Detection and interpretation of geological linear features on the satellite images by using gradient filtering and principal component analysis

Abstract

Digital image processing has the ability to detect geological linear features on the image by using some algorithms. The most common algorithm which is used for this purpose is edge filtering. Gradient filtering is one type of edge filtering and is based on the calculation of the gradient along two perpendicular directions and obtain the largest absolute value of the two results. The directions of geological linear features are random in nature. One gradient filtering can be applied to only one specific direction for each time. So at least four or eight linear feature directions can be obtained for one specific area. The major dominant direction must be determined from these different directions. The filtered Landsat images are used for each directions. The principal component analysis is applied to all these filtered images for the detection of major direction. In the end, the result is verified with the tectonic lineaments of the studied area