On-Position Tool Wear Detection Based On Machine Vision

Tool wear is one of the key factors affecting the quality and machining efficiency. With the development of science and technology, on-process tool wear detection has become more important, in order to meet the higher requirement of the surface quality and economy. In comparison with the traditional methods, machine vision technology is a simple and high precision quick solution for the tool wear detection, without contact, and distortion.Based on its evolution, this paper summarized the advantages and disadvantages of the various tool wear detection technologies. The following exploratory works have been done in order to solve the problems exist in the current research of tool wear detection technology:The tool wear characteristics and morphology have been analyzed and the wear mechanism of conventional and micro-milling tool has been investigated. Besides, it has been determined that the influence of tool geometry on the processing quality. It has been proposed, that the tool wear area, tool diameter, rake angle and free angle should be employed as abrasion index for judging.A new hardware configuration has been proposed for the detection system hardware configuration and the frame work of software has been established for various function modules. The principle of tool wear detection system has been analyzed.In this paper, an image preprocessing techniques of tool wear detection has been employed and the adaptive median filter tool is used to denoise the image. And then by binarized,the image was then used the Canny edge detection technique of pixel-level to extract the rough edge.In addition, in order to improve the measurement accuracy of the tool geometry, this paper proposed the sub-pixel edge detection algorithm.We use the sub-pixel corner extraction algorithm to realize the sub-pixel location of the Canny pixel-level edge. It has been proved, that the theoretical error is up to ±0.5μm when using the sub-pixel corner algorithm. During the research, gauge blocks were used as an specimen and the true measurement error can be controlled within± 0.5m. All of the testing results provide the technical foundation for micro milling.Finally, with the combination of the robot and the optical measurement system, the testing platform can simultaneously determine the abrasion on the tool bottom and the side of the drilling tool.Also the milling tests were carried out and the testing system has been verified. With the detection system and its related morphology, various tool geometries and abrasion area can be determined, including the rake angle, the free angle and tool diameter. The research mentioned in this paper provides a way to further improve the form accuracy and surface quality of high-performance precision machining.