Image Processing Methods of LANDSAT TM Data for Lineament Analysis

Abstract

Lineaments in Landsat TM images were analyzed using a image processing system consisting of apersonal computer, PIAS LA-500 and full color board. We studied the granitic body of stockform and cretaceous rocks in the northern part of Kumamoto prefecture. The first analysis of the digital terrain model of the granitic body revealed two prominent trends of maximum dip of slopes, NE-SW and NW-SE. The second revealed contrast modifications of the original TM images. The best method was obtained by flattening the frequency of the gray level (encoded brightness) in the original images. False color composition, which clarifies the geological structures of the granitic body. was carried out using these modified TM images. While TM images have 7 bands of wavelength, the data of band 5 corresponds well with the characteristics of the digital terrain model.
Furthermore, we evaluated a method of automatic detection of lineaments with a corrective function and differential filters. This method reduced the effect of the sun azimuth on the contrast of TM images and showed fracture patterns in the granitic body in detail. The fracture patterns were similar to those obtained by the analysis of aerophotographs of the studied area.