Development of Processing System for Composite Images of SPOT and LANDSAT for the Analyzes of Earth Environment and Natural Resources

点击翻译按钮获取中文摘要

The purpose of this research is to detect a subsurface fracture system and to estimate an environmental variation in urban areas by using SPOT images and LANDSAT TM images. The following is the summary of research results. (1) Geologic faults often correspond with lineaments appearing on satellite images. We developed a simple and effective algorithm, Segment Tracing Algorithm, which adopts a non-filtering approach and a directionally variable threshold for the exact extraction of lineaments. This method can detect lineaments existing in weak contrast area and constituting linear valleys. (2) The method to estimate the strike and dip of the fault constituting a lineament on an image was proposed. This calculation is based on a relationship between the curvature of lineaments and the dip of mountain slopes. It was shown using the method that the dominant azimuth of lineaments on the LANDSAT TM image and the SPOT image of the granitic massif agree well with the dominant azimuth of the joints appearing on the drifts in the massif. Consequently the self-similarity on azimuth distribution in the fracture system was revealed. (3) The neural network method was applied to classify a satellite image based on the superficial geology or the land-use pattern. (4) The reduction of ground-water level is partly due to development of the urban area, which is confirmed by processing of the two LANDSAT TM images for different years. (5) Degree of weathering, and the effect of surface temperature and water content of rock materials appear remarkably on the change of reflectance at 2215 nm. It was revealed by using a scanning visible to infrared radiometer that the reflectance of granite in infrared region is affected strongly by water content, the reflectance of basalt is affected strongly by surface temperature, and that of green tuff is not changed by neither of them.