Showing papers in "Photogrammetric Engineering and Remote Sensing in 2001"

Completion of the 1990s national land cover data set for the conterminous united states from landsat thematic mapper data and ancillary data sources

Abstract: In late 2000, the U.S. Geological Survey (USGS) EROS Data Center completed the circa 1992 National Land Cover Data set (NLCD). The NLCD, derived from early 1990s Landsat Thematic Mapper (TM) imagery and other sources of digital data, represents an intermediate-scale national land cover data set. The resolution of this data set lends itself to many regional to national scale investigations, including analyses of water quality, ecosystem health, wildlife habitat, land cover assessment, land use planning, urban studies, urban sprawl, and other land management issues. The article provides information about the materials and methods used to collect and process the data, the accuracy of the data, and uses for the NLCD.

A Comprehensive Study of the Rational Function Model for Photogrammetric Processing

Abstract: The rational function model (RFM) has gained considerable interest recently mainly due to the fact that Space Imaging Inc. (Thornton, Colorado) has adopted the RFM' as an alternative sensor model for image exploitation. The RFM has also been implemented in some digital photogrammetric systems to replace the physical sensor mode for photogrammetric processing. However, there have been few publications addressing the theoretical properties and practical aspects of the RFM until recently. In this paper a comprehensive study of the RFM is reported upon. Technical issues such as the solutions, feasibility, accuracy, numerical stability, and requirements for control information are addressed. Both the direct and iterative least-squares solutions to the RFM are derived, and the solutions under terrain-dependent and terrain-independent computation scenarios are discussed. Finally, evaluations of the numerous tests with different data sets are analyzed. The objective of this study is to provide readers with a comprehensive understanding of the issues pertaining to applications of the RFM. Background

Measurement and monitoring of urban sprawl in a rapidly growing region using entropy

Abstract: Rapid urban development and dramatic change of landscape have been recently witnessed in some developing countries as a result of rapid economic development. The measurement and monitoring of land-use changes in these areas are crucial to government officials and planners who urgently need updated information for planning and management purposes. This paper examines the use of entropy in the measurement and monitoring of urban sprawl by the integration of remote sensing and GI~. The advantages of the entropy method are its simplicity and easy integration with GIs. The measurement of entropy is devised based on two locational factors-distances from town centers and roads-to capture and reveal spatial patterns of urban sprawl. The entropy space can be conveniently used to differentiate various kinds of urban growth patterns. The application of the method in the Pearl River Delta, one of the fastest growing regions in China, has demonstmted that it is very useful and effective for the monitoring of urban sprawl. It provides a useful tool for the quantitative measurement that is much needed for rapidly growing regions in identifying the spatial variations and temporal changes of urban sprawl patterns. lntroductlon

Rule-Based Classification Systems Using Classification and Regression Tree (CART) Analysis

Abstract: Incorporating ancillary data into image classification can increase classification accuracy and precision. Rule-based classification systems using expert systems or machine learning are a particularly useful means of incorporating ancillary data, but have been difficult to implement. We developed a means for creating a rule-based classification using classification and regression tree analysis (CART), a commonly available statistical method. The CART classification does not require expert knowledge, automatically selects useful spectral and ancillary data from data supplied by the analyst, and can be used with continuous and categorical ancillary data. We demonstrated the use of the CART classification at three increasingly detailed classification levels for a portion of the Greater Yellowstone Ecosystem. Overall accuracies ranged from 96 percent at level 1, to 79 percent at level 2, and 65 percent at level 3.

Harmonic analysis of time-series AVHRR NDVI data

Abstract: Harmonic analysis of a one-year time series (26 periods) of NOAA AVHRR NDVI biweekly composite data was used to characterize seasonal changes for natural and agricultural land uselland cover in Finney County in southwest Kansas. Different crops (corn, soybeans, alfalfa) exhibit distinctive seasonal patterns of NDVI variation that have strong periodic characteristics. Harmonic analysis, also termed spectral analysis or Fourier analysis, decomposes a time-dependent periodic phenomenon into a series of sinusoidal functions, each defined by unique amplitude and phase values. The proportion of variance in the original time-series data set accounted for by each term of the harmonic analysis can also be calculated. Amplitude and phase angle images were produced from analysis of the time-series NDVI data and correlated with information on crop type and extent for the region to develop a methodology for crop-type identification. Crop types occurring in southwest Kansas, including corn, winter wheat, alfalfa, pasture, and native prairie grasslands, were characterized and identified using this technique and biweekly AVHRR composite data for 1992. For crops with a simple phenology, such as corn, the majority of the variance was captured by the first and additive terms of the harmonic analysis, while winter wheat exhibited a bimodal NDvI periodicity with the majority of the variance accounted for by the second harmonic term.

