Research on the landslide evolution mechanism driven by ground radar images of slopes
摘要
Aiming at the difficulty in predicting deformation and evolution of high and large slopes in open-pit mines, this study proposes a new approach. Based on periodic ground radar monitoring images, we applied OpenCV technology and introduced a novel indicator—Cumulative Difference Degree of Image Data-Time (CDD-T)—to construct a landslide evolution description method. Verified at the south slope deformation area of an open-pit coal mine (Xilingol League, China), the method identified dangerous areas via nn×nn refined segmentation and CDD-T curve analysis: slope toe (+ 984 to + 960) and southwest side (+ 1030 to + 984), with potential traction-type landslide mechanism confirmed by CDD-T heat maps. CDD-T curves were highly consistent with cumulative displacement-time curves (minimum Pearson coefficient: 0.9326; average: 0.9629). Notably, 89% of CDD-T curves provided earlier warnings, offering new insights for open-pit mine landslide research.