spectral mixture analysis, continuum removal, and LiDAR derived canopy height in a decision tree format with an overall accuracy greater than 90%. Water indices such as NDWI (Normalized Difference Water Index), continuum removal of water absorptions and average reflectance in the SWIR (shortwave infrared) were found to be the most important inputs for differentiating water hyacinth from other co-occuring species due to the high water content in its leaves. Water hyacinth maps from five years of data (2004 to 2008) showed that water hyacinth area significantly decreased between June of 2004 and 2005 and between June of 2007 and 2008. The decrease was larger in trapped areas (with restricted movement of water) compared to free-flowing channels. Three continuous weeks of frost in January of 2007 might have critically reduced water hyacinth cover across the Delta. There was no lasting effect of chemical control from one year to the next since no significant difference was found in water hyacinth cover in treated vs. untreated sites. The main objective of controlling water hyacinth -- to keep water channels clear for recreational and navigation use -- was not being met. Change detection analysis on the classification maps with six classes revealed that SAP (Submerged Aquatic Plants) was the most likely class to be colonized by water hyacinth and the most likely to spread into areas cleared of water hyacinth. The spread of water hyacinth into an empty niche (open water) and the return of an area after a decline in water hyacinth to open water were less common and more site dependant than with SAP. Water hyacinth interaction with other floating species was relatively limited, site dependant and more variable. The majority of water hyacinth cleared area was colonized by non-native species rather than returning to open water.
Article
Segmentation has already been recognized as a valuable and complementary approach that performs a region-based rather than a point-based evaluation of high-resolution remotely sensed data. An approach to segmentation of multispectral IKONOS image based on texture marker-controlled watershed transform is presented. Primarily the texture and edge features are extracted from the response of log Gabor filtering. The texture features are obtained from the amplitude response, and phase congruency is introduced to detect invariant edge features. Then a method for multispectral IKONOS image segmentation based on band feature combination is demonstrated. After that an algorithm to combining texture with edge features is presented and used to implement the marker-controlled watershed segmentation. Finally empirical discrepancy is calculated to evaluate the segmentation results. It shows that the precision of right segmentation rate is up to 75% to 85%.
Article
This paper addresses change detection in multitemporal remote sensing images. After a review of the main techniques developed in remote sensing for the analysis of multitemporal data, the attention is focused on the challenging problem of change detection in very-high-resolution (VHR) multispectral images. In this context, we propose a framework that aims at defining a top-down approach to the design of the architecture of novel change-detection systems for multitemporal VHR images. The proposed framework explicitly models the presence of different radiometric changes on the basis of the properties of multitemporal images, extracts the semantic meaning of radiometric changes, identifies changes of interest with strategies designed on the basis of the specific application, and takes advantage of the intrinsic multiscale/multilevel properties of the objects and the high spatial correlation between pixels in a neighborhood. This framework defines guidelines for the development of a new generation of change-detection methods that can properly analyze multitemporal VHR images taking into account the intrinsic complexity associated with these data. In order to illustrate the use of the proposed framework, a real change-detection problem has been considered, which is described by a pair of VHR multispectral images acquired by the QuickBird satellite on the city of Trento, Italy. The proposed framework has been used for defining a system for change detection in the two images. Experimental results confirm the effectiveness of the developed system and the usefulness of the proposed framework.
Conference Paper
Object-oriented change detection is to detect change by comparing the features of image objects. Its key is how to describe the images at two temporal phases with the same segmentation boundries in order to realize the one to one comparision between the two sets of segments from two different images. In this paper, the authors put forward a novel change detection method based on unitemporal image segmentation. This method can well address this problem even without the GIS auxiliary data, and performance well in the change detection of fully polarimetric SAR images.
