Study on the identification of karst surface collapse via UAV photography based on the YOLO model

Surface collapse is a relatively common geological disaster in karst areas, characterized by sudden occurrence and continuous development. Large-scale ground construction operations need to identify it in advance to reduce its adverse effects. This paper, based on the Zala Mountain photovoltaic project of the Jinping Hydropower Station in Liangshan Prefecture, Sichuan Province, conducts research on the identification and detection of karst surface collapse. Field topographic and geomorphic data are obtained through Unmanned Aerial Vehicle photography and augmented processing. The You Only Look Once model is used for target recognition of karst surface collapse, and the effectiveness of the model is evaluated and analyzed based on indicators such as recall rate and precision rate, which can provide technical support for engineering construction in karst subsidence area.