Key Technology Research Of Tool Condition Detection On-machine Based On Machine Vision

Automated production process need to change tool automatically, when the tool shouldbe changed and if it is the correct size are determined by monitoring tool conditionon-machine. However, monitoring tool condition on-machine is still depend on indirectmethods, such as cutting forces monitory, acoustic emission monitory, spindle current/power monitory and so on. It is difficult to set accurate models with tool shape/size and toolwear condition. It is a viable technology method to acquire tool condition through processtool image on-machine between gaps in production process, This method can achieveproduction and monitoring integration. In this context, the key technology research of toolcondition detection on-machine based on machine vision is represented in this paper. Analyzethe significance and constraints of tool condition detection on-machine, take drilling andmilling tool as the object of the research, take CNC machining center space as working space,take intelligent, integration and oriented automated production process as the goal. Design antool condition detect system which is multi-faceted and cross-scale system. In order to matchwith the detect system, design a software to process tool image of real-time.According the optical properties of tool wear, the surroundings inside machine tool, andthe requirements of detected object, combined with flat surface light, fiber optic light, ringlight, and image fusion technology, an illuminating system is designed. And design an deviceto installation and adjust light source.After calculation, select the appropriate industrial camera, lens, and camera then circle,build a machine vision device to acquire macro-scale contour image and microscopic toolwear image. And design a electromechanical device for installing machine vision sensor, it islight, flexible, and convenient to be fixed, and can acquire macro and micro images of tool inthe limited machine space.The image process algorithm is specially researched. Because of much interferenceinside machine tool, the quality of the images acquired on-machine is poorer than imagesacquired in laboratory. Therefore, the image process algorithm is specially researchedaccording the characteristics of images acquired on-machine. In this paper, the diameter oftool is calculated according to the extracted curve contour of tool; the edges are enhanced byimage enhance technology; on account of accuracy and real time requirement, the sub pixelcontour of tool is extracted by the Lanser-facet operator for the calculation of tool diameter;to detect the wear of rake face and back face, the contrast of image is increase by contrast stretch technology; according to different images, the tool wear region is extracted byimproved two-dimensional Otsu operator and region growing respectively, the area of wearregion is calculated.Based on the above research, a software to measure diameter and wear of cutting toolson-machine is developed in Labview platform with Halcon and Matlab image processingfunction library. The experiments shows the error of measuring tool diameter is in the rangeof3%, the error of detecting tool wear is about7%. The system can basically meet the needsof industrial applications.