Comparison of change-detection techniques for monitoring tropical forest clearing and vegetation regrowth in a time series

Abstract: The once remote and inaccessible forests of Guatemala's Maya Biosphere Reserve (MBR) have recently experienced high rates of deforestation corresponding to human migration and expansion of the agricultural frontier. Given the importance of land-cover and land-use change data in conservation planning, accurate and efficient techniques to detect forest change from multi-temporal satellite imagery were desired for implementation by local conservation organizations. Three dates of Landsat Thematic Mapper imagery, each acquired two years apart, were radiometrically normalized and preprocessed to remove clouds, water, and wetlands, prior to employing the change-detection algorithm. Three change-detection methods were evaluated: normalized difference vegetation index (NDVI) image differencing, principal component analysis, and RGB-NDVI change detection. A technique to generate reference points by visual interpretation of color composite Landsat images, for Kappa-optimizing thresholding and accuracy assessment, was employed. The highest overall accuracy was achieved with the RGB-NDVI method (85 percent). This method was also preferred for its simplicity in design and ease in interpretation, which were important considerations for transferring remote sensing technology to local and international non-governmental organizations.

Spatial Prediction of Fire Ignition Probabilities: Comparing Logistic Regression and Neural Networks

Abstract: The objective of this work was to develop and validate models to predict spatially distributed probabilities of ignition of wildland fires in central Portugal The models were constructed by exploring relationships between ignition location/cause and values of geographical and environmental variables using logistic regression and neural networks The conclusions are that (1) the spatial patterns of fire ignition identified can be used for prediction, (2) the spatial patterns are different for the different causes, (3) the logistic models and the neural networks both reveal acceptable levels of predictive ability but the neural networks present better accuracy and robustness, (4) the maps produced by the two methods are similar, and (5) the information contained in the spatial position of ignition events can be used to gain predictive capability over an important phenomenon that is difficult to characterize and, for that reason, has not been included in most of the currently used fire danger estimation systems

Land Subsidence Monitoring with Differential SAR Interferometry

Abstract: This article investigates whether differential synthetic aperture radar (SAR) interferometry can be used to monitor land subsidence. The principle of the technique and the approach used on a specific case are presented. The high potential of differential SAR interferometry to monitor a wide range of deformation velocities ranging from fast (m/year) to slow (mm/year) was demonstrated by the generation of subsidence maps for sites in Germany, Mexico, and Italy. The SAR interferometric displacement maps are validated with available leveling data. Differential SAR interferometry is suitable for operational monitoring of land subsidence due to the accuracy of the maps produced, the extensive SAR data archive over the past 10 years, the expected continued availability of SAR data, and the maturity of the required processing techniques. A strategy for the integration of leveling, global positioning systems, and SAR data is proposed in order to achieve an accurate, rational and cost-effective monitoring.

Sensitivity of landscape pattern metrics to map spatial extent

Abstract: Computation of landscape pattern metrics from spectrally classified digital images is becoming increasingly common, because the characterization of landscape spatial structure provides valuable information for many applications. However, the spatial extent (window size) from which pattern metrics are estimated has been shown to influence and produce biases in the results of these spatial analyses. In this study, the sensitivity of eight commonly used landscape configuration metrics to changes in map spatial extent is analyzed using simulated thematic landscape patterns generated by the modified random clusters method. This approach makes it possible to control and isolate the different factors that in-fluence the behavior of spatial pattern metrics, as well as taking into account a wide range of landscape configuration possi-bilities. Edge Density is found to be the most robust metric and is recommended as a fragmentation index where the effect of spatial extent is concerned. The metrics that attempt to quantify the irregularity and complexity of the shapes in the pattern (Mean Shape Index, Area Weighted Mean Shape Index, and Perimeter Area Fractal Dimension) are by far the most sensitive. In particular, it is suggested that the Mean Shape Index should be avoided in further landscape studies. For the eight analyzed pattern metrics, quantitative guidelines are provided to estimate the systematic biases that may be introduced by the use of a given extent, so that the metric values derived from data of different spatial extents can be properly compared.

