ASPRS

PE&RS August 2008

VOLUME 74, NUMBER 8
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
JOURNAL OF THE AMERICAN SOCIETY FOR PHOTOGRAMMETRY AND REMOTE SENSING

Peer-Reviewed Articles

973 MODIS-based Change Detection for Grizzly Bear Habitat Mapping in Alberta
Alysha D. Pape and Steven E. Franklin

Abstract  Download Full Article
Coarse resolution data from the Moderate Resolution Imaging Spectroradiometer (MODIS) was used to test the effectiveness of 250 m data to detect forest disturbances and update an existing, large-area (150,000 km2), 30 m Landsat ETM+ and TM land-cover map product used in Grizzly Bear (Ursus arctos) habitat analysis. A Landsat-derived polygon layer was applied to the MOD13Q1 data product to create a polygon-based, mean NDVI time series (2000 to 2005). Image differencing of the dataset produced multiple-scale layers of change including a two-date, five-year change and a fiveyear composite of annual changes. Accuracy assessments based on available GIS data showed an overall accuracy as high as 59 percent. Results also show that disturbance patches larger than 15 ha were represented with an accuracy of 75 percent or higher. This offers an alternative to higher spatial resolution data for the identification of larger features and also provides general change information for those areas that may be suitable for analysis with higher spatial resolution data.

987 Quantitative Mapping of Hydrodynamic Vegetation Density of Floodplain Forests Under Leaf-off Conditions Using Airborne Laser Scanning
Menno W. Straatsma

Abstract  Download Full Article
In this paper a method is presented to extract hydrodynamic vegetation density from airborne laser scanner data, relevant for exceedance levels of embankments of lowland areas. Two indices to predict vegetation density from the laser data were considered: (a) Percentage Index (PI) of points in the height interval inundated by the water, and (b) the Vegetation Area Index (VAI) that corrects for occlusion from the crown area. A computer simulation, using a digital forest model, showed a sensitivity of the indices for laser pulses that were sent out, but not detected by the laser receiver. The locations of these invalid points were therefore reconstructed. Two different assumptions were tested to assign new coordinates to these so-called invalid points. Percentage Index, with the invalid points reconstructed by means of thresholding the point density ratio, proved the best predictor (R2 = 0.66) of vegetation density of deciduous floodplain forests under winter conditions..

999 Urban Change Detection Based on Coherence and Intensity Characteristics of SAR Imagery
Mingsheng Liao, Liming Jiang, Hui Lin, Bo Huang, and Jianya Gong

Abstract  Download Full Article
In this paper, an unsupervised change-detection approach was proposed to detect new urban areas from multi-temporal SAR images. The novelty of the proposed approach is the joint use of coherence and intensity characteristics of SAR imagery. The approach involves two main steps: (a) the extraction of difference feature containing information on changed areas, and (b) the unsupervised two-dimensional (2D) thresholding. First, two difference features based on the concepts of long-term coherence and backscattering temporal variability are extracted from a series of multitemporal SAR images. Then, the resulting features that represent the INSAR signal temporal variability of changed areas are merged, and a 2D thresholding technique based on the maximum 2D Renyi’s entropy criterion is developed to obtain the change-detection results. The effectiveness of the proposed approach is confirmed with experimental results obtained from a set of six ERS-1/2 SLC SAR images acquired in Shanghai, China.

1007 Factors Affecting Spatial Variation of Classification Uncertainty in an Image Object-based Vegetation Mapping
Qian Yu, Peng Gong, Yong Q. Tian, Ruiliang Pu and Jun Yang

Abstract  Download Full Article
Much effort has been spent on examining the spatial variation of classification accuracy and associated factors on a per-pixel basis. In the past few years, object-based classification has attracted growing interest. This paper examines factors affecting the spatial variation of classification uncertainty in an object-based vegetation mapping. We studied six categories of factors in an object-based classification: general membership, topography, sample object density, spatial composition, sample object reliability, and object features. First, classification uncertainty (classification accuracy on a per-case basis) is derived with a bootstrap method. Then, six categories of factors are quantified by categorical or continuous variables. In this step, the appropriate radius for calculating the spatial composition metrics of sample objects is also discussed. Finally, classification uncertainty is modeled as a function of those factors using a mixed linear model. The significant factors are identified and their parameters are estimated from restricted maximum likelihood fit. The modeling results show that elevation, sample object size, sample object reliability, sample object density, and sample spatial composition significantly influence the object-based classification uncertainty. Many of these factors are closely related to the object-based approach. The result of this study helps in understanding classification errors and suggests further improvement of the classification.

