ABSTRACT
Industrialization is criticized for posing environmental concerns despite its contribution to local economic growth and modern urbanization. Accurately mapping industrial zones is a necessity for the strategic planning of modern industrialization. Unfortunately, industrial zones mapping by remotely sensed imagery over large areas remains methodologically challenging, as the spectral and texture features of industrial zones resemble those of other man-made surfaces. To fill the gap, this paper presents the first attempt at industrial zones mapping through identifying the ‘blue steel tile’-roofed buildings (BSTBs). Specifically, we develop a BSTB index (BSTBI) for extracting the BSTBs in Landsat images. The new index combines the BSTB feature of ‘V-shaped valley’ over visible-near infrared reflection (NIR) span and the most separable band of Landsat sensors (shortwave infrared-2 band). Through the experiments conducted in Linyi city, China, the new index is demonstrated a good performance in detecting BSTBs with an overall accuracy larger than 85% and Kappa coefficient over 0.75 under the support of atmospherically corrected Landsat data.