Processing landslide hazard prediction using satellite image segmentation

Remote sensing technology has become one of the indispensable techniques for the investigation of geologic disasters at home and abroad. However, the interpretation of bare ground classification by satellite multispectral images is often interfered by atmospheric clouds in mountainous areas, shadows on the mountains, and the phenomenon of Speckle Signals, which seriously affects the correctness of the interpretation images, and is prone to errors such as misjudgment or omission. In this study, the SPOT 5 satellite images of the study area were segmented by applying the Multiphase Level Set Framework using MATLAB computation. The results show that the problem of the speckle signals can be eliminated efficiently and quickly, and the complex terrain and features can be segmented into simple objects, which can solve the problem of the spatial recognition rate without being affected by the resolution of the image itself, and improve the image accuracy. This can solve the problem of spatial recognition without being affected by the resolution of the image itself, and improve the accuracy of satellite image interpretation, which can be used as a possible landslide trend.