Earthquake landslide detection method combining post-earthquake coherence increase and polarized vegetation damage

Rapid detection and accurate assessment of earthquake landslides are of great significance for disaster response and mitigation efforts. Although the method for detecting increased coherence in post-earthquake landslide areas can effectively capture surface changes caused by landslides, relying solely on information about changes in coherence still carries the risk of false detections and missed detections. Therefore, a landslide detection method based on Combining Coherence Increase and Polarized Vegetation Damage (CCI-PVD) is proposed. The method first optimizes coherence data to minimize interference from decoherence factors, then introduces the dual-polarization radar vegetation index (DpRVI) to quantify post-earthquake vegetation damage, and finally uses wavelet transform (WT) to fuse this information with coherence increase data for improved detection accuracy. Applied to the 2022 Luding earthquake (Sichuan, China), CCI-PVD achieves a 66.4% correct detection rate (CDR), significantly outperforming Co-seismic Coherence Loss (CSCL; 37.6%) and PECI (47.1%). It also reduces false detection rates (FDR: 25.8% vs. CSCL’s 53.1% and PECI’s 41.3%) and missed detection rates (MDR: 33.6% vs. CSCL’s 62.3% and PECI’s 50.6%). Statistical validation (95% confidence intervals, risk difference, and relative risk) confirms CCI-PVD’s superior accuracy and robustness, demonstrating its potential for improving post-earthquake landslide detection capabilities.