Algorithm for Automatic Extraction of Lineaments from SAR Image Data
Yoichi Kageyama, Makoto Nishida, Tomihiro Utsumi · IEEJ Transactions on Electronics Information and Systems · 1998
This paper proposes a new method for the automatic extraction of the lineaments on the SAR (Synthetic Aperture Radar) image data. It is well known that the lineament patterns are the important feature reflect the structures of the subsurface. A lineament map has been drawn by the fieldworks of experts and the analysis of the enhanced image data using visual interpretation. It is therefore indicated that automatic extraction of the lineaments on the SAR image data is useful and effective for understanding the geological structure in the study area.This paper generates the edge data which contains the lineaments using some operators and template matching at first. Next, the emphasis of edge with the scored method is carried out. It is assumed that the edge pixel having high scored value can be much related to the lineaments. Finally, the trace and connection of the edge is conducted with the above assumption. As a result, it is confirmed that the results with the proposed method correspond to the lineament map drawn by experts. Therefore, it can be concluded that the proposed method is the effective technique for automatic extraction of the lineaments on the SAR image data.