3D Reconstruction Of Soft Tissue Image Based On Binocular Endoscope

With the development of medical endoscopic equipment,Minimally invasive surgery(MIS)has gradually become an important technical means in daily medical practice.Minimally invasive surgery has been widely used in recent years due to small wounds,slight pain,and rapid postoperative recovery.However,while benefiting patients,minimally invasive surgery puts forward higher requirements for surgical operators.Minimally invasive surgery is a multi-person indirect eye-hand collaboration process using an endoscope as a medium.The limited field of view of the endoscope and the lack of direct depth perception bring difficulties to the doctor's intraoperative positioning,which greatly affects the accuracy and safety of the surgical operation.In response to these problems,this paper has conducted an in-depth study on the three-dimensional reconstruction method of soft tissue images based on binocular endoscope.The main research work is as follows:1)The binocular stereoscopic endoscopy imaging model is studied.The stereoscopic endoscope was calibrated using Zhang Zhengyou's camera calibration method.Based on the calibration results,the stereo image pairs were corrected,the feature matching algorithm was researched,and the soft tissue 3D reconstruction based on stereo vision was realized.2)An improved shape-from-shading(SFS)3D reconstruction algorithm is proposed.The Classical algorithm used Lambertian reflection model and Blinn-Phong reflection model and other empirical illumination reflection models are studied.In order to solve the problem of the limited expressiveness of the empirical illumination reflection model in the real MIS environment,this paper uses an irradiance theory and micro surface theory-based reflection model used for SFS reconstruction.The comparison experiment with the classical light reflection model shows that the light emission model in this paper can more effectively reconstruct the three-dimensional information of the target surface.3)A 3D reconstruction framework combining stereo vision and SFS is proposed.based on the improved SFS and classical stereo vision matching methods,a joint objective function is constructed,and a depth map optimization algorithm under the joint objective function is given.Finally,the joint reconstruction algorithm was experimentally verified.Based on the silicone heart model image data set and the real heart image data set,qualitative and quantitative comparison and analysis were carried out to verify the effectiveness of the joint reconstruction framework.Experimental results show that the joint reconstruction framework can reconstrut the shape of the heart surface while reconstruct local details.Compared with the classical stereo vision and SFS methods,the joint reconstruction method proposed in this paper has better robustness and reconstruction accuracy.