Feature Matching Based Conveyor Belt Anomaly Detection

Abstract

In the field of industrial automation, computer vision is becoming more and more popular. At present, the traditional conveyor belt anomaly detection will consume a huge amount of manpower and material resources, this paper proposes a way that can effectively reduce manpower and improve the correct rate. The algorithm first adopts multiple cameras, which are used as the acquisition medium for feature extraction, and the collected images on the conveyor belt, image processing and feature extraction techniques to find the unique features on the conveyor belt for preservation, and through these unique features on the conveyor belt, carry out the subsequent feature comparison, and if the subsequent detection of the existence of images with similar features, the algorithm will carry out the comparison, and if it exceeds the threshold, it will carry out the abnormality alarm. Experiments have proved that the algorithm has good performance in conveyor belt anomaly detection.