Abstract: The research and development of the coal mine sump dredging robot can greatly improve the dredging efficiency of the water sump and ensure the reliability of the water sump. In the research and development process of the coal mine water sump dredging robot, the identification of the sludge interface has become a key basic problem in the research and development of the coal mine sump dredging robot, and a method combining 2D lidar and depth camera was proposed to achieve accurate identification of the sludge interface. Firstly, the internal parameters and external parameters of the camera were obtained by the joint calibration of the lidar and the depth camera in ROS, and the transformation relationship between the lidar data and the camera data was established. Then, the noise of the 3D point cloud obtained by the depth camera is reduced by Gaussian filtering, and the density is reduced by means filtering. Finally, by superimposing the converted lidar data with the camera data, the 3D point cloud data of the depth camera was filtered through the depth mean of the lidar sampling points to realize the identification of the sludge interface. Experimental results show that the method can effectively identify the sludge interface, and lay a foundation for the intelligent and mechanized cleaning of water sump.