Research On Quality Inspection Technology Of Twist Drill Based On Vision

Machine vision-based non-contact part size detection has been widely used in automated production due to its advantages of fast detection speed,high accuracy,and strong adaptability.In manufacturing,drilling is a very common and important step in the machining process.Among the drilling tools,the twist drill is often used as the first choice tool because of its low price and good drilling performance,so it is widely used.Due to the complex spiral geometry of the twist drill itself,the traditional twist drill quality inspection method is inefficient and cannot meet the needs of the automatic production of twist drills.Therefore,it is proposed to apply visual inspection technology to the quality inspection of twist drills.The research content of the thesis mainly involves the following aspects:Firstly,the overall structure of the bit quality inspection system based on vision was introduced,and the appropriate camera,lens and light source were selected according to the measurement requirements of the bit edge and side parameters.Considering the length of system development cycle and the difficulty of integrating with other control systems,LabVIEW virtual instrument and NI vision development kit were selected as the development platform of quality inspection system.Secondly,the image processing technology was studied.According to the different effects produced by different algorithms,the image processing algorithm most suitable for the twist bit quality inspection system was selected.In order to improve the system accuracy on the basis of saving the system development cost,this thesis used sub-pixel edge detection algorithm to locate the edge of the twist drill bit.Subsequently,the parameter detection method of the twist drill and the camera calibration process were introduced.Aiming at the problem that the angular position of the twist drill is not fixed during the measurement process,a ROI setting method based on geometric matching was proposed.Used the image calibration operator in the image function of NI Vision Assistant to complete the camera calibration process.After experimental analysis,this method can meet the system accuracy requirements.Finally,the modules of the twist drill quality inspection system software were introduced,and the measurement accuracy and accuracy of the twist drill quality inspection system were verified through experiments.The experimental results show that the maximum measurement error of the quality inspection system is 0.0037 mm,and the PTR(ratio of measurement accuracy to dimensional design tolerance)of each measurement item is less than 0.3.The measurement accuracy and accuracy of the twist drill quality inspection system meet the requirements,can meet the actual production needs.