Improved forest change detection with terrain illumination corrected Landsat images Bin Tan a,b, ⁎, Jeffrey G. Masek b , Robert Wolfe b , Feng Gao c , Chengquan Huang d , Eric F. Vermote d , Joseph O. Sexton d,f , Greg Ederer b,e a Earth Resources Technology, Inc., Laurel, MD, 20707, United States b NASA Goddard Space Flight Center, Greenbelt, MD 20771, United States c USDA-ARS Hydrology & Remote Sensing Lab, Beltsville, MD 20705, United States d Department of Geography, University of Maryland, College Park, MD 20742, United States e Sigma Space Corp., Lanham, MD 20706, United States f Global Land Cover Facility, University of Maryland, College Park, MD 20742, United States abstract article info Article history: Received 24 May 2012 Received in revised form 17 May 2013 Accepted 19 May 2013 Available online xxxx Keywords: Illumination correction Topographic effect Landsat LEDAPS TDA-SVM An illumination correction algorithm has been developed to improve the accuracy of forest change detection from Landsat-derived reflectance data. This algorithm is based on an empirical rotation model and was tested on Landsat image pairs over the Cherokee National Forest, Tennessee; Uinta–Wasatch–Cache National Forest, Utah; San Juan National Forest, Colorado; and Sinkyone Wilderness State Park, California. The illumination correction process successfully eliminated correlation between Landsat reflectance and illumination condi- tion. Comparison to forest-change maps derived from uncorrected images showed significant disagreement, ranging from 23% to 45%. Validated against high-resolution (1 m or less) time-serial images, the illumination correction decreased overestimation of forest gains and losses and improved specificity in detection of major forest changes. The overall accuracy increases 34% at the Cherokee Forest site and about 10% at the other three sites. The disagreement rate between change maps from the original and corrected Landsat images in- creased with increasing terrain inclination angle, with the relationship between illumination condition and the disagreement rate following a V-shaped curve that varied among sites. The lowest disagreement rate oc- curred when illumination condition was slightly smaller than that of a horizontal field. The correction for to- pographic illumination should be considered as a standard pre-processing step for land cover classification and land use change detection, especially for mountainous areas. © 2013 Elsevier Inc. All rights reserved. 1. Introduction Landsat imagery is widely used to monitor changes in land surface conditions, including changes in forest cover, which impact Earth's energy balance, carbon cycle, water cycle and biogeochemical pro- cesses (Band, 1993; Huang et al., 2008; Pandey, 2002). To quantify such changes, two Landsat images acquired before and after the forest change are typically examined by visual interpretation or automated analysis. With the Landsat archive becoming freely available, the main challenge for generating continental or global forest change maps at Landsat resolution (30 m) is an effective and accurate change detection algorithm (Huang et al., 2008). Various computer based change-detection algorithms have been developed (Foody & Mathur, 2004; Huang et al., 2010a, 2010b; Masek et al., 2008; Townshend et al., 2012). The key process in these algorithms is the spectral analysis of a set of training pixels and discriminating appropriate spectral thresholds that can be applied to the whole image scene or multiple scenes to define the area of forest change. One important assumption underlying these algorithms is that the spectral characteristics of the training pixels represent those of the forest pixels within the study region. However, topographic illumina- tion effects (shadow, slope, etc.) negate this assumption. Varying illumination conditions due to topography lead to significant changes in the observed spectral characteristics of a group of neighboring pixels, even in the absence of variations in land cover type or condi- tion. Therefore, illumination correction, also known as topographic correction or topographic normalization, is an important step in pre-processing high-resolution remote sensing data for forest change detection studies. Illumination correction refers to the compensation for solar irradiance to minimize the variability of observed reflectance for similar targets due to topography and Bidirectional Reflectance Distribution Function (BRDF) effects. Remote Sensing of Environment 136 (2013) 469–483 ⁎ Corresponding author at: Code 619, NASA GSFC, Greenbelt, MD 20707, United States. Tel.: +1 3016145965. E-mail address: bin.tan@nasa.gov (B. Tan). 0034-4257/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.rse.2013.05.013 Contents lists available at SciVerse ScienceDirect Remote Sensing of Environment journal homepage: www.elsevier.com/locate/rse