Novel monitoring method for belt wear state based on machine vision and image processing under grinding parameter variation

The wear state of an abrasive belt is one of the important factors affecting the grinding precision of belt grinding processes. At present, there are two problems associated with the monitoring method of the wear condition of abrasive belts: (1) there are no uniform wear criteria to classify the wear condition of abrasive belts, and the segmentation threshold of the wear condition is affected by the change in the grinding parameters; and (2) an abrasive belt wear model based on indirect sensor monitoring of signals is affected by the change in the grinding parameters of abrasive belts; therefore, it is only suitable for abrasive belt wear monitoring under specific grinding parameters. This paper introduces a method of belt wear state monitoring based on machine vision and image processing. Surface images of an abrasive belt during its entire life are captured using a noncontact electron microscope. Three image features related to the wear state are selected: first-order distance of color component R, entropy of the horizontal subgraph, and vertical subgraph of the texture feature. Moreover, the wear state is classified into three categories based on the selected features. Using the selected features and the random forest classification algorithm, an abrasive belt wear state classifier is established. The performance of the classifier is verified and evaluated using a data subset of different images. The results show that the proposed method has high recognition accuracy for belt wear state, which can reach 99% in the accelerated wear stage. The proposed method solves the problem of the dependence and sensitivity of the monitoring model on the variation in the grinding parameters in the process of abrasive belt wear monitoring, and it improves the adaptability and versatility of the monitoring method.