Conveyor belt can enhance transportation efficiency and reduce labor intensity, which is extensively applied in material transportation. However, the deviation of conveyor belt commonly leads to malfunctions in the transmission system, affecting routine transportation function. Therefore, it is crucial to accurately identify whether the conveyor belt deviates and rectify anomalies promptly. Previously, the on-site camera have to be arranged in the appropriate position to ensure conducive recognition results for deviation discrimination, which is probably constrained by environmental conditions during implementation. In this paper, a combined method of detection and segmentation with deblurring function (CDSwD) is presented for conveyor belt deviation discrimination, which can avoid the impact of the camera position on the deviation discrimination results by combining roller information and conveyor belt contour information. Initially, the deblurring approach is adopted to remove the motion blur from the images. Subsequently, the object detection technology and the semantic segmentation technique are utilized to generate the detection results along the lower line of the rollers and the segmentation results of the overall conveyor belt contour. Finally, the outcomes will be combined for conveyor belt deviation discrimination. Comparisons have demonstrated that the proposed method can accurately determine the deviation of the conveyor belt, which is beneficial for maintaining the normal transportation and improving work efficiency.
Read full abstract