Chapter
Synthetic aperture radar (SAR) imagery has been widely used in the field of remote sensing image change detection. However, its disadvantage of strong coherent multiplicative noise reduces the accuracy of change detection results. This paper proposes a novel SAR image change detection method, which is mainly comprised of three steps. Firstly, the difference image (DI) which is generated by log-ratio operator is segmented into superpixels by Simple Linear Iterative Clustering (SLIC) Algorithm. Secondly, superpixels are encoded uniformly in order to be utilized as the training samples, and deep neural network is used to extract deep features of DI. Finally, this paper designs an improved clustering algorithm which is optimized by Non-dominated Sorting Genetic Algorithm (NSGA-II). When the deep features of DI are used to cluster, Bhattacharyya distance between two categories of samples is selected as the similarity measurement. Taking the logarithmic likelihood function of clustering algorithm and the Bhattacharyya distance between the two categories as two optimization objectives, NSGA-II algorithm is used to optimize the model, and a set of pareto optimal solutions are thus generated. Compared with various indexes for accuracy evaluation, the map which has the highest accuracy is the final change detection map. Experimental results on real synthetic aperture radar datasets show that the proposed method is superior to other classical change detection methods, which demonstrates its effectiveness, feasibility, and superiority of the proposed method.
Chapter
Urban areas are a challenging environment because of their ever changing structure and the different temporal behaviors and spatial patterns. In this chapter a detailed analysis of some of the questions arising from the use of remotely sensed data in urban area for change detection are addressed. Specifically, the role of very high resolution sensors and their relevance with respect to either fast or slow changes in human settlement is analyzed, with specific stress on rapid mapping in specific sites (hotspots), e.g. for post-disaster damage assessment. Similarly, the possibility to exploit long temporal sequences of coarser resolution data is also explored and discussed, since the availability of huge archives is nowadays a reality that may be used to look for interesting interrelationships between urban area pattern changes and environmental changes, at both the local (town), regional and global level. Examples related to a so-called “hypertemporal” sequences of EO data are offered, and show the great potentials of these data sets.
Article
Full-text available
Change detection techniques based on high-spatial resolution imageries have been widely applied in environment monitoring, land management, dynamic monitoring of the military battlefield. In this paper, two methods including (1) object-oriented change detection based on post-classification comparison (CDBPC), (2) object-oriented change detection based on multi-feature (CDBMF) were put forward and compared to determine which method was more effective. The ample spectral information, textural information, structure information of high-spatial resolution SPOT 5 imageries were utilized synthetically in both two methods. In contrast to CDBPC, CDBMF did change classification only once, thus it avoided the accumulated classification error. Accuracy assessment shows that CDBMF is more favorable for land-use/cover change detection, and the overall accuracy has been improved significantly from 80.00% to 86.67%.
Conference Paper
In this paper, an unsupervised change detection method based on conditional random fields with texture feature (TFCRF) is designed for high spatial resolution (HSR) remote sensing images in order to make better use of the spatial information of HSR imagery. We firstly use the change vector analysis (CVA) method to calculate the difference image, and the texture features are extracted from the difference image with the help of gray level cooccurrence matrix (GLCM). Two initial change detection probabilistic maps are then acquired using the expectation maximization (EM) algorithm based on spectral and extracted texture information, respectively. Those two probabilistic maps are fused into the TFCRF algorithm using a probabilistic ensemble model to get the final binary change map. The experimental results on QuickBird and eCognition test images have shown the potential of the proposed TFCRF method in the field of change detection for HSR remote sensing images.
Article
In the process of object-oriented change detection, the determination of the optimal segmentation scale is directly related to the subsequent change information extraction and analysis. Aiming at this problem, this paper presents a novel object-level change detection method based on multi-scale segmentation and fusion. First of all, the fine to coarse segmentation is used to obtain initial objects which have different sizes; then, according to the features of the objects, the method of change vector analysis is used to obtain the change detection results of various scales. In order to improve the accuracy of change detection, this paper introduces fuzzy fusion and two kinds of decision level fusion methods to get the results of multi-scale fusion. Based on these methods, experiments are done with SPOT5 multi-spectral remote sensing imagery. Compared with pixel-level change detection methods, the overall accuracy of our method has been improved by nearly 10%, and the experimental results prove the feasibility and effectiveness of the fusion strategies.
