Photogrammetric Engineering & Remote Sensing

The objective was to test GEographic Object-based Image Analysis (GEOBIA) techniques for delineating neighborhoods of Accra, Ghana using QuickBird multispectral imagery. Two approaches to aggregating census enumeration areas (EAs) based on image-derived measures of vegetation objects were tested: (1) merging adjacent EAs according to vegetation measures and (2) image segmentation. Both approaches exploit readily available functions within commercial GEOBIA software. Image-derived neighborhood maps were compared to a reference map derived by spatial clustering of slum index values (from census data), to provide a relative assessment of potential map utility. A size-constrained iterative segmentation approach to aggregation was more successful than standard image segmentation or feature merge techniques. The segmentation approaches account for size and shape characteristics, enabling more realistic neighborhood boundaries to be delineated. The percentage of vegetation patches within each EA yielded more realistic delineation of potential neighborhoods than mean vegetation patch size per EA.

Remotely sensed imagery has been used to update and improve the spatial resolution of malaria transmission intensity maps in Tanzania, Uganda, and Kenya. Discriminant analysis achieved statistically robust agreements between historical maps of the intensity of malaria transmission and predictions based on multitemporal meteorological satellite sensor data processed using temporal Fourier analysis. The study identified land surface temperature as the best predictor of transmission intensity. Rainfall and moisture availability as inferred by cold cloud duration (ccd) and the normalized difference vegetation index (ndvi), respectively, were identified as secondary predictors of transmission intensity. Information on altitude derived from a digital elevation model significantly improved the predictions. "Malaria-free" areas were predicted with an accuracy of 96 percent while areas where transmission occurs only near water, moderate malaria areas, and intense malaria transmission areas were predicted with accuracies of 90 percent, 72 percent, and 87 percent, respectively. The importance of such maps for rationalizing malaria control is discussed, as is the potential contribution of the next generation of satellite sensors to these mapping efforts.

Complex forest structure and abundant tree species in the moist tropical regions often cause difficulties in classifying vegetation classes with remotely sensed data. This paper explores improvement in vegetation classification accuracies through a comparative study of different image combinations based on the integration of Landsat Thematic Mapper (TM) and SPOT High Resolution Geometric (HRG) instrument data, as well as the combination of spectral signatures and textures. A maximum likelihood classifier was used to classify the different image combinations into thematic maps. This research indicated that data fusion based on HRG multispectral and panchromatic data slightly improved vegetation classification accuracies: a 3.1 to 4.6 percent increase in the kappa coefficient compared with the classification results based on original HRG or TM multispectral images. A combination of HRG spectral signatures and two textural images improved the kappa coefficient by 6.3 percent compared with pure HRG multispectral images. The textural images based on entropy or second-moment texture measures with a window size of 9 pixels × 9 pixels played an important role in improving vegetation classification accuracy. Overall, optical remote-sensing data are still insufficient for accurate vegetation classifications in the Amazon basin.

Traditional change detection approaches have been proven to be difficult in detecting vegetation changes in the moist tropical regions with multitemporal images. This paper explores the integration of Landsat Thematic Mapper (TM) and SPOT High Resolution Geometric (HRG) instrument data for vegetation change detection in the Brazilian Amazon. A principal component analysis was used to integrate TM and HRG panchromatic data. Vegetation change/non-change was detected with the image differencing approach based on the TM and HRG fused image and the corresponding TM image. A rule-based approach was used to classify the TM and HRG multispectral images into thematic maps with three coarse land-cover classes: forest, non-forest vegetation, and non-vegetation lands. A hybrid approach combining image differencing and post-classification comparison was used to detect vegetation change trajectories. This research indicates promising vegetation change techniques, especially for vegetation gain and loss, even if very limited reference data are available.

The classification of image-objects is usually done using parametric statistical measures of central tendency and/or dispersion (e.g., mean or standard deviation). The objectives of this study were to analyze digital number histograms of image objects and evaluate classifications measures exploiting characteristic signatures of such histograms. Two histograms matching classifiers were evaluated and compared to the standard nearest neighbor to mean classifier. An ADS40 airborne multispectral image of San Diego, California was used for assessing the utility of curve matching classifiers in a geographic object-based image analysis (GEOBIA) approach. The classifications were performed with data sets having 0.5 m, 2.5 m, and 5 m spatial resolutions. Results show that histograms are reliable features for characterizing classes. Also, both histogram matching classifiers consistently performed better than the one based on the standard nearest neighbor to mean rule. The highest classification accuracies were produced with images having 2.5 m spatial resolution.

