Remote Sensing and GIS Accuracy Assessment - Chapter 19

CHAPTER 19 A Technique for Assessing the Accuracy of Subpixel Impervious Surface Estimates Derived from Landsat TM Imagery S. Taylor Jarnagin, David B. Jennings, and Donald W. Ebert CONTENTS 19.1 Introduction...........................................................................................................................269 19.2 Methods................................................................................................................................270 19.2.1 Study Area................................................................................................................270 19.2.2 Data...........................................................................................................................272 19.2.3 Spatial Processing.....................................................................................................272 19.2.4 Statistical Processing................................................................................................274 19.3 Results and Discussion.........................................................................................................275 19.4 Conclusions...........................................................................................................................277 19.5 Summary...............................................................................................................................278 Acknowledgments..........................................................................................................................279 References......................................................................................................................................279 19.1 INTRODUCTION An emerging area in remote sensing science is subpixel image processing (Ichoku and Karnieli, 1996). Subpixel algorithms allow the characterization of spatial components at resolutions smaller than the size of the pixel. Recent studies have shown the general effectiveness of these techniques (Huguenin, 1994; Huguenin et al., 1997). The importance of subpixel methods is particularly relevant to the field of impervious surface mapping where the predominance of the “mixed pixel” in medium-resolution imagery forces the aggregation of urban features such as roadways and rooftops into general “developed” categories (Civco and Hurd, 1997; Ji and Jensen, 1999; Smith, 2001). The amount of impervious surface in a watershed is a landscape indicator integrating a number of concurrent interactions that influence a watershed’s hydrology, stream chemical quality, and ecology and has emerged as an important landscape element in the study of nonpoint source pollution (NPS) (USEPA, 1994). As such, Schueler (1994) proposed that impervious surfaces should be the single unifying environmental theme for the analysis of urbanizing watersheds. Effectively extracting the percentage of impervious surface from medium-resolution imagery would provide 269 © 2004 by Taylor & Francis Group, LLC 270 REMOTE SENSING AND GIS ACCURACY ASSESSMENT a time and cost savings as well as allowing the assessment of these landscape features over extensive geographic areas such as the Chesapeake Bay. As part of the Multi-Resolution Landscape Charac-terization 2000 program (MRLC 2000, 2002), the United States Geological Survey (USGS) has embarked on an effort to map impervious surfaces across the conterminous U.S. utilizing subpixel techniques. This study proposes to produce a spatial and statistical framework from within which we can investigate subpixel-derived estimates of a material of interest (MOI) utilizing multiple accuracy assessment strategies. Traditional map accuracy assessment has utilized a contingency table approach for assessing the per-pixel accuracy of classified maps. The contingency table is referred to as a confusion matrix or error matrix (Story and Congalton, 1986). This type of assessment is a “hit or miss” technique and produces a binary output in that a pixel is either “correct” or “not correct.” The generally accepted overall accuracy level for land-use (LU) maps has been 85% with approximately equal accuracy for most categories (Jensen, 1986). While alternative techniques to assess the accuracy of land-cover (LC) maps using measurement statistics such as the Kappa coefficient of agreement have been proposed, most methods still rely on the contingency table and use per-pixel assessments of the thematic map class compared to “truth” sample points (Congalton and Green, 1999). However, as noted by Ji and Jensen (1999), this classic “hit or miss” approach is problematic with respect to assessing the accuracy of a subpixel-derived classification. A subpixel algorithm allows the pixel to be classified based on the percentage of a given MOI such that for any given pixel the “fit” to truth can be assessed. A level of accuracy can be still be obtained from a pixel that “misses” the truth. The derivation of a percentage of a MOI per-pixel allows for alternative accuracy assessment approaches such as aggregate whole-area assessments (i.e., watershed) and correlations (Ji and Jensen, 1999). These alternative approaches may produce adequate accuracies despite the fact that a lower