Article
Change detection is one of the most important applications of remote sensing techniques due to its capability of repetitive acquisition imageries with consistent image quality, at short intervals, on a global scale, and during complete seasonal cycles. This paper uses two Landsat ETM+ imageries acquired in 2000 and 2002 respectively to detect change of Guangzhou in southern China during two years using post classification comparison method. Firstly, two remote sensing data are precision geometrically corrected to UTM projection with a root mean square error (RMSE) of 0.3 pixels, and then they are classfied using Maximum Likelihood method respectively. Images are classified into four classes which are water, forest, grass or crop and building,soil or unused land. Sencondly, two classified images are calculated by band geometric algorithm pixel by pixel using programming. The class value of pixel in different year is the same, and then the processed pixel is zero, whereas the processed pixel is assigned to a certain value which represents change from the one land cover type to another during two years. Finally, statistic analyses of change information during two years are computed and the post classification comparison change detection image is outputted. It concludes that the largest change areas are exchanges of building, soil or unused land with grass land, and land covers in Baiyun district are changed mostly from 2000 to 2002.
Chapter
The chapter addresses the theoretical and practical aspects of the scene change detection problem with the use of computer vision techniques. It means detecting new or disappeared objects in images registered at different moments of time and possibly in various lighting, weather, and season conditions. In this chapter, we propose the new scheme of Comparative Morphology (CM) as a generalization of the Morphological Image Analysis (MIA) scheme originally proposed by Pyt’ev. The CMs are the mathematical shape theories, which solve the tasks of the image similarity estimation, image matching, and change detection by means of some special morphological models and tools. The original morphological change detection approach is based on the analysis of difference between the test image and its projection to the shape of reference image. In our generalized approach, the morphological filter-projector is substituted by the comparative morphological filter with weaker properties, which transforms the test image guided by the local shape of reference image. Following theoretical aspects are addressed in this chapter: the comparative morphology, change detection scheme based on morphological comparative filtering, diffusion morphology, and morphological filters based on guided contrasting. Following practical aspects are addressed: the pipeline for change detection in remote sensing data based on comparative morphology and implementation of change detection scheme based on both guided contrasting and diffusion morphology. The chapter also contains the results of qualitative and quantitative experiments on a wide set of real images including the public benchmark.
Chapter
This chapter discussed the major challenges and limits when using very-high-spatial-resolution (VHR) optical remote sensor data for monitoring human settlements. Specifically, the discussion centers on the spectral aspects, mapping limits and challenges in connection to urban change analysis. Then, an integrated adaptive spatial approach is proposed to help deal with the challenges, followed by an example of the major research lines still to be pursued in this active field. Finally, the chapter discusses some interesting results using VHR data for damage mapping in connection to major events.
Article
Full-text available
This article presents a method of object-based image analysis (OBIA) for landslide delineation and landslide-related change detection from multi-temporal satellite images. It uses both spatial and spectral information on landslides, through spectral analysis, shape analysis, textural measurements using a gray-level co-occurrence matrix (GLCM), and fuzzy logic membership functionality. Following an initial segmentation step, particular combinations of various information layers were investigated to generate objects. This was achieved by applying multi-resolution segmentation to IRS-1D, SPOT-5, and ALOS satellite imagery in sequential steps of feature selection and object classification, and using slope and flow direction derivatives from a digital elevation model together with topographically-oriented gray level co-occurrence matrices. Fuzzy membership values were calculated for 11 different membership functions using 20 landslide objects from a landslide training data. Six fuzzy operators were used for the final classification and the accuracies of the resulting landslide maps were compared. A Fuzzy Synthetic Evaluation (FSE) approach was adapted for validation of the results and for an accuracy assessment using the landslide inventory database. The FSE approach revealed that the AND operator performed best with an accuracy of 93.87% for 2005 and 94.74% for 2011, closely followed by the MEAN Arithmetic operator, while the OR and AND (*) operators yielded relatively low accuracies. An object-based change detection was then applied to monitor landslide-related changes that occurred in northern Iran between 2005 and 2011. Knowledge rules to detect possible landslide-related changes were developed by evaluating all possible landslide-related objects for both time steps.