Color Figures (Adobe PDF format):

[figure 1.] [figure 7.]

1019 Evaluating Neural Networks and Evidence Pooling for Land Cover Mapping
M.J. Aitkenhead, S. Flaherty, and M.E.J. Cutler

Abstract  Download Full Article
The diversity of data sources, analysis methodologies, and classification systems has led to a number of new techniques for monitoring land-cover change. However, this wide choice means that it is difficult to know which solution to choose. A system capable of integrating the results of different analyses and applying them to land-cover mapping would therefore be extremely useful. This study investigates the use of evidence pooling and neural networks in land-cover mapping. Neural networks were used to classify land-cover using evidence from spectral (Landsat-7 ETM1), textural, and topographic information. Mapping was performed using combinations of evidence source and evidence pooling techniques. The best performance was achieved using all available information with a method that summed evidence directly instead of categorizing it. While the methodology failed to reach the level of accuracy recommended elsewhere, a comparison of the number of classes used with other methods showed that the system performed better than these approaches.

1033 Lidar-based Mapping of Forest Volume and Biomass by Taxonomic Group Using Structurally Homogenous Segments
Jan A.N. van Aardt, Randolph H. Wynne, and John A. Scrivani

Abstract  Download Full Article
This study evaluated the potential of an object-oriented approach to forest type classification as well as volume and biomass estimation using small-footprint, multiple return lidar data. The approach was applied to coniferous, deciduous, and mixed forest stands in the Virginia Piedmont, U.S.A. A multiresolution, hierarchical segmentation algorithm was applied to a canopy height model (CHM) to delineate objects ranging from 0.035 to 5.632 ha/average object. Per-object lidar point (per return height and intensity) and CHM distributional parameters were used as input to a discriminant classification of 2-class (deciduous-coniferous) and 3-class (deciduous-coniferous-mixed) forest definitions. Lidar point-height-based and CHM classifications yielded overall accuracies of 89 percent and 79 percent, respectively. Volume and biomass estimates exhibited differences of no more than 5.5 percent compared to field estimates, while showing distinctly improved precisions (up to 45.5 percent). There were no significant differences between accuracies for varying object sizes, which implies that reducing the lidar point coverage would not affect classification accuracy. These results lead to the conclusion that a lidar-based approach to forest type classification and volume/biomass assessment has the potential to serve as a single-source inventory tool.

1045 Comparison of Spectral Analysis Techniques for Impervious Surface Estimation Using Landsat Imagery
Fei Yuan, Changshan Wu, and Marvin E. Bauer

Abstract  Download Full Article
Various methodologies have been used to estimate and map percent impervious surface area (%ISA) using moderate resolution remote sensing imagery (e.g., Landsat Thematic Mapper). There is, however, a lack of comparative analyses among these methods. This study compares three major spectral analysis techniques (regression modeling, regression tree, and normalized spectral mixture analysis (NSMA)) for continuous %ISA estimation using Landsat imagery for 1986 and 2002 for the seven-county Twin Cities Metropolitan Area of Minnesota. Our study showed that all three techniques demonstrate the capability for estimating %ISA accurately, with RMSE ranging from 7.3 percent to 11 percent and R2 of 0.90 to 0.96 for both years. Comparatively, regression modeling and regression tree methods produced similar results; however, both of them are highly dependent on accurate masks to differentiate urban impervious surfaces from bare soil. Within the urban mask, the regression tree-based estimates were the most accurate. In terms of time and cost, the NSMA approach is most efficient, but it tends to underestimate the percent imperviousness for highly developed areas. Findings from the study provide guidance for the selection of %ISA estimation techniques using moderate resolution remote sensing data, along with information for further methodological improvements.

Top Home