Research On Toothbrush Detection And Location Technology Based On Machine Vision

The detection and localization of the toothbrush production process rely mainly on artificial or complicated mechanical structures now,which have the disadvantages of low efficiency and high cost.The use of machine vision instead of the traditional methods can improve the automation and intelligence of toothbrush manufacturing equipment,which is of great significance.The common quality defect such as toothbrush bristles leakage and burr after toothbrush grafting process was introduced.Aiming at the defect of bristles leakage,local threshold method was used to extract the target region,and the defect region was screened by Blob analysis;Aiming at the defect of bristles burr,the region of interest of the image is calculated by the gray projection method,and then the target region is obtained by the Sobel edge detection,finally the features such as area are used to identify defects.The experimental results prove the effectiveness of the two defect detection algorithms.The requirements of the toothbrush positioning and sorting in the production line are analyzed,and the hardware system is designed.Through the camera calibration and position calibration between vision system and conveyor system,the coordinate conversion relation between the image coordinate system and the conveyor belt coordinate system is established.The toothbrush image preprocessing and segmentation method was studied,and the holes region is filled based on the flood fill method.In view of the shape characteristics of the toothbrush,a method for calculating the center and orientation of the toothbrush based on the circumscribed rectangle and the scan line is proposed.The experimental results show that the positioning error of the toothbrush is less than 0.6mm and the orientation error is less than 0.2 ?.The method of using support vector machine(SVM)to classify the toothbrush's posture is studied and the toothbrush product feature database is established.According to the characteristics of different toothbrush varieties,the suitable feature combination is selected as the input feature vectors of the SVM classifier.The experimental results show that this method can achieve the accuracy of more than 95% for posture classification of most varieties of toothbrushes.In the realization of the above algorithm,based on the Qt Creator development platform,using the C++ programming language combined with Qt and OpenCV function library to develop a toothbrush detection and location software.The software adopts the principle of modular design to facilitate commissioning and maintenance.