Objective Underground spatial surveying of coal mine is an important part of transparent geological modelling for coal mine. However, the complex environment of coal mine roadway, incomplete information acquisition, insufficient accuracy of sensed data and other difficulties need to be urgently solved.

Methods First, the signal characteristics of millimeter wave radar under the complex environmental factors such as dust, water mist, surrounding rock structure, etc. were studied in depth, the millimeter wave signal attenuation model was established for the surrounding rock of coal mine roadway, and the influence mechanism of complex environmental factors for millimeter wave radar was comparatively analyzed. Second, in view of the problem of millimeter wave radar based digital roadway modelling in the complex environment of coal mine, the Poisson surface reconstruction method based on millimeter wave radar point cloud for the roadway space was proposed. Through experimental tests and simulated roadway environment verification, the spatial surveying and digital modelling reconstruction of the coal mine roadway was achieved on the basis of revealing the millimetre-wave radar sensing mechanism under the complex environmental conditions of the coal mine.

Results and Conclusions  The results show that: (1) The millimetre-wave radar could adapt to the underground coal mine roadway environment with a lot of dust, water mist and rough surrounding rock, and provide effective data for the spatial reconstruction of the coal mine roadway. (2) The Poisson surface reconstruction method could fully display the real information on surrounding rock of roadway, and the average absolute error percentage is 0.59% for the reconstructed total roadway width, and 0.78% for the total roadway height. Generally, the study of millimeter-wave radar spatial surveying characteristics and reconstruction methods in the complex environment of coal mines provides the spatial surveying data of roadway for transparent geological modelling of coal mines, which is of great significance in promoting the intelligent mining of coal mines.