Abstract: The underground coal mine environment is dim, humid and dusty, and the collected video images are blurred, which makes it difficult to identify the longitudinal tear of the conveyor belt. The existing research results only focus on whether there is longitudinal tear, without involving the location of tear damage and trend tracking. This paper proposes a method for longitudinal tearing detection of belt conveyors based on multi-channel linear laser. This method uses a mine intrinsically safe structured light emitter to project multiple linear lasers onto the surface of the conveyor belt, and uses a mine intrinsically safe industrial camera to take linear laser stripe images. By extracting the center line of the laser stripe and analyzing its characteristics, this method determines whether there is longitudinal tear damage in a single frame image. When there is damage, this method can search for damage boundary points and calculate damage width and depth characteristic values. The method fuses multi-frame image detection results and speed sensor values to calculate the complete longitudinal tear damage length, average width and average depth. By detecting the position of the marker on the conveyor belt, the method can locate the longitudinal position of the tear damage and locate the lateral position by finding the bit width ratio of the starting point of the damage on the conveyor belt. And the damage trend tracking is realized based on the longitudinal position. The method is tested by setting the conveyor belt longitudinal tear damage length to 0.73, 0.95 m, the average width to 0.01 m, the average depth to 0.008 m, and the sampling rate to 25 frames/s. The results show that the method can accurately detect whether the belt conveyor has longitudinal tears, the average error of damage length calculation is 0.06 m, the average error of damage average width and average depth calculation is 0.001 m, and the error of longitudinal positioning is less than 0.1 m. It can determine the development trend of longitudinal tears accurately.