per-pixel accuracy is derived from the standard error matrix. An accuracy assessment of subpixel data is largely dependent upon high-resolution planimetric maps or images to provide reference data. Concurrent with the emergence of subpixel techniques has been a trend in the production of high-resolution data sets, including high-resolution multispectral satellite imagery, GIS planimetric data, and USGS Digital Ortho Quarter Quads (DOQQs). All these data sources can be readily processed within standard GIS software packages and used to assess the accuracy of subpixel estimates, as derived from Landsat data, over large geographic regions. In this study we compared classified subpixel impervious surface data derived from Landsat TM imagery and planimetric impervious surface maps produced from photogrammetric mapping processes. Comparisons were performed on the classified subpixel (30 m) using planimetric refer-ence data in a raster GIS overlay environment. Our goal was to produce a spatial framework in which to test the accuracy of subpixel-derived estimates of impervious surface coverage. In addition to a traditional per-pixel assessment of accuracy, our technique allowed for a correlation assessment and an assessment of the whole-area accuracy of the impervious surface estimate per unit area (i.e., watershed). The latter is important for ecological and water quality models that have percentage of impervious surface as a variable input. 19.2 METHODS 19.2.1 Study Area Our study area was the Dead Run watershed, a small, 14-km2 subwatershed located 9 km west of Baltimore, Maryland (Figure 19.1). The Dead Run subwatershed is a portion of the greater Baltimore Long Term Ecological Research (LTER) area located in Baltimore County, Maryland, and resides within the coastal plain and piedmont geologic areas of the Mid-Atlantic physiographic region. Previously produced planimetric and subpixel data sets were available for the area. © 2004 by Taylor & Francis Group, LLC Figure 19.1 Location of the 14-km2 Dead Run subwatershed approximately 9 km west of Baltimore, Maryland. © 2004 by Taylor & Francis Group, LLC 272 REMOTE SENSING AND GIS ACCURACY ASSESSMENT Table 19.1 The University of Maryland Mid-Atlantic RESAC Impervious Surface Percentage per Pixel Classes Impervious Class 0 1–10 11–20 21–30 31–40 41–50 51–60 61–70 71–80 81–90 91–100 Note: Classes are represented in the raster data as 10, 20, etc., such that class 1–10 = 10, 11–20 = 20, etc. 19.2.2 Data Subpixel impervious surface cover data derived from TM imagery were provided by the University of Maryland’s Mid-Atlantic Regional Earth Sciences Application Center (RESAC) impervious surface mapping effort (Mid-Atlantic RESAC, 2002). The Mid-Atlantic RESAC process utilized a decision tree classification system to map 11 different levels of impervious surface percentage per 30-m pixel (Table 19.1) (Smith, 2001). Reference data were obtained using photo-grammetrically derived GIS planimetric vector data provided by Baltimore County, Maryland. The vector data included anthropogenic features such as roads, parking lots, and rooftops but did not include driveways associated with single-family homes. The lack of compiled driveways was a limitation of the truth set and has the potential to be a source of error. The Dead Run subwatershed was delineated using USGS Digital Raster Graphics (DRG) and “heads-up” digital collection methods. The compiled Dead Run subwatershed was subsequently utilized to clip both the Mid-Atlantic RESAC raster data and the Baltimore county impervious surface planimetric data. This produced a spatially coincident Dead Run 30-m subpixel estimate GRID and a Dead Run impervious surface truth vector file (Figure 19.2). All data were processed in the UTM Zone 18, NAD83 projection. The respective data sets were independently registered (prior to our study) and no attempt was made to coregister the data via image-to-image methods. 19.2.3 Spatial Processing GIS raster overlay techniques were utilized to compute the reference values for percentage of impervious surface for each 30-m grid cell within the Dead Run subwatershed. The process was a modified form of zonal analysis. Here, however, the zones are the individual 30-m classified pixels as opposed to individual land LU/LC zones. This method was a variation of the overlay processes reported by Prisloe et al. (2000) and Smith (2001) and included the following analysis procedures: · A vector-to-raster conversion of the Dead Run impervious surface reference data was performed to produce a high-resolution (3-m) impervious surface grid cell (0 = nonimpervious, 1 = impervious). · A comparison of the classified 30-m Dead Run data with the 3-m impervious surface reference data was performed using an overlay process, which calculated the number of reference data cells spatially coincident with the classified data (Plate 19.1). The count of coincident reference data cell percentage for each Dead Run grid cell was tallied. © 2004 by Taylor & Francis Group, LLC Figure 19.2 Two graphics of the Dead Run subwatershed showing the separate data types of (a) subpixel derived estimates of impervious surface percentage and (b) truth impervious surface vector file. © 2004 by Taylor & Francis Group, LLC ... - tailieumienphi.vn