A fast and automatic algorithm for built-up areas classification in high-resolution SAR images based on geostatistical texture

Nowadays, main methods used to SAR imagery built-up areas classification are GLCM (gray-level co-occurrence matrix) textural analysis, Markov random field, etc. They are extraordinarily time consumption and need for manual interaction. In this paper, a new scheme for fast and automatic classification of built-up areas is presented. It is based on geostatistical texture analysis and mainly consists of four parts: semivariogram calculation, best lag distance finding, FCM (Fuzzy C-Mean) clustering, and edge detection. The experimental results show that it is robust, fast and accurate.