Abstract: We proposed a multi-directional mine rigid tank deformation diagnosis and location method based on machine vision and laser fusion. We used CCD camera, laser, screen, LabVIEW software and PC deformation diagnosis and real-time positioning system to collect light spot image on the screen, carried out image enhancement, threshold segmentation, image matching and tracking, pixel centroid calculation and other processes for the collected spot image, displayed the status and position of tank channel on the PC monitoring interface. The recognition degree of the step bulge height is 88.43%, the length recognition degree of the step bulge is as high as 99.15%, and the depth recognition rate is as high as 99.52%. The experiment proves that the technology can accurately diagnose the deformation of the rigid tank channel in the mine and locate it quickly.