A new machine vision detection method for identifying and screening out various large foreign objects on coal belt conveyor lines

Abstract Large foreign object transporting by coal mine conveyor belt may lead to production safety hazards. To reduce accidents during mining, a large detection method based on machine vision is proposed in this paper. An adaptive weighted multi-scale Retinex (MSR) image enhancement algorithm improve the captured quality of line. improved template matching designed combining frame difference and area methods screen identify objects mixed coals. The multi-layer perceptron (MLP) network optimized Gray Wolf introduced objects. Experimental results show that identification accuracy reaches 98.8% for Furthermore, industrial field test carried out Gaoyang mine, filed demonstrate more than 95%. Hence, meets requirements can be used practices detecting