BDTM-Net: A tool wear monitoring framework based on semantic segmentation module

Abstract

The integration of advanced manufacturing and the new generation of information technology promotes the development of intelligent manufacturing. In the cutting process, the condition of cutting tools is a critical factor that profoundly affects product surface quality and machining efficiency. Tool Condition Monitoring (TCM) can reduce the cost of processing and improve the quality of processing. It is one of the important technologies to realize intelligent manufacturing. To better identify the amount of tool wear in the cutting process, this research constructs a tool wear detection framework based on a semantic segmentation module. The semantic segmentation task of tool surface wear image collected in a complex environment is carried out by using image pixel information for tool wear monitoring. Because of the uneven illumination of the edge of the wear area and the unclear edge boundary, the self-learning parameters are used to separate the foreground and background of the image and amplify the subtle difference information. While enhancing the feature information of the tool wear image, the detection efficiency is improved. At the same time, to meet the needs of detail segmentation, a dual attention module is introduced to improve the performance of the model. The accuracy of the model is verified by orthogonal experiments and the model is comprehensively compared based on common evaluation indicators. The accuracy rate of 95.34 % in segmenting the tool wear images, demonstrating that the developed detection framework is suitable for accurate and efficient tool wear condition monitoring. This research not only proposes a new semantic segmentation model but also provides valuable insights into key information during the cutting process, validates the patterns of tool wear, and reasonably promotes the development of Tool Condition Monitoring and Remaining Useful Life.