Chapter
Bangladesh possesses enormous areas of wetlands including rivers and streams, fresh water lakes and marshes, haors, baors, beels, water storage reservoirs, fishponds, flooded cultivated fields, and estuarine systems with extensive mangrove swamps. The haors, baors, beels, and jheels are of fluvial origin and are commonly identified as freshwater wetlands. Wetlands are subject to periodic inundation, changing from shallow to deep water during the wet monsoon. Most of the wetlands are inland wetlands located in the northeastern part of Bangladesh. The total area of the wetlands in this country has been estimated at 7–8 million ha, or about 50% of the total land surface. The wetlands provide habitats for many special plants, birds, mammals, reptiles, amphibians, fish, and invertebrate species. The wetlands are critically important in Bangladesh for human settlements, biodiversity, fisheries, agricultural diversity, irrigation, navigation, communication, and ecotourism. Wetlands help to reduce the impacts of flooding, maintain good water quality in rivers, recharge groundwater, store carbon, stabilize climatic conditions, and control pests. Wetlands also improve water quality by trapping sediments, filtering out pollutants, and absorbing nutrients that would otherwise result in poor water quality for downstream users.
Chapter
This chapter provides a short overview of the principles of remote sensing outlines current studies focused on the Euphrates River Basin (ERB) and presents a survey of the literature available on the topics that the thesis covers. Within the confines of this study, remote sensing is defined as the measurement of emitted or reflected electromagnetic radiation, or spectral behaviors, from a target object by a multispectral satellite sensor. This thesis contains four main sections: land use/land cover classification, the mapping of irrigated areas, irrigated agriculture mapping (especially crops classification), and land use/land cover change detection mapping. A great number of papers have been published on the above four topics. In this section a small range is given, based on significance and likeness to this thesis, with the goal of providing no wide-ranging survey, but of giving an experience of the techniques, applications and performances found in the literature.
Article
In this paper, the performance of truncated gray-coded bit-plane matching (T-GCBPM) based motion estimation in the newest video coding standard known as HEVC is assessed. Truncated gray approach is based on the usage of the most significant bit-planes. Video coding performance between truncated gray approach and MAD measure is compared. As can be seen from the experimental results that low bit-depth representation based truncated gray approach shows a minor performance decrease about 0.2-0.25 dB (PSNR) at different bit-rates compared to MAD measure. It is possible to obtain a state of the art video coding module, which has low power consumption with truncated gray-based motion estimation if implemented in hardware.
Article
In order to increase image registration accuracy in change detection, from the view of multi-scale fusion, a method was proposed for region variation detection, which applied the multi-scale analysis to remote sensing image. First of all, Wavelet transform was adopted to decompose original images, then Mahalanobis distance decision function was used to detect the changes of different scale images, and finally Markov random field was applied to fuse different scale change detection results. Since Markov random field fusion method took full account of the correlation between the adjacent pixels and the links of different scale change detection results, the fusion results were more accurate and practical. The testing results showed that this method is effective and robust.
Chapter
With its synoptic view and the repeatability, satellite remote sensing can provide timely, accurate and consistent information about earth’s surface for costeffective monitoring of environmental changes. In this chapter, recent development in change detection techniques using multitemporal remotely sensed images were reviewed. The chapter covers change detection methods for both optical and SAR images. Various aspects of change detection processes were presented including data preprocessing, change image generation and change detection algorithms such as unsupervised and supervised change detection as well as pixel-based and objectbased change detection. The review shows that significant progress has been made in the field of change detection and innovative methods have been developed for change detection using both multitemporal SAR and optical data. Attempts have been made for change detection using multitemporal multisensor/cross-sensor images. The review also identified a number of challenges and opportunities in change detection.
