The Research On Key Technology Of 3D Body Reconstruction Based On The Virtual Positioning Of Patients

In the process of heavy ion radiotherapy,the patient needs to be positioned several times during the fractionation process with the guidance of CT images,which takes a long time and has a high radiation dose.In order to reduce the radiation dose to the patient and improve the positioning efficiency of the fractionated treatment,this study provides a technical guarantee for the improvement of the image guided accurate positioning efficiency by performing 3D scan reconstruction of the human body before the operation,using the model as a positioning reference,and achieving virtual positioning through model comparison and position adjustment during the fractionated treatment,ensuring the minimum change of the position between the human body and the bed plate before and after the fractionated treatment and positioning.The technology guarantees to improve the efficiency of accurate positioning of image guidance.How to achieve 3D human reconstruction through data acquisition,point cloud calculation and point cloud alignment is the focus of this paper.The main research contents and results of this paper are as follows.(1)In view of the complexity of human body point cloud data,based on the Microsoft Kinect camera,such as easy operation,low cost and development,this paper proposes a method to obtain human body point cloud data based on Kinect "six directions" method.Before data acquisition,its imaging model and the mapping relationship between various coordinate systems are analyzed,and finally the camera calibration is performed to obtain the focal length,principal point and radial and tangential aberration parameters of the camera and the average error of image corner points,respectively,and the tangential aberration parameters are analyzed to meet the experimental requirements of 3D reconstruction.(2)For the point cloud data acquisition process,there are non-human outside the field point cloud data and noise problems,a straight-through filtering algorithm based on the improvement of the statistical filter to build a two-pass filtering denoising algorithm.The straight-through filter is mainly used to remove the redundant point clouds in the xyz direction outside the human body,and then the statistical filter is used to denoise the human point cloud data to complete the redundant data removal,which improves the denoising effect and saves time.(3)The alignment study of the pre-processed human point clouds was carried out according to the alignment process of stepwise alignment proposed in this paper.Through the experimental data analysis,the coarse alignment is mainly based on the SAC-IA dependent fast feature descriptor FPFH coarse alignment algorithm,and combined with the extraction of key points to improve the best initial position between the point cloud pairs;the complete alignment is mainly based on the combination of k-d tree improved ICP alignment and fast feature descriptor FPFH algorithm to minimize the alignment error between the point cloud pairs and complete the final Human point cloud alignment.(4)The Poisson surface-based reconstruction and triangular greedy meshing reconstruction algorithms are analyzed.For the characteristics of human point cloud data in this paper,the experimental results are given by using the improved triangular greedy meshing reconstruction algorithm based on the removal of redundant points from human alignment point clouds,uniform point cloud data and point cloud smoothing,respectively.(5)In order to verify the effectiveness of virtual positioning technology,this paper adopts the method of soft plastic material system combined with body surface scribing for patient position fixation,independently builds the experimental platform based on medical flexible robotic arm and carbon fiber bed plate,develops the test software,and verifies the virtual positioning effect.In this paper,we focus on the key technology of 3D body reconstruction before radiotherapy to improve the positional efficiency and reduce the positional errors,in order to address the problems of accurate patient positioning during heavy ion radiotherapy.This paper has implications for the engineering applications of reverse engineering technology in the medical field.