Research On 3D Modeling Method Of Human Lumen Based On PTAM

The three-dimensional modeling method of human lumen based on PTAM(Parallel Tracking And Mapping)has the characteristics of simple system structure and high real-time performance.It can be applied to the three-dimensional measurement of partial abdominal cavity and cranial cavity surgery,and plays an important role in the three-dimensional reconstruction of human chest and abdomen.Therefore,To this end,the paper focuses on the PTAM human lumen threedimensional modeling method,mainly for the analysis and research of PTAM system construction and error analysis,lumen image preprocessing,feature point detection and matching design.First,the paper analyzed the corresponding basic imaging model of the camera used in PTAM 3D reconstruction technology.The mathematical model of PTAM system is established,and the internal relationship between the structural parameters and the measured object in the system model is analyzed.Combined with the characteristics of the internal cavity environment,the PTAM three-dimensional reconstruction system is constructed.According to the 3D reconstruction method designed in this paper,the camera calibration method is researched and the camera calibration process is completed.Secondly,aiming at the complexity and particularity of the internal cavity environment,the paper detects the specular reflection area existing in the cavity image and preprocesses the image.The Harris algorithm,FAST algorithm and SURF algorithm in image feature point detection and matching are deeply studied.Design simulation experiments verify the advantages and disadvantages of the above algorithms.Aiming at the characteristics of luminal vessels,this paper designs a feature point detection method based on vascular directional gradient.This method can detect vascular feature information better.The method and FAST algorithm are applied to luminal feature point detection,and use the SURF algorithm performs feature point matching.The experimental results show that the combined algorithmsatisfies the requirements of real-time and stability of the system.Finally,the paper designs a three-dimensional modeling method of human lumen based on PTAM,and carries out practical experiments on the camera calibration method and feature point detection and matching method designed in this paper.The calibration process of the camera in the PTAM-based 3D reconstruction method is completed.The PTAM feature tracking and map construction process is realized.After the map is initialized,the feature point detection algorithm based on the vascular direction gradient and the FAST algorithm are used to detect the intraluminal image feature points.The feature point matching is completed by the SURF algorithm;the internal cavity three-dimensional map model is constructed by continuously adding the feature points in the new key frame to the initialization map.The experimental results present the three-dimensional map information in the inner cavity environment,which verifies the prospect of the PTAM-based 3D modeling method proposed in this paper.