Book
Full-text available
Edition 1 Since its beginning, image analysis research has been evolving from heuristic design of algorithms to systematic investigation of approaches. Researchers have realized: (1) The solution to a vision problem should be sought based on optimization principles, either explicitly or implicitly, and (2) contextual constraints are ultimately necessary for the understanding of visual information in images. Two questions follow: how to define an optimality criterion under contextual constraints and how to find its optimal solution. Markov random field (MRF), a branch of probability theory, provides a foundation for the characterization of contextual constraints and the derivation of the probability distribution of interacting features. In conjunction with methods from decision and estimation theory, MRF theory provides a systematic approach for deriving optimality criteria such as those based on the maximum a posteriori (MAP) concept. This MAP-MRF framework enables us to systematically develop algorithms for a variety of vision problems using rational principles rather than ad hoc heuristics. For these reasons, there has been increasing interest in modeling computer vision problems using MRF’s in recent years. This book provides a coherent reference to theories, methodologies, and recent developments in solving computer vision problems based on MRF’s, statistics, and optimization. It treats various problems in low- and high-level computational vision in a systematic and unified way within the MAP-MRF framework. The main issues of concern are how to use MRF’s to encode contextual constraints that are indispensable to image understanding; how to derive the objective function, typically the posterior distribution, for the optimal solution to a problem; and how to design computational algorithms for finding the optimal solution. As the first thorough reference on the subject, the book has four essential parts for solving image and vision analysis problems using MRF’s: (1) introduction to fundamental theories, (2) formulations of various image models in the MAP-MRF framework, (3) parameter estimation, and (4) optimization methods.
Chapter
Visual modeling in this book is addressed mainly from the computational viewpoint. It concerns how to define an objective function for the optimal solution to a vision problem and how to find the solution. The reason for defining the solution in an optimization sense is due to various uncertainties in vision processes. It may be difficult to find the perfect solution and we usually look for some optimal solution to optimally satisfy certain constraints.
Book
Planning is a part of everybody’s activities in daily life. We need to plan our actions from the moment we get up in the morning. We plan what to do for the day, what to eat for lunch, whom to speak to, how to get to the party in time in the evening, etc. When we wish to go somewhere, we need to plan our route to get there. For example, suppose we are in a supermarket and wish to buy some vegetables. Usually, we cannot just move straight to the vegetable section from the entrance of the supermarket because we would probably run into something, e.g., a shelf, a person, or a wall. We realize that it is necessary to plan a proper motion, i.e., path, to reach our goal. This process is usually called route planning. In robotics where the process is automated, the terms path planning or motion planning are used to refer to this type of computational process of moving an object from one place to another in the presence of obstacles. This monograph focuses on computational aspects of motion planning in dynamic domains, i.e., how to plan a motion, when the environment is changing over time.
Article
A novel automatic change detection procedure has been developed using multi-spectral imagery for a number of important applications including the surveillance of enemy military installations, detection of military vehicle movements, battle damage assessments and monitoring of environmental changes. The proposed approach consists of key algorithm components that includes data normalization, image registration and distance measurement. The algorithm is selected based on performance and computational considerations for near real-time implementation. Limited study on real multi-spectral data has shown that the performance of our proposed change detection approach is far superior to that of standard techniques. Example results using M7 imagery are presented to illustrate the performance improvements of this approach over the other techniques. Bibtex entry for this abstract Preferred format for this abstract (see Preferences) Find Similar Abstracts: Use: Authors Title Abstract Text Return: Query Results Return items starting with number Query Form Database: Astronomy Physics arXiv e-prints
Article
A continuous two-dimensional region is partitioned into a fine rectangular array of sites, or ‘pixels', each pixel having a particular '‘colour’ belonging to a prescribed finite set. The true colouring of the region is unknown but, associated with each pixel, there is a possibly multivariate record which conveys imperfect information about its colour according to a known statistical model. The aim is to reconstruct the true scene, with the additional knowledge that pixels close together tend to have the same or similar colours. In this paper, it is assumed that the local characteristics of the true scene can be represented by a non-degenerate Markov random field. Such information can be combined with the records by Bayes' theorem and the true scene can be estimated according to standard criteria. However, the computational burden is enormous and the reconstruction may reflect undesirable large-scale properties of the random field. Thus, a simple, iterative method of reconstruction is proposed, which does not depend on these large-scale characteristics. The method is illustrated by computer simulations in which the original scene is not directly related to the assumed random field. Some complications, including parameter estimation, are discussed. Potential applications are mentioned briefly.