Modeling the population of China using DMSP operational linescan system nighttime data

Abstract: Radiance-calibrated DMSP-OLS nighttime lights data of China acquired between March 1996 and January-February 1997 were evaluated for their potential as a source of population data at the provincial, county, and city levels The light clusters were classified into six categories of light intensity, and their areal extents were extracted from the image Mean pixel values of light clusters corresponding to the settlements were also extracted A light volume measure was developed to gauge the three-dimensional capacity of a settlement A density of light cluster measure known as percent light area was also calculated for each spatial unit Allometric growth models and linear regression models were developed to estimate the Chinese population and population densities at the three spatial levels using light area, light volume, pixel mean, and percent light area as independent variables It was found that the DMSP-OLS nighttime data produced reasonably accurate estimates of non-agricultural (urban) population at both the county and city levels using the allometric growth model and the light area or light volume as input Non-agricultural population density was best estimated using percent light area in a linear regression model at the county level The total sums of the estimates for non-agricultural population and even population overall closely approximated the true values given by the Chinese statistics at all three spatial levels It is concluded that the 1-km resolution radiance-calibrated DMSP-OLS nighttime lights image has the potential to provide population estimates of a country and shed light on its urban population from space

Mapping Seasonal Flooding in Forested Wetlands Using Multi-Temporal Radarsat SAR

Abstract: Eleven Radarsat scenes imaged between 22 September 1996 and 28 February 1998 were analyzed to delineate flood inundation in the forests of the Roanoke River floodplain, North Carolina. Threshold σ° values distinguishing flooded from nonflooded forests were identified using classification trees. Data from 13 U.S. Geological Survey (USGS) wells located throughout the floodplain were used to validate the flood mapping with an overall accuracy of 93.5 percent. Images from both leaf-on and leaf-off periods were acceptable for detecting flooding, although the leaf-off scenes were classified with higher accuracy than were the leaf-on scenes (98.1 percent versus 89.1 percent). In addition, threshold σ° values were lower for leaf-on scenes. The results also indicate that Radarsat data can be used to detect minimal flood levels-sites with water stages between 10 cm below and 10 cm above the forest floor were classified with 90.6 percent accuracy. Radarsat data are effective and appropriate for flood inundation mapping in forests, regardless of season or water level.

Detecting the nature of change in an urban environment : A comparison of machine learning algorithms

Abstract: The performance of different machine learning algorithms for detecting the nature of change was compared. To choose the best algorithms for a specific task, users are advised to consider not only classification accuracy, but also comprehensibility, compactness, and robustness in training and classification. The purpose of this paper is to provide information to users so that they can choose the best algorithms for specific tasks. To alleviate the problem of obtaining enough training data, simulated training data were generated from single-date images. A one-pass classification with four machine learning algorithms, namely, Multi-Layer Perceptrons (MLP), Learning Vector Quantization (LVQ), Decision Tree Classifiers (DTC), and the Maximum-Likelihood Classifier (MLC), were tested. Recognition rates, ease of use, and degree of automation of the four algorithms were assessed. The results showed that the incorporation of cross-combined simulated training data enhanced the detection of nature of change. Compared to conventional post-classification comparison methods, LVQ and DTC did better in terms of overall accuracy. In terms of average accuracy of the change classes, LVQ was the best performer. DTC was the easiest to use and the most robust in training. MLP procedures were the most difficult to replicate.

