The hiddenness and unexpectedness of dangerous rock mass in high-steep rock slopes makes its quick identification and stability assessment in the early stage one of the major technical problems in geological disaster management. Taking Shizitou in Wanzhou, Chongqing as an example, micro unmanned aerial vehicles (UAVs) were used to reasonably plan the route to obtain high-definition (HD) images and generate two- and three-dimensional models. The combination of remote sensing information extraction technology and geological disaster analysis provides basic data for the study of dangerous rock mass stability. Use of image recognition technology was applied to identify the structural surface of dangerous rock mass, and basic data such as the coordinates extracted earlier was used to calculate its occurrence. Finally, the stability of typical dangerous rock mass was analyzed by erythroid projection. The research results provide a new idea for the non-contact measurement, acquisition of fine geological information and quick stability assessment of high dangerous rock mass.