Automatic Landslide Detection

One of the most frequent natural disasters is a landslide. It is defined as the downward movement of soil and rock due to gravitational force. It leads to loss of people and property. So, landslide detection plays a crucial role in saving people's life. In this project work, we took sentinel-l satellite dataset of Kodagu(formally known as Coorg) in Karnataka, India. The images are pre-processed using CLAHE(Contrast Limited Adaptive Histogram Equalization) and pixels based classification will be applied to extract the low-level features. The image background is removed by Digital Elevation Model(DEM). Using morphological analysis, the small objects that distinguish landslides from another are removed by taking maxima and minima points. Landslides are detected from that image using H-BEMD(Hue-Bi-Directional Empirical Mode Decomposition). H-BEMD is the improved version of BEMD(Bi-Directional Empirical Mode Decomposition) along with Hue Channel.