Extraction of vegetation cover rate in urban areas based on mixel analyses of Landsat data

摘要

This study aims the development of a practical method for the extraction of vegetation cover rate (VCR) in urban areas using Landsat TM/ETM+ data. The linear mixture model in which two main categories, vegetation and non-vegetation, have two sub-categories respectively is employed combined with the least square estimation using six bands data of TM/ETM+ data. The experimental results show fairly good coincidence between the VCRs from Landsat and the ground surface conditions from ground survey, and high correlation between the VCRs from Landsat and from Quickbird. The experiments also show the possibility to generate VCR distribution maps in urban areas, and to extract their yearly changes. In addition the relationships between the VCRs estimated in this study and some vegetation indices are investigated. Finally the algorithm is modified to extract VCR conditions in wider urban areas, Tokyo-metropolis and Kinki-district, the biggest and the secondary biggest urban areas in Japan respectively.