Digital modelling method of coal-mine roadway based on millimeter-wave radar array
Xusheng Xue,Xingyun Yang,Jianing Yue,Qinghua Mao,Yihan Qin,Hongwei Ma,Jianxin Yang,Huahao Wan,Enqiao Zhang,Junbiao Qiu,Xiaopeng Li,Rongquan Wang
DOI: https://doi.org/10.1038/s41598-024-69547-5
2024-08-10
Abstract:The roadway space of coal mine is narrow, and the illumination is low and uneven. Dynamic mining is accompanied by dust and water mist. The accuracy and reliability of roadway data collected by vision and laser sensors are poor. Based on this, a digital modeling method of coal mine roadway based on millimeter-wave radar array is proposed. Firstly, aiming at the problem of complex environmental interference, combined with the characteristics of small amount of single frame data of millimeter-wave point cloud, a multi-layer filtering noise reduction processing and dynamic subgraph registration method of millimeter-wave point cloud is proposed to filter out interference points and realize single radar point cloud registration. Secondly, aiming at the problem that a single millimeter-wave radar cannot scan the complete roadway information at one time, combined with the characteristics of small elevation field of view of millimeter-wave radar, a millimeter-wave radar array acquisition system is built, and an improved iterative closest point (ICP) registration algorithm based on point cloud features is established to construct the roadway point cloud fusion model. Finally, aiming at the problem of uneven and sparse point cloud after array fusion, a Poisson surface reconstruction method based on point cloud density weighted interpolation is proposed to refine the roadway structure and realize the accurate reconstruction of digital roadway model. The experimental results show that the digital modeling method of millimeter-wave radar array can accurately obtain the information of roadway surrounding rock, the density of roadway point cloud is increased by 22.4%, and the average absolute error percentage of the width and height of the reconstructed roadway model is only 0.82% and 0.72%, which provides a new research method for the reconstruction of underground roadway and an important basis for the digital modeling method of coal mine roadway.