Remote sensing of clear-water, shallow, gravel-bed rivers using digital photogrammetry

Abstract: The digital elevation model (DEM] quality that can be obtained from a digital photogrammetric survey of a reach of the clear water, shallow, gravel-bed North Ashburton River, New Zealand is assessed. An automated correction procedure is used to deal with point errors associated with submerged topography, based on a correction for refraction at an air-water interface. The effects of collection parameter variation upon DEM quality are also considered. The accuracy and precision of DEMs of submerged topography are evaluated using an independent data set. Results show that digital photogrammetry, if used in conjunction with image analysis techniques, can successfully be used to extract high-resolution DEMs of gravel riverbeds, but that the quality of submerged topographic representation is heavily dependent upon the water depth at the time of image acquisition. It is suggested that differences between the digital photogrammetric surface and the "actual" riverbed surface (as determined by terrestrial ground survey] will, in part, reflect the problem of defining what is the true elevation of a gravel-covered surface. A digital photogrammetric survey will generally see the tops of gravel cobbles, while a hand-held survey staff will tend to record the elevation between stones. The nomenclature of errors is also discussed, and it is concluded that the measure of surface quality adopted should be consistent with the application for which the DEM is to be used.

Statistical Rigor and Practical Utility in Thematic Map Accuracy Assessment

Abstract: Quantifying map accuracy provides important descriptive information to assess the utility of a map for a specified application. This article focuses on site-specific, thematic accuracy in which accuracy is defined by comparing the map attribute and the actual attribute for a sample of areal units. Although statistical rigor and practical utility have been advocated as desirable features of map accuracy assessment protocols, specific criteria defining these features have not been elucidated. Two criteria are proposed for statistical rigor: probability sampling and consistent estimation. Practical utility is synonymous with cost, and because cost is directly related to quality, decisions regarding practical utility may be evaluated in terms of their effect on quality. Four criteria are proposed to define quality: the precision of the accuracy estimates, the population to which sampling inference is justified, the assumptions needed to justify inference, and the accuracy of the reference data. The first step in planning a statistically rigorous, practical accuracy assessment is to construct an efficient, probability-sampling-based strategy permitting inference to the full map population. Modifications of this strategy to enhance practical utility (i.e. reduce cost of the assessment) should be evaluated using the criteria defined for quality and statistical rigor.

Spectral Separability among Six Southern Tree Species

Abstract: Spectroradiometer data (350 to 2500 nm) were acquired in late summer 1999 over various forest sites in Appomattox Buckingham State Forest, Virginia, to assess the spectral differentiability among six major forestry tree species: loblolly pine (Pinus taeda), Virginia pine (Pinus virginiana), shortleaf pine (Pinus echinata), scarlet oak (Quercus coccinea), white oak (Quercus alba), and yellow poplar (Liriodendron tulipifera). Data were smoothed and curve shape was determined using first- and second-difference operators. Stepwise discriminant analysis was used to decrease the number of independent variables, after which a canonical discriminant analysis and a normal discriminant analysis were performed. Cross-validation accuracies varied from 99 percent to 100 percent (hardwood versus pine groups), 62 percent to 84 percent (within pine group), and 78 percent to 93 percent (within hardwood group). The second difference of a nine-point weighted average proved most accurate overall, with cross-validation accuracies of 84 percent (within pine separability), 93 percent (within hardwood separability), and 100 percent (between group separability). Landsat simulation data had lower accuracies, varying from 93 percent to 96 percent (hardwood versus pine groups), 45 percent to 60 percent (within pine group), and 54 percent to 70 percent (within hardwood group). The relatively low accuracies for Landsat simulation data indicate the need for high spectral resolution data for within group separability. The variables significant in defining spectral separability within and between groups were largely located in the visible (350- to 700-nm) and shortwave infrared 1 (700- to 1850-nm) regions of the spectrum, with markedly less representation in the shortwave infrared I1 (1 700- to 2500nm) region. Some wavelengths related to nitrogen concentration and 0-H bond regions were evident, but not dominant.

Using spatial Co-occurrence texture to increase forest structure and species composition classification accuracy

