The Quality Evaluation Of 3D Printing Medical Guide Combining 3D Model And Computer Vision

With the rapid development of 3D printing technology in recent years, 3D printing products have been utilized in various applications, such as medicine, industry and so on. Maxillofacial tumors medical guide plate made by 3D printer is an important auxiliary equipment of tumor radiation therapy. However there is not a reasonable product inspection solution for 3D printing products because of the personality of such products. And the general sampling inspection solution is not suitable to them. Therefore, product inspection is one of important researches in 3D printing.In this thesis, we mainly study the product inspection scheme to 3D printing maxillofacial tumor medical guide by combining 3D model and machine vision. Based on the scheme, we design the prototype machine. The main work is as follows in details:1. We propose a product inspection scheme to 3D printing maxillofacial tumor medical guide by combining the 3D model, computer vision and mechanical system. The position and direction of guide hole are the factor which could influence the radiation therapy effect. Deal with high personalized characteristics of 3D printing products and there is not unified specification, therefore, 3D printing product inspection is to detect the consistence between the printed product and the 3D model. We first make the alignment between the 3D model and the printed medical guide. Then we design the inspection scheme based on the 3D model. And then the guide is moved to the suitable by the mechanical system. And the quality of the guide hole is evaluated by image analysis.2. We design a 3D model initialization method based on pseudo binocular stereo vision from the single camera. 3D model initialization is to align the 3D model and the printed guide. Under the pseudo binocular stereo vision system, we could compute the coordinates of the mark points in the world coordinates. Furthermore, the transformation matrix could be resolved from the coordinates of the marked points. Using this idea, the 3D model and the printed guide could be put into the same coordinates. The method make 3D measurement from single camera with the supporting of accuracy mechanical system. It is simple, practical and the precision is satisfying.3. We put forward a cross section of guide hole detection algorithm based on the geometric feature filter and clustering. We first vertically project3 D model onto a plane and transform the plane detection in 3D space into region detection in the 2D plane. Then we detect the cross section by image segmentation. And we obtain the center of the guide hole using Hough transformation.4. We achieve the guide hole quality inspection based on image features. We first extract the contour of the guide hole using the image procession algorithms such as edge detection, binarization and so on. Then we evaluation the hole based on the geometric features such as area, circularity.The testing on the prototype device show that the precision of maxillofacial tumor medical guide made by 3D printer quality testing and trying to puncture system based on the 3D model and machine vision is enough to finish quality detection.