A method for determining longitudinal tear of conveyor belt based on feature fusion

Conveyor belt is often damaged by longitudinal rip while working in transport, which can lead to the scrapping of expensive machines, and sometimes causing fire and casualties. This paper presents a method for determining longitudinal tear of conveyor belt based on feature fusion. We analyze the main reasons of longitudinal rip and the image characteristics when the belt is ripped apart correspond with real operational conditions. Then the geometric characteristic and template matching feature based on gray image are selected to describe the characteristics of longitudinal tearing crack of conveyor belt. Finally, longitudinal tear of conveyor belt is determined based on DS evidence theory. The detection system based on computer vision is designed, which includes information collection apparatus, the image detecting apparatus, PC operating system and data transmission interface and the software is programmed by using MATLAB graphical user interface (GUI) technique. The presented method is useful for avoiding the risks of belt longitudinal rip under real operational conditions and for improving the production of transport’s efficiency and effectiveness.