Abstract: should be augmented with texture measures (St-Onge and The analysis of forest structure and species composition with Cavayas, 1995). Texture can be used in the classification or, high spatial resolution (5 1 m) multispeckal digital imagery alternatively, the process of per-pixel classification can be supis described in an experiment using spatial co-occurrence plemented or replaced with another approach, such as texture texture and maximum-likelihood classification. The segmentation (Lobe, 1997). Much work remains to be done on objective was to determine if higher forest species composition how and where texture analysis can be effective in fores~ classification accuracies would result in comparison to the remote sensing applications (Lark, 1996; ~ulder et al., 1996; use of spectral response patterns alone. Increased accuracy Bmiquel-Pinel and Gastellu-EtchegO~* lgg8). was obtained when using texture at all levels of a classifi'cation A large degree of variability can exist in the development hierarchy. At the stand level, accuracies were on the order of of a classification signature using high spatial resolution image 75 percent in agreement with field surveys, an improvement data. In Figure 1, approximately 1-m pixels are shown of a forof 21 percent over the accuracy obtained using spectral data est stand adjacent to a logging road Tintersection. Large Stanalone; in stands grouped according to species dominancelco- dard deviations, relative to the mean spectral response, typidominance, the accuracy improved still firther to 80 percent. cally result in forested scenes in most spectral bands. The diffiThe overall classification accuracy in a highly generalized culties in using spectral signatures comprised of the mean and lifeform classification was 100 percent. This represented a 33 standard deviation are obvious, giving rise to the notion that percent increase in accuracy over that which could be some measure of spatial variability would be useful in signaobtained, in a classic spectral "signature" classification ture generation (Figure 2). Texture analysis attempts to meaapproach, using spectral response patterns alone. sure this scene variance for use in the classification process. Traditionally, texture has been defined as the spatial variation

Laser altimetry : From science to commercial lidar mapping

Abstract: Laser altimetry, which is also referred to in the commercial sector as lidar mapping, can rapidly generate dense, accurate, digital models of the topography and vertical structure of a target surface. It is becoming a popular operational tool in the fields of remote sensing, photogrammetry, surveying and mapping because it offers unique technical capabilities, lower field-operation costs, and reduced post-processing time and effort compared with traditional survey methods. However, although the adoption of laser altimetry by the commercial sector has grown rapidly during the past five years, the transition from science to an operational remote sensing tool has not been easy. Lidar mapping is still regarded with skepticism by many mainstream end users because of early experiences of poor data, failure to meet specifications, repeatedly missed schedules and hidden costs. This article provides an overview of commercial lidar mapping development, specifications and pricing. Five key areas that will have significant impact in the next five years on laser altimetry from both a commercial and academic point-of-view are highlighted: higher density data, commercial off-the-shelf software, waveform capture, laser bathymetry, and platform options. Barriers to greater adoption of lidar mapping include the high capital costs to acquire or develop a sensor, a lack of published guidelines or standards, the perceived threat to existing value chains posed by the new technology, conflicting opinions about the achievable performance of the sensors, and a poor fit between the nature of lidar data ad the existing mapping standards developed for older technologies.

Differential snakes for change detection in road segments

Abstract: Suetens et al. (1992), Gruen et al. (1995b), Gruen et al. (1997), The automation of object extraction from digital imagery has and Lukes (1998). been a key research issue in digital photogrammetry and com- The majority of current object extraction methodologies puter vision. In the spatiotemporal context of modern GIS, with are semi-automatic, whereby a human operator provides manuconstantly changing environments and periodic database re- ally some approximations (e.g., by selecting points on a monivisions, change detection is becoming increasingly important. tor display) and an automated algorithm uses these points as In this paper, we present a novel approach for the integration input to extract a complete object outline. Considering roads of object extraction and image-based geospatial change de- and similar linear features, these approximations may be in the tection. We extend the model of deformable contour models form of an initial point and an approximate direction. This (snakes) to function in a differential mode, and introduce a information is used as input to automated algorithms that pronew framework to differentiate change detection from the ceed by profile matching (Vosselman and de Knecht, 1995), recording of numerous slightly different versions of objects that edge analysis (Nevatia and Babu, 1980), or even combinations may remain unchanged. We assume the existence of prior of both (McKeown and Denlinger, 1988). Alternatively, the information for an object (an older record of its shape available human operator may provide a set of points that roughly in a GIS) with accompanying accuracy estimates. This infor- approximate the road from start to end, e.g., a polygonic mation becomes input for our “differential snakes” approach. approximation of a long road segment. This information is In a departure from standard techniques, the objective of our used by automated methods like dynamic programming and object extraction is not to extract yet another version of an deformable contour models, i.e., snakes (Gruen and Li, 1997; object from the new image, but instead to update the pre- Li, 1997). Full automation is pursued by automating the selecexisting GIS information (shape and corresponding accuracy). tion of the above-mentioned necessary initial information By incorporating accuracy information in our technique, we (e.g., node locations, road orientation). Examples of substantial identify local or global changes to this prior information, and efforts towards full automation may be found in Baumgartner et

