Research On The Detection Method For Positioning Accuracy Of 3D Phantom System Based On Machine Vision

It is an urgent problem in the field of radiotherapy to realize the detection of the positioning accuracy of 3D phantom.Although there are some detection schemes,they all have some problems,such as cumbersome detection process and difficult construction of detection equipment.Therefore,a simple detection method for the positioning accuracy of 3D phantom is needed.In order to quickly complete the detection of 3D phantom.Based on the above reasons,a detection method for positioning accuracy of 3D phantom system based on machine vision is proposed in this study.Specific research contents are as follows:Firstly,a scheme to detect the positioning accuracy of the 3D phantom is designed.The technical index of the detection method is put forward,the measurement process of the machine vision in the detection for positioning accuracy of the 3D phantom is described,the detection device is designed and built,and the software of the detection for positioning accuracy of the3 D phantom based on visual method is written.Secondly,the corner detection algorithm in the camera calibration process is optimized,and the automatic recognition of target points in binocular matching is completed.In order to solve the problem that the traditional corner detection algorithm makes errors or cannot be recognized when the rotation angle of the calibration board is too large,this paper processes the image through convolution filtering,subpixelates all the corners,and uses the minimum convex hull principle to complete the corner sorting.Finally,the same image is used to compare the corner detection effect of the algorithm in Open CV and the improved corner detection algorithm in this paper.Finally,the test of the detection method for positioning accuracy of 3D phantom system based on visual method and the uncertainty evaluation of measurement were completed.According to the planned movement path and the calculation method for evaluation parameters of positioning performance,the detection measurement for positioning accuracy of 3D phantom system based on monocular vision and binocular vision were carried out respectively.Based on the data of laser tracker,the experimental results show that the measurement accuracy of monocular vision method and binocular vision method can be consistent with the accuracy of laser tracker.Then,the uncertainty of monocular vision method and binocular vision method is analyzed.The combined uncertainty of monocular vision method is 0.0187 mm,and the combined uncertainty of binocular vision method is 0.0396 mm.The above experiments prove that machine vision can meet the requirements of three-dimensional water model positioning performance detection,but the monocular vision method is too long to meet the needs of daily detection in hospitals.The speed of optimized binocular vision measurement is faster.Binocular vision is more suitable for the rapid detection of 3D phantoms.The above experiments prove that this study provides an effective method for detecting the positioning accuracy of 3D phantoms and can complete the detection task of 3D phantoms.