关键词: 井下狭小空间, 无人机, 自主定位, SLAM定位算法

Abstract: The area spaces of Meishan Iron Mine refered to the shaft of the underground air flow shaft，the empty area formed by blasting，the underground ore bin，etc. are relatively small and the loca‐ tions are rather dangerous. There is a risk of edge operation or operation close to unknown empty areas. Per‐ sonnel or instruments and equipment are exposed to injuries such as falling from heights and being hit by falling objects. The measurement operation is rather dangerous and difficult. However，accurately obtaining the 3D model of the confined space underground is of great significance for guiding the scientific mining of mines. The measurement technicians of Meishan Iron Mine proposed to effectively combine (UAV) with 3D laser scanning radars，fully leveraging the flexible and efficient features of UAV to carry 3D laser scanning radars and collect point clouds of physical objects in confined spaces. By developing the SLAM positioning algorithm for UAV and the broadband wireless AD hoc network data transmission system，the tasks of data collection and transmission are accomplished. Through data thinning and modeling，the deformation of the data model can be effectively prevented and the data accuracy can be improved. Using this method，after multiple production chiller shafts were evacuated underground，the surveying personnel of Meishan Iron Mine collected the solid point cloud data of the inner walls of the shafts. Through data processing，model‐ing，and generating longitudinal and transverse sectional views of the shafts for specific data analysis，they obtained the specific data information of each chiller shaft，providing detailed first-hand data for the main‐ tenance and daily management of underground chiller shafts. By collecting the physical point cloud data of the underground blasting empty area through unmanned aerial vehicles and then establishing a 3D model， the length，width，height and specific shape of the blasting empty area can be accurately grasped，provid‐ ing a basis for accurately handling faults in the blasting empty area and eliminating potential safety hazards in production. By using UAV with autonomous positioning to carry 3D laser scanners，it is possible to com‐ plete the data collection and 3D modeling of underground narrow and complex space entities，achieving im‐ provements in operation efficiency，safety factors，and data accuracy.

Key words: underground confined spaces, UAV, autonomous localization, SLAM positioning algo‐ rithm