Optimal Methodology for Detecting Land Cover Change in a Forestry, Lakeside Environment Using NAIP Imagery

Abstract

Mapping land cover change is useful for various environmental and urban planning applications, e.g. land management, forest conservation, ecological assessment, transportation planning, and impervious surface control. As the optimal change detection approaches, algorithms, and parameters often depend on the phenomenon of interest and the remote sensing imagery used, the goal of this study is to find the optimal procedure for detecting urban growth in rural, forestry areas using one-meter, four-band NAIP images. Focusing on different types of impervious covers, the authors test the optimal segmentation parameters for object-based image analysis, and conclude that the random tree classifier, among the six classifiers compared, is most optimal for land use/cover change detection analysis with a satisfying overall accuracy of 87.7%. With continuous free coverage of NAIP images, the optimal change detection procedure concluded in this study is valuable for future analyses of urban growth change detection in rural, forestry environments.