Projection based Image Processing for Automatic Tool Wear Detection

Tool wear monitoring is a fundamental aspect in the evolution of modern machining operations. As the quality of the tool is directly related to the quality of the product, automatic tool wear evaluation is vital to ensure smooth automatic control of machining processes [1]. In this paper, an automatic tool wear detection procedure is presented, where real time collected sequence images of the cutting tool during turning tests are subjected to image processing. The procedure can be divided into four steps: (1) Standardization of tool images; (2) Adjustment of pixel color density to obtain clear edges of the tool wear region; (3) Drawing of the projection line to find out the coordinates of the edge lines; (4) Calculation of tool wear development characteristics. The tool wear characteristics and extension area are evaluated on the basis of the processed edge. One by one, sequence images are processed to automatically provide the growth of tool wear characteristics. In order to verify the procedure accuracy, the tool wear characteristics and extension area are also evaluated manually using a grid number counting technique on paper images. The comparison of the results confirmed that the automatic evaluations achieve a high degree of accuracy. The experimental results indicate that the proposed procedure can be a viable technique to assess the tool wear development characteristics on an on-line basis.