Millimeter-wave Imaging Radiometer (MIR) data (ranging in frequency from 89 GHz to 325 GHz) collected from NASA ER-2 flights over Alaska in April 1995, are used to identify clouds, vegetation type, and snow cover. The procedure used is as follows: 1.) Determine whether a purely MIR-based cloud detection scheme is possible over a snow-covered surface; 2.) Analyze the influence of changing vegetation type on the brightness temperatures; and 3.) Compare completely snow-covered scenes with partially snow-covered and snow-free regions for cloudy and clear sky periods to determine whether varying snow conditions affect the MIR data. Results show that the determination of cloudy pixels over a snow-covered surface is not possible using a simple brightness temperature threshold technique. Furthermore, it is concluded that while no statistical discrimination between specific vegetation classes can be made, statistical significance is obtained when the vegetation is grouped into two classes only, for example vegetated and barren. It is also shown that the state of the snow cover (complete coverage; melting; or patchy) has a distinct affect on these results

We developed a technique for mapping wildland fire in California, USA, using NOAA-14/AVHRR/HRPT data. We integrated a modified active fire detection algorithm and a modified HANDS algorithm. The technique was applied to map wildland fires in California during the 1999 fire season. Reliable results were produced. Most of the wildfires mapped were correct through visual inspection of satellite composite images, validation based on limited ground truth data from the California Departemnt of Forestry and Fire Protection and interpreted burned areas from TM scenes

In recent decades, rapid rates of population growth and urban expansion have led to widespread conversion of natural ecosystems and agricultural lands to urban land cover. The amount and rate of this land conversion affects local and regional ecosystems, climate, biogeochemistry, as well as food production. The main objective of the research described in this paper is to improve understanding of the methodological and validation requirements for mapping urban land cover over large areas from coarse resolution remotely sensed data. A technique called boosting is used to improve supervised classification accuracy and provides a means to integrate MODIS data with the DMSP nighttime lights data set and gridded population data. Results for North America indicate that fusion of these three data types improves urban classification results by resolving confusion between urban and other classes that occurs when any one of the data sets is used by itself. Traditional measures of accuracy assessment as well as new, maplet-based methods demonstrate the effectiveness of the methodology for creating maps of cities at continental scales.

Ground subsidence due to underground mining has posed a constant threat to the safety of surface infrastructure such as motorways, railways, power lines, and telecommunications cables. Traditional monitoring techniques like using levels, total stations and GPS can only measure on a point-by-point basis and hence are costly and time-consuming. Differential interferometric synthetic aperture radar (DINSAR) together with GPS and GIS have been studied as a complementary alternative by exploiting multi-source satellite SAR images over a mining site southwest of Sydney. Digital elevation models (DEMs) derived from ERS-1 and ERS-2 tandem images, photogrammetry, airborne laser scanning, and the Shuttle Radar Topography Mission were assessed based on ground survey data using levelling as well as GPS-RTK. The identified high quality DEM was then used in the DINSAR analysis. Repeat-pass acquisitions by the ERS-1, ERS-2, JERS-1, RADARSAT-1 and ENVISAT satellites were used to monitor mine subsidence in the region with seven active mine collieries. Sub-centimeter accuracy has been demonstrated by comparing DINSAR results against ground survey profiles. The ERS tandem DINSAR results revealed mm-level resolution.

Effects of the 1988 drought on crops in the U.S. Corn Belt were assessed and monitored by the Foreign Crop Condition Assessment Division (FCCAD), U.S. Department of Agriculture. The primary data were vegetation index numbers (VINs), each of which was calculated as an average vegetation index of a geographically referenced cell of AVHRR pixels. Using VINs, the FCCAD was able to detect the existence of drought early in the season, monitor changing conditions, and provide objective assessments of the drought's extent and severity. Field observations confirmed the image analyses, and underlined the importance of the timing of extreme weather events with respect to crop stages for interpreting VINs. The analyses were conducted in an operational environment, providing a unique test of the AVHRR data for large area, near real-time crop monitoring. Because large area, operational remote sensing of crops is quite different from traditional, controlled, small plot research studies, more work is needed to link the two; this would improve crop assessment capabilities.

