Micro Vision-based Sharpening Quality Detection of Diamond Tools

Abstract

AbstractUltra-precision grinding is the last critical step in diamond tool machining. At present, most of the tool quality detection methods in the grinding process are offline measurement, which will reduce production efficiency. The general tool condition monitoring (TCM) method is designed for the use of the tool for processing and production, and there are relatively few studies on the detection in the process of producing diamond tools. Referring to the general TCM method, this paper adopts the machine vision detection method based on deep learning to realize the on-machine detection of the grinding quality of diamond tools. The method proposed in this paper is optimized for the recognition of small-sized defect targets of diamond tools, and the impact of less data on the training of the recognition network is improved. First, an imaging system is built on an ultra-precision grinder to obtain information suitable for detecting defective targets. These tool images are then used to train an optimized object recognition network, resulting in an object recognition network model. Finally, the performance of the network on the task of diamond tool defect recognition is verified by experiments. The experimental results show that the method proposed in this paper can achieve an average accuracy of 87.3% for diamond tool detection and a 6.0% improvement in position accuracy.KeywordsDiamond toolsVisual inspectionTool condition monitoring