The Research On 3D Reconstruction Technology Of Heart Soft Tissue Based On Binocular Stereo Vision

Currently,the depth of the lesion obtained by the doctor through binocular vision is dependent on the passive synthesis of the human eye in the process of using the surgical robot.Which means we can't provide 3D measurement data for the surgical robot system directly.This situation is limiting the development of surgical robots to intelligent direction.To solve this problem,this paper presents the Research on 3D reconstruction technology of heart soft tissue based on binocular stereo vision and the details are as follows.1.This paper constructs the imaging model of monocular endoscope which includs the aperture imaging model and the nonlinear geometric distortion correction model.Based on this,the stereo endoscopic vision system model is constructed.2.The joint calibration algorithm of stereo endoscope is designed based on the single-camera calibration algorithm based on the checkerboard plane template.First,complete the calibration of the left and right endoscopes.Then,get the external parameters of the stereo endoscope according to the conversion relationship between the two endoscopic coordinate system and the world coordinate system.Complete the calibration experiment of the stereo endoscope and get the re-projection error finally.The experimental results show that the calibration accuracy of the stereo calibration method is reliable.3.A stereo endoscopic image correction algorithm is designed based on the known stereoscopic endoscopic parameters.Which corrects the stereoscopic image by means of parameter reconstruction,back projection,rotation,de-distortion,re-projection and bilinear interpolation.The image correction algorithm is finally validated by using the template images and the Heart soft tissue images.The experimental results show that the vertical parallax between the left and the right images is eliminated basically.4.In order to avoid the defects of the stereo matching algorithm based on the joint similarity measure and the adaptive weight.The improved joint similarity measure and the improved adaptive weight are proposed.Which uses the color images for stereo matching directly and the matching accuracy is improved greatly.The stereo matching algorithm is finally validated by using the heart model images and the CT scan data which are provided by the British Imperial University.The experimental results show that the improved similarity measure and the improved adaptive weight all can improve the matching accuracy.The matching algorithm in this paper is batter than the algorithm uses a single similarity measure,and it has a good robustness to light distortion.5.3D reconstruction of the surface of heart soft tissue includes the 3D reconstruction of spatial points and surface reconstruction.Getting the 3D points according to the stereo endoscopic parameters and the disparity map bteween the heart soft tissue images.After filtering and points cloud streamlining,do the triangulation to the 3D points and achieve the surface reconstruction of the heart soft tissue.The experimental results show that the reconstructed 3D points is reliable and the reconstructed surface conforms to the contour information of the corresponding interest surface basically.