Successful aerial photography depends on aerial cameras that provide acceptable photographs within the cost restrictions of the job. For topographic mapping where ultimate accuracy is required, only large-format mapping cameras will suffice. For mapping environmental patterns of vegetation, soils, or water pollution, 9-inch cameras often exceed accuracy and cost requirements, and small formats may be an overall better choice. In choosing the best camera for environmental mapping, relative capabilities and costs must be understood. This study compares resolution, photo interpretation potential, metric accuracy, and cost of 9-inch, 70 mm, and 35 mm cameras for obtaining simultaneous color and color-infrared photography for environmental mapping purposes.

This paper confirms the need for detailed 3D models for the simulation of high-resolution SAR images in order to support Persistent Scatterer Interferometry (PSI) focused on single urban objects. Using a building model enhanced by facade grammar, multiple reflections at building facades are analyzed using ray tracing techniques and scatterers are localized in azimuth, range, and elevation. In a case study, salient signatures of a TerraSAR-X image are analyzed based on simulated SAR reflectivity maps. Phase centers of trihedral reflections are mapped onto the building model and the physical correspondence of scatterers to building features is investigated. Surfaces contributing to salient scatterers are identified at the building model. Eventually, the use of SAR simulation to support PSI is shown from two aspects: (a) for providing a-priori information about building layover, and (b) for extending knowledge about the nature of dominant scatterers.

Geophysical and Environmental Research Imaging Spectrometer (GERIS) 63-channel scanner data covering the spectral region 0.4 to 2.5 microns were analyzed for the Cuprite mining district, Esmeralda and Nye Counties, Nevada. Individual and spatially averaged spectra extracted from the GERIS data were used to identify the minerals alunite, kaolinite, buddingtonite, and hematite by their spectral characteristics. The images were classified in the spectral domain to produce color-coded image maps of mineral distribution that clearly show the zoned nature of the hydrothermal system. Comparison of the thematic mineral maps with existing geologic and alteration maps demonstrates the utility of imaging spectrometers for producing detailed maps for mineral exploration.

Two digital elevation models are compared for the Echo Mountain SE quadrangle in the Cascade Mountains of Oregon. Comparisons were made between 7.5-minute (1:24,000-scale) and 1-degree (1:250,000-scale) images using the variables of elevation, slope aspect, and slope gradient. Both visual and statistical differences are presented.

A south Texas rangeland area was used as a study site to test the use of microdensitometry on 70-mm color-infrared and black-and-white photographs (scale 1:19,000) for distinguishing among 11 range sites (two brushland, seven grassland, two barren land) during the winter (February), spring (May), and summer (August) of 1976. Color-infrared photographs were also taken at a scale of 1:42,000 for the summer date. Film optical density readings were made on one color-infrared film with white light only. The best separations among density readings for all range sites were obtained using white light exposed on color-infrared film in the summer when vegetation was at peak foliage development. Results from this study indicate that 70-mm aerial color-infrared photography at a scale of 1:19,000 or 1:42,000 has good potential for identifying range sites in large and inaccessible areas, and could be a useful tool for range management.

Many of the world's rangelands are affected by salinity. The detection of these areas is important to range and resource managers who are concerned with productivity, condition, and animal carrying capacity. The reported study was conducted along a north-to-south flight line 24 km long and 1.6 wide in Starr County, TX. Everitt et al. (1977) described seven different native range sites (four nonsaline and three saline) along this flight line. The study showed that photointerpretation by microdensitometry could be used to identify saline range sites quantitatively on CIR (0.50 to 0.90 micrometers) aerial film (scales 1:19,000, 1:42,000, and 1:80,000) exposed in May 1976, June 1976, and June 1979. Microdensitometer readings made on CIR film using white or blue light generally gave the best separation between saline and nonsaline range sites. The differences in microdensitometry readings among saline and nonsaline range sites were caused by less plant cover on the saline sites.

