A novel vision sensor for longitudinal tear detection of conveyor belt

In view of problem of unclear image of existing longitudinal tear detection of conveyor belt, a novel vision sensor based on infrared and visible image fusion was proposed. In the proposed sensor, coaxial light from the same lens is projected to infrared CCD and visible CCD respectively by use of beam splitter prism. Then, infrared image and visible image are collected by the infrared CCD and the visible CCD respectively at the same time. Finally, fusion image is obtained by use of pixel-level fusion method. The experimental results show that the fusion image collected by the proposed sensor can meet requirement of high-quality collection of conveyor belt image with clear tear information and detailed background information.