Investigation On The Technology Of Tool Wear Detection Based On Machine Vision

During the processing of aeronautic structure,the removal rate of material and the usage of tool are large.The geometrical parameters and the worn state of the cutting tool have important influence on the processing efficiency and machining precision.Real-time detection of the geometrical parameters could prevent wrong tool caused by tool changes frequently.Real-time monitoring of tool wear condition could prevent too much or too little wear mass loss caused by the experience of workers.The potential of tool life cannot be used effectively by the way of the workers’experience.Machine vision has the advantages of non-contact,simple and high precision in the tool detection,with overcoming the shortcomings of traditional methods.However,detection system based on machine vision still has some problems such as off-line detection and complicated processing algorithms.In this paper,the research of on-site tool detection based on machine vision is carried out.The developed background,the detection of tool geometric parameters and tool wear are expounded.The detection system of geometric parameters and tool wear based on the industrial camera is designed,built and verified.Meantime,the corresponding software of processing tool images matching the system is developed to realize automatic manufacturing and intelligent detection.In the study of system with detecting tool geometric parameters and tool wear condition,some solutions are proposed to design,build and adjust based on tool optical property,imaging environment and lighting system.The solutions include a combination of flat light backlit illumination and ring light forward illumination,a comparison between ordinary lens and bi-telecentric lens and an integration of bottom and flank edges by triple prism.On the MATLAB software platform,the algorithm of detecting the tool geometric parameters is developed.An algorithm that bounding rectangle of connected component is proposed.The results show diameter error was less than 0.01mm and hanging length error was less than 0.08mm.The method of image processing is used to detect the tool wear condition.A self-matching algorithm is proposed.Bottom edge is separated by method of connected domain labeling to analyze wear condition.Then separated edges are arranged regularly by location method of rotation to prepare for cutting edge extracted.The cutting edge is extracted by method of partial angle threshold to fit.Wear width is calculated with cutting edge compared with fitting edge.For the bottom wear,the results show the absolute values of errors in the width of maximum wear are less than 0.007 mm.In the case of severe wear and breakage,absolute values of errors in the width of maximum wear are less than 0.057 mm.For the flank wear,the absolute values of errors in the width of maximum wear are less than 0.03 mm.