In-line monitoring of belt transport with adaptive bandwidth Mean-Shift hazard

Abstract

For the disadvantage of long online monitoring processing time of hazards of coal transportation belt infrared image, the adaptive bandwidth Mean-Shift monitoring is proposed. To establish an infrared spectroscopic imaging hazard model, according to the difference of hazards and the coal reflective background radiation in the best bands, hazards image be extract. Using the automatic bandwidth selection method based on backward tracking and object centroid registration, gray histogram is established only for hazards within the kernel bandwidth and tracking it. Experiments show that this method can effectively identify and track hazards and reduce the processing time compared to the conventional image for the hazard line monitoring with strong Real-Time.