Abstract: At present, the automation level of coal mine auxiliary operation in China is not high, and the degree of intelligence of relevant equipment is low. Among them, the installation equipment of underground auxiliary pipelines in coal mines are rarely studied, most mining enterprises still use traditional manual installation, which consumes a lot of labor. For these problems, the study of the control system of multi-axis mechanical arm in coal mines is carried out, including two aspects: (1) Based on embedded Linux operating system, multi-source sensor signal acquisition, analysis and fusion processing, multi-axis hydraulic manipulator and its control system, intelligent algorithm and other technologies, the perception and control system of an automatic installation robot of underground pipelines is designed and developed; (2) Based on the infrared video image and laser radar point cloud data collected by the perception and control system of multi-axis mechanical arm, the pipe support target recognition methods based on image and point cloud data are proposed respectively. In the actual underground coal mine tunnel pipe support scene, the proposed target recognition methods can effectively identify and tag the pipeline and pipe support in the tunnel, which shows the potential of the proposed methods to be applied to the actual coal mine automation auxiliary installation of pipeline.