Landslide detection with multi-dimensional histogram equalization for multispectral remotely sensed imagery

Landslides are serious natural hazards. They usually cause significant properties damage and sometimes also fatalities. For reconstruction and rescue purposes, landslide detection is very important. Normalized Difference Vegetation Index (NDVI) is widely used for this purpose. If NDVI significantly decreases within a short period, that area will be considered as landslide candidate. However, remote sensing images always contain atmospheric effects which affect the NDVI estimation. In this study, we propose a multi-dimensional histogram equalization algorithm as a pre-process step. It modifies multispectral images collected under different atmospheric conditions to have similar spectrum for the same land cover. A set of SPOT images is adopted for experiments and preliminary results show the proposed method can reduce the misclassification rate.