Article
Thesis (Ph. D.)--University of Maryland, College Park, Md., 2000. Thesis research directed by Dept. of Electrical and Computer Engineering. Includes bibliographical references (leaves 170-176).
Article
The fully adaptive hypothesis testing algorithm developed by I.S. Reed and X. Yu (1990) for detecting low-contrast objects of unknown spectral features in a nonstationary background is extended to the case in which the relative spectral signatures of objects can be specified in advance. The resulting background-adaptive algorithm is analyzed and shown to achieve robust spectral feature discrimination with a constant false-alarm rate (CFAR) performance. A comparative performance analysis of the two algorithms establishes some important theoretical properties of adaptive spectral detectors and leads to practical guidelines for applying the algorithms to multispectral sensor data. The adaptive detection of man-made artifacts in a natural background is demonstrated by processing multiband infrared imagery collected by the Thermal Infrared Multispectral Scanner (TIMS) instrument
Article
Gaussian Markov random field texture models and multivariate parametric clustering algorithms have been applied extensively for segmentation, restoration, and anomaly detection of single-band and multispectral imagery, respectively. The present work extends and combines these previous efforts to demonstrate joint spatial-spectral modeling of multispectral imagery, a multivariate (vector observations) GMRF texture model is employed. Algorithms for parameter estimation and image segmentation are discussed, and a new anomaly detection technique is developed. The model is applied to imagery from the Daedalus sensor. Image segmentation results from test images are discussed and compared to spectral clustering results. The test images are collages, with known texture boundaries constructed from larger data cubes. Anomaly detection results for two Daedalus images are also presented, assessed using receiver operating characteristic (ROC) performance curves, and compared to spectral clustering models. It is demonstrated that even the simplest first-order isotropic texture models provide significant improvement in image segmentation and anomaly detection over pure spectral clustering for the data sets examined. The sensitivity of anomaly detection performance to the choice of parameter estimation method and to the number of texture segments is examined for one example data set
Article
The detection of subpixel targets with unknown spectral signatures and cluttered backgrounds in multispectral imagery is a topic of great interest for remote surveillance applications. Because no knowledge of the target is assumed, the only way to accomplish such a detection is through a search for anomalous pixels. Two approaches to this problem are examined in this paper. The first is to separate the image into a number of statistical clusters by using an extension of the well-known k-means algorithm. Each bin of resultant residual vectors is then decorrelated, and the results are thresholded to provide detection. The second approach requires the formation of a probabilistic background model by using an adaptive Bayesian classification algorithm. This allows the calculation of a probability for each pixel, with respect to the model. These probabilities are then thresholded to provide detection. Both algorithms are shown to provide significant improvement over current filtering techniques for anomaly detection in experiments using multispectral IR imagery with both simulated and actual subpixel targets
A constant false alarm rate (CFAR) detection algorithm (see J.Y. Chen and I.S. Reed, IEEE Trans. Aerosp. Electron. Syst., vol.AES-23, no.1, Jan. 1987) is generalized to a test which is able to detect the presence of known optical signal pattern which has nonnegligible unknown relative intensities in several signal-plus-noise bands or channels. This test and its statistics are analytically evaluated, and the signal-to-noise ratio (SNR) performance improvement is analyzed. Both theoretical and computer simulation results show that the SNR improvement factor of this algorithm using multiple band scenes over the single scene of maximum SNR can be substantial. The SNR gain of this detection algorithm is compared to the previously published one. It illustrates that the generalized SNR of the test using the full data array is always greater than that of using partial data array. The database used to simulate this adaptive CFAR test is obtained from actual image scenes
Long-interval chronochrome target detec-tion
  • A Schaum
  • A Stocker
A. Schaum and A. Stocker, " Long-interval chronochrome target detec-tion, " in Proc. 1997 Int. Symp. Sectral Sensing Research, San Diego, CA, 1998.
SAR image understanding: High-speed target detection and site-model-based exploitation
  • W Phillips
W. Phillips, " SAR image understanding: High-speed target detection and site-model-based exploitation, " Univ. Maryland, College Park, Apr. 1998.
Multi-spectral change detection
  • A E Iverson
  • S S Shen