Research on Cutting Tool Wear Condition Monitoring Based on Workpiece Surface Image

It is very important to improve the automation machining efficiency and quality through the monitoring technology of cutting tool. And cutting tool condition monitoring based on image processing has great advantages which traditional monitoring methods can not provide. In this thesis, the turning is regarded as the research object, the internal relation between the tool wear and the txture images on the workpiece surfaces is analyzed, the key tool wear condition monitoring technology based on the workpiece surfaces images is researched by the method of image processing, improved Hough transform, fractal theory and HMM. These have great important to guide the tool wear condition Monitoring.Based on the machined surface images of tool wear monitoring, the experimental system is built,and based on the workpiece surface image in this thesis, the usual pre-processing method of image, was researched, analyzed and collated. Then the preprocessing method adapted to workpiece surface image was discovered. These lay the foundation for realizing the image characteristic extraction about the tool wear condition monitoring.In view of traditional Hough transform’s shortcomings, an improved Hough transform algorithm is proposed in this thesis, then the improved algorithm is used to process images of workpiece surface. The average length of line segments and the average angle between line segments and direction of cutting velocity which attained from the workpiece surface are used to be feature parameters;The high-order fractal feature and the multifractal analysis algorithm is employed to analyze the workpiece surface, a new fractal feature-average lacunarityη,multifractal entropy Hm are defined as feature parameters in this thesis. The experimental results show that the closer correlations exist between the tool wear condition and the four feature parameters; the tool wear condition monitoring can be realized based on the variation law of the feature parameters.Based on the characteristic parameters of machined surface image, the hidden Markov model of tool wear condition monitoring is built in this thesis. The tool wear HMM’s training and recognition are completed, and the specific implementation steps are given. The experimental results indicate that probabilities of unknown observation series can reflect the statistic similarities of observation series in different tool wear status and track the development trend of tool wear. Use threshold can identify the tool wear condition effectively.