Monitoring the magnitude of land-cover change around the southern limits of the Sahara

Abstract: Studies of land-cover change using satellite remote sensing are often constrained to depict land-cover conversions only, with the equally important modifications undetected or misrepresented, resulting in significant error. Desert fluctuations within the Sahel were examined using an approach that indicated the magnitude of land-cover changes. This showed that the conventional post-classification comparison method of change detection appeared to underestimate the area of land-cover change and, where a change was detected, typically overestimate its magnitude. At the regional scale, the land-cover changes detected were strongly related to rainfall variability. This relationship did not, however, explain changes at a finer spatial scale and indicated that dryland degradation, and its causes, may remain far from understood.

Remote sensing and cast shadows in mountainous terrain

Abstract: In mountainous environments with high relief, topographymay cause cast shadows due to the blocking of direct solar mdiation. Optical-injinred remote sensing images of these landscapes display reduced values of reflectance for shadowed areas compared to non-shadowed areas with similar surjace cover characteristics. Different approaches to dealing with cast shadows are possible, although a common step in various active approaches is first to delineate the shadows using an automated algorithm and a digital elevation model. This article demonstmtes a common confusion caused by cast shadows and describes a quantitative spatial evaluation of a cast shadow delineation algorithm in comparison to human interpretation of a Landsat TM image. It is shown that 86 percent of cast-shadow pixels were correctly marked by the algorithm. The causes of differences between the algorithm and human interpretation are discussed, and alternatives are considered for dealing with cast shadows in classification studies using optical-infrared images of mountainous terrain.

Effects of Differential Single- and Dual-Frequency GPS and GLONASS Observations on Point Accuracy under Forest Canopies

Abstract: A 20-channel, dual-frequency receiver observing dual-fiequency pseudorange and carrier phase of both GPS and GLONASS was used to determine the positional accuracy of 29 points under tree canopies. The mean positional accuracy based on differential postprocessing of GPS+GLONASS singlefrequency observations ranged from 0.16 m to 1 .I 6 m for 2.5 min to 20 min of observation at points with basal area ranging from <20 m2/ha to 230 m2/ha. The mean positional accuracy of differential postprocessing of dual-frequency GPS+GLONASS observations ranged from 0.08 m to 1.35 m. Using the dualfrequency carrier phase as main observable and fixing the initial integer phase ambiguities, i.e., a fixed solution, gave the best accuracy. However, searching for fixed solutions increased the risk of large individual positional errors due to "false" fixed solutions. The accuracy increased with decreasing density of forest, increasing length of observation period, and decreasing a priori standard error as reported by the postprocessing software.

Texture-Integrated Classification of Urban Treed Areas in High-Resolution Color-Infrared Imagery

Abstract: lkaditional multispectral classification methods have not provided satisfying results for treed area extraction from highresolution digital imagery because trees are characterized not only by their spectral but also by their textural properties. Treed areas in urban regions are especially dificult to extract due to their small area. Many other urban objects, such as lawn and playgrounds, cause confusion because they display similar, even identical, spectral properties. In this study a texture integrated classification method is proposed. To effectively extract tree textural features and eliminate noise, an algorithm of conditional variance detection is developed, which consists of a directional variance detection and a local variance detection. This algorithm detects tree features with higher accuracy than common texture algorithms. By integrating the new algorithm with traditional multispectral classification, treed areas in urban regions can be extracted with sufficiently high accuracy. Application of the new approach in different urban areas indicates that the average accuracy of treed area extraction was increased from 67 percent, using a multispectral classification, to 96 percent, using the texture integrated classification.

Analysis of GAC NDVI Data for Cropland Identification and Yield Forecasting in Mediterranean African Countries

Abstract: The utilization of NOAA-AVHRR NDVI data for crop yield forecasting is of particular importance in semiarid regions where there are strong inter-year yield fluctuations due to meteorological vagaries. The present work deals with the use of monthly GAC NDVI data for the early estimation of cereal crop yield in Mediterranean African countries. A preliminary analysis showed that relatively high correlations were present between crop yield and mean NDvI values of specific months computed at national levels. The stratification of the countries according to the ~SGS global land-cover map brought only marginal correlation increases. Greater improvements were instead reached by a statistical method which allows the estimation of the per-pixel fractions of agricultural and nonagricultural vegetation. When compared to available independent maps, the areas identified in this way were confirmed to be mainly covered by crop and forest land, respectively The methodology for cropland identification and yield forecasting was finally evaluated for operational applications.

