Research And Implementation Of 3D Reconstruction Technology Based On Binocular Vision

In recent years,with the development of computer vision technology,3D reconstruction has been widely used in many fields,such as aerial mapping,visual navigation,medical diagnosis and industrial testing.Binocular stereo vision is an important form of computer vision,based on two cameras,whose parameters are same or different,to get two images of the same 3D scene in a different perspective simulating the human visual mechanism,to restore the 3D scene depth information using parallax principles and mathematical methods.So it has some research value and broad application prospects to study for 3D reconstruction based on binocular stereo vision.This thesis mainly focuses on improving 3D reconstruction technology based on binocular stereo vision from the binocular stereo camera calibration,stereo matching,depth information acquisition and so on.The hardware experiment platform for the binocular stereo vision was set up through theoretical analysis firstly.Then the existing stereo calibration algorithm was improved to calculate the camera internal and external parameters.After getting binocular images,the improved stereo matching algorithm was used to obtain the disparity map.Then the 3D shape of the object was restored according to the depth information and the principle of the triangulation.In this thesis,the research mainly focuses on the following aspects.(1)The parallel binocular structure was chosen to be the research object and appropriate cameras and lens were selected to set up the hardware platform of binocular stereo vision.(2)For stereo camera calibration,an improved stereo calibration method based on epipolar rectification was proposed and the results were optimized using re-projection error.Firstly rectify the two images using camera parameters,then calculate the rectification error,finally optimize the parameters matrix between cameras using the error.The experiments show that it has higher precision using the improved calibration algorithm.(3)In the aspect of image feature matching,the improved stereo matching algorithm based on SURF feature descriptor was proposed.Establish the k-d tree structure using BBF method for nearest neighbor search.Finally adopt some strategies to eliminate the false matches.The experimental results proved that the improved SURF algorithm is of lesscomputing time and higher precision.(4)In the aspect of image region matching,the improved fast NCC matching algorithm based on edge feature and morphology was proposed.The experiments show that it has higher speed of calculating the disparity map by using the improved matching algorithm.(5)The experimental process of 3D reconstruction was designed based on binocular stereo vision.Then the algorithm was verified on MATLAB platform through several experiments and finally excellent reconstruction results were acquired.