Abstract: With the continuous deepening of mining activities,the problem of mining subsidence has become increasingly
prominent,seriously threatening environmental safety and social stability in mining area. In response to the shortcomings of traditional
subsidence monitoring methods,such as high cost,low efficiency,and delayed data updates,a method for identifying
subsidence areas in mines by integrating unmanned aerial vehicle (UAV) oblique photogrammetry technology and the K-means
clustering algorithm is proposed. Firstly,a UAV equipped with oblique photogrammetry equipment is used to collect high-resolution
images of the mining area from multiple angles and at multiple times. Then,a three-dimensional point cloud model of the
mining area is constructed through oblique photogrammetry technology to accurately locate the subsidence areas. On this basis,
the K-means clustering algorithm is applied to classify the point cloud data and automatically identify the subsidence areas.
Tests were conducted using the goaf area of the Guandi Coal Mine of Xishan Coal and Electricity as an example and compared
with 3D laser scanning technique,support vector machine (SVM),random forest (RF),DBSCAN clustering and convolutional
neural network. The results show that the accuracy rate of sinkhole identification by the proposed method reached 92. 5%. It
significantly outperformed the other algorithms in terms of identification accuracy and algorithm running efficiency,indicating
that this method has achieved significant improvements in both efficiency and accuracy compared to traditional algorithms,providing
technical support for mining sinkhole monitoring.

Key words: mining subsidence,unmanned aerial vehicle oblique photogrammetry,K-means clustering,identificaiton of , subsidence area,3D laser scanning technique,support vector machine