Model-based conifer-crown surface reconstruction from high-resolution aerial images

Abstract: Knowledge of tree-crown parameters such as height, shape, and crown closure is desimble in forest and ecological studies, but those pammeters are difficult to measure on the ground. The stereoscopic capability of high-resolution aerial images provides a method for crown-surface reconstruction. However. 'existing digital photogrammet~packages, designed to map terrain surfaces. cannot accuratelv extmct tree-crown surfaces. particularb for-conifer-crowns dth steep vertical profilis. ' In this paper, we integrate crown features derived from images with stereo matching, and develop a model-based approach for reconstructing conifer-crown sqfaces. The model is based on the fact that most conifer crowns are a form of solid geometry. We model a conifer crown as a genemlized hemi-ellipsoid, establish the optimal tree model using a geometric equation, and apply the optimal tree model to guide a conventional pyramidal image matching in crown-surface reconstruction. The effectiveness of the approach is illustmted using an example of a redwood tree on 1:2,400-scale aerial photographs.

Landsat TM-Based Forest Area Estimation Using Iterative Guided Spectral Class Rejection

Abstract: In cooperation with the USDA Forest Service Southern Research Station, an algorithm has been developed to replace the current aerial-photography-derived FIA Phase I estimates of forest/ non-forest area with a Landsat TM-based forest area estimation. Corrected area estimates were obtained using a new hybrid classifier called Iterative Guided Spectral Class Rejection (IGSCR) for portions of three physiographic regions of Virginia. Corrected area estimates were also derived using the Landsat TM-based Multi-Resolution Land Characteristic Inter-agency Consortium (MRLC) cover maps. Both satellite-based corrected area estimates were tested against the traditional photo-based estimates. Forest area estimates were not significantly different (at the 95 percent level) between the traditional FIA, IGSCR, and MRLC methods, although the precision of the satellite-based estimates was lower. Map accuracies were not significantly different (at the 95 percent level) between the IGSCR method and the MRLC method. Overall accuracies ranged from 80 percent to 89 percent using FIA definitions of forest and non-forest land use. Given standardization of the image rectification process and training data properties, the IGSCR methodology is fast, objective, and repeatable across users, regions, and time, and it outperforms the MRLC for FIA applications.

Mapping Gold-Bearing Massive Sulfide Deposits in the Neoproterozoic Allaqi Suture, Southeast Egypt with Landsat TM and SIR-C/X SAR Images

Abstract: Meta-volcanic sequences in the Neoproterozoic AmbianNubian Shield host auriferous massive sulfide deposits with su$ace expression in the form of clay and iron alteration zones. These are large enough (few hundreds of meters across) and have distinctive characteristic reflectance spectra to be mapped with the 30-m spatial resolution Landsat Thematic Mapper [TM) images. Landsat TM together with Shuttle Imaging Radar [SIR)-CIX-Synthetic Aperture Radar (SAR) images are used for finding and mapping the alteration zones in the western part of the east-trending Allaqi suture in the Southeastern Desert of Egypt. The 517-511-314*5/4 band-ratio Landsat TM image shows the Umm Garaiat alteration zone as sugary white; this appears red in the 517-415-311 band-ratio Landsat TM image. Geological and geochemical data indicate that the Umm Garaiat alteration zone is the surface expression of a massive sulfide deposit that contains up to 12 glt gold. Density slicing using 517 and 311 Landsat TM band-ratios effectively maps clay and iron alteration. The 517 density slicing Landsat TM image suggests that the Umm Garaiat alteration zone is dominated by clay minemls. The 311 density slicing Landsat TM image shows little evidence for FeO minerals associated with the Umm Garaiat alteration zone. Analysis of Landsat TM images with supervised classification techniques using the Umm Garaiat alteration zone as the training site helped identify previously unknown alteration zones at Wadi Marahiq. Interpretation of Chh-Lhh-Lhv SIR-CIX-SAR images helped in understanding the lithological and structural controls on massive sulfide deposits in the study area. This demonstrates the utility of orbital remote sensing for finding ore deposits in arid regions.