Airborne Visible/Infrared Imaging Spectrometer data covering the wavelength range between 2000 and 2400 nm are examined for their ability to display the diagnostic mineral absorption features of certain alteration minerals, employing various data processing techniques. The techniques may be separated into two broad categories: scene based techniques that use parameters derived from the data themselves, and correction techniques utilizing external information such as solar/atmospheric models. Results indicate that the data corrected utilizing the LOWTRAN 7 atmospheric transfer code constrained with local weather station data are the most effective at showing the diagnostic absorption features of the regions of known mineralogy and introduce the least number of artifacts into the data.

By means of two simple sampling plans suggested in the accuracy-assessment literature, it is shown how one can use knowledge of map-category relative sizes to improve estimates of various probabilities. The fact that maximum likelihood estimates of cell probabilities for the simple random sampling and map category-stratified sampling were identical has permitted a unified treatment of the contingency-table analysis. A rigorous analysis of the effect of sampling independently within map categories is made possible by results for the stratified case. It is noted that such matters as optimal sample size selection for the achievement of a desired level of precision in various estimators are irrelevant, since the estimators derived are valid irrespective of how sample sizes are chosen.

A Landsat Image Data Quality Analysis (LIDQA) Program is conducted by NASA. One part of this program forms studies which are being performed with the objective to evaluate the geometric fidelity of Landsat-4 and Landsat-5 Thematic Mapper (TM) data in computer tape (CCT-pt) formats. It is pointed out that the Landsat-4 and Landsat-5 systems provide image data of significantly better geometric fidelity than were obtained from the earlier Landsat missions. Attention is given to the factors which influence the geometric fidelity of the Landsat TM data, the study areas and data sets, the rectification procedures, the rectification of Landsat-4 TM data and comparisons of the Scrounge and the TM Image Processing System (TIPS), the rectification of system and scene corrected Landsat-5 data processed on TIPS, and the cartographic potential of TM data.

Adaptive box-filtering algorithms to remove random bit errors and to smooth noisy data have been developed. For both procedures, the standard deviation of those pixels within a local box surrounding each pixel is used. A series of two or three filters with decreasing box sizes can be run to clean up extremely noisy images and to remove bit errors near sharp edges. The second filter, for noise smoothing, is similar to the 'sigma filter' of Lee (1983). The technique effectively reduces speckle in radar images without eliminating fine details.

An earlier study under somewhat clinical laboratory conditions has suggested the possibility of using smaller scales of forest photography without serious information loss. The present paper subjects this idea to a rigorous field test by a number of experienced user-cooperators. Various combinations of summer black-and-white infrared and color infrared aerial photography at scales of 1:15,840, 1:24,000, 1:31,680, and 1:80,000 were taken over forested portions of Minnesota. Major conclusions are that 1:15,840 is the preferred working photo scale, and that instead of 1:15,840 a scale of 1:20,000 is considered an acceptable substitute.

The Universal Soil Loss Equation (USLE) is a widely accepted tool for erosion prediction and conservation planning. Solving this equation yields the long-term average annual soil loss that can be expected from rill and inter-rill erosion. In this study, manual interpretation of color and color infrared 70 mm photography at the scale of 1:60,000 is used to determine the cropping management factor in the USLE. Accurate information was collected about plowing practices and crop residue cover (unharvested vegetation) for the winter season on agricultural land in Pheasant Branch Creek watershed in Dane County, Wisconsin.

Estimates of per cent ground cover made by ground observers were compared with independent estimates made on the basis of low-altitude (640-1219 m) aerial photographs of the same fields. Standard statistical simple correlation and linear regression analyses revealed a high correlation between the two estimation methods. In crops such as grain, sorghum, corn, and forage sorghum, in which the broadest part of the leaf canopy is near the top of the plant, there was a tendency to overestimate the per cent ground cover from aerial photographs.

Determination of aerial camera shutter and aperture settings to produce consistently high-quality aerial photographs is a task complicated by numerous variables. Presented in this article are brief discussions of each variable and specific data which may be used for the systematic control of each. The variables discussed include sunlight, aircraft altitude, subject and season, film speed, and optical system. Data which may be used as a base reference are included, and encompass two sets of sensitometric specifications for two film-chemistry processes along with camera-aircraft parameters, which have been established and used to produce good exposures. Information contained here may be used to design and implement an exposure-determination system for aerial photography.

The objective of this project is to determine the feasibility of an operational inventory utilizing predawn aerial thermal infrared data for censusing Canada geese. -from Authors