Landslide Susceptibility Mapping Based on Multitemporal Remote Sensing Image Change Detection and Multiexponential Band Math

Landslides pose a great threat to the safety of people’s lives and property within disaster areas. In this study, Zigui Badong section Three Gorges Reservoir is used as study area, land use (LU), change (LUC) band math (band) factors from 2016–2020 along with six selected commonly are form factor combination (LUFC), (LUCFC) (BMFC). An artificial neural network (ANN), support vector machine (SVM) convolutional (CNN) chosen three models for landslide susceptibility mapping (LSM). The results show that BMFC generally better than LUFC LUCFC. For validation set, highest simple ranking scores were obtained (37.2, 32.8 39.2), followed by (28, 26.6 31.8) LUCFC (26.8, 28.6 20); is, band-based predictions those based on LU LUC, CNN model provides best prediction ability. According four groups experimental ANNs, compared easier access, yields higher predictive performance, stronger stability. Thus, can replace LUC certain extent provide automatic real-time monitoring.