An accuracy assessment of various gis-based viewshed delineation techniques

Abstract: The widespread availability of geographic information systems allows spatial analysts to predict what can be seen from various locations, thereby allowing them to find advantageous sites for facilities requiring wide fields of view and for facilities whose visual impacts are to be minimized. This study uses standard surveying techniques to delineate actual viewsheds in the field, and then compares these actual viewsheds to predicted viewsheds generated by computer analysis using various commonly used elevation data sets, data models and visibility criteria. Results showed that some data sets produced predicted viewsheds that mimicked surveyed viewsheds much more accurately than other data sets, and that some combinations of data model and visibility criteria produced much greater matching accuracies than other combinations. However, even under the best of conditions, the average level of agreement between predicted viewsheds and field surveyed viewsheds was only slightly higher than 50 percent. Plausible causes of these predicted viewshed inaccuracies include likely errors in the elevation data used in creating the predicted viewsheds and insufficient spatial resolution of the elevation data.

Introducing New Indices for Accuracy Evaluation of Classified Images Representing Semi-Natural Woodland Environments.

Abstract: A range of accuracy indices for determining the optimal outputs from the classification of multispectral remotely sensed data is evaluated. Airborne Thematic Mapper imagery of semi-natural woodland was used in conjunction with an in situ data set. Indices of classification accuracy were unable to distinguish substantial differences in classified images because they are based only on errors of omission, accounting for only a proportion of the errors in classification. The Classification Success Index (CSI) is introduced here to estimate the overall effectiveness of classification, considering all output classes and using both errors of omission and commission from the error matrix. The Individual Classification Success Index (ICSI) is introduced which accounts for the classification success of a specific class. Finally, the Group Classification Success Index (GCSI) measures classification success for the most important classes in the area of interest. These new indices were found to offer considerable improvement over existing approaches.

Combining location and classification error sources for estimating multi-temporal database accuracy

Abstract: Detection and quantification of temporal change in spatial objects is the subject of a growing number of studies. Much of the change shown in such studies may be an artifact of location error and classification error. The basic units of these two measures are different (distance units for location error and pixel counts for classification error). The lack of a single index summarizing both error sources poses a constraint on assessing and interpreting the apparent change. We present an error model that addresses location and classification error jointly. Our approach quantifies location accuracy in terms of thematic accuracy, using a simulation of the location error process. We further develop an error model that combines the location and classification accuracy matrices into a single matrix, representing the overall thematic accuracy in a single layer. The resulting time-specific matrices serve to derive indices for estimating the overall uncertainty in a multi-temporal dataset. In order to validate the model, we performed simulations in which known amounts of location and classification error were introduced into raster maps. Our error model estimates were highly accurate under a wide range of parameters tested. We applied the error model to a study of vegetation dynamics in California woodlands in order to explore its value for realistic assessment of change, and its potential to provide a means for quantifying the relative contributions of these two error sources.

Irrigated Crop Area Estimation Using Landsat TM Imagery in La Mancha, Spain

Abstract: A methodology for supervised classification to identify irrigated crops with Landsat TM imagery in a semiarid zone (La Mancha, Spain] is presented. The discrimination procedure is based on the diffrent crop spectral responses through time according to their phenological evolution. Our multitemporal supervised classification includes maximumlikelihood algorithms, decision-tree criteria, and context classifiers. We have applied the procedure to two sets of scenes obtained for the growing seasons of 3996 and 1997, respectively The resulting classification accuracy was 93.1 percent for 1996 and 90.21 percent for 1997. We have estimated the areas occupied by each crop class by means of intersecting the TM-derived land-use raster map and the digital rural cadastre vector map in a geographic information system. We have assessed the accuracy of the crop area estimation from the classified image by comparing these areas with those calculated from the digital rural cadastre. A median filter applied to the final classification improves the agreement of the estimated crop areas with the cadastre data. Additional post-classification methods to correct crop areas did not bring any significant further improvements. Therefore, we conclude that the context classifier is a useful and sufficient tool to improve sudace