Research On 3D Point Cloud Reconstruction Technology About Potato With Binocular Vision Based On OpenCV

In recent years, machine vision technology has developed rapidly,and gradually become an independent discipline. As computer technology and digital signal technology become more mature, machine vision can use the camera to read the environment, and convert it into digital signals. With the gradual deepening of the study of human eye imaging model, binocular stereoscopic vision technology can be the closest to the realization of machine vision to human visual bionic. So it became one of the popular research contents in machine vision.This paper mainly studies the following aspects:(1) Firstly, the research on the imaging principle of the camera model, the computer coordinate system, the image coordinate system, the camera coordinate system and the world coordinate system are studied. Using the transformation relationship between the above coordinate systems, a binocular imaging model was obtained.(2) Then, the binocular camera calibration was studied. In order to improve the accuracy of calibration, when we were experimenting, the first single standard set, and then binocular calibration. In this experiment, the camera was calibrated using a calibration plate. The experiment results show that the calibration is very accurate.(3 ) In order to improve the matching efficiency, the image was corrected before the stereo matching. Theoretically, three matrix transformations (perspective transform, affine transformation, and misalignment) can be used to complete the correction of stereoscopic images.But since the rotation matrix and the translation vector of the camera were obtained during the calibration process. So we use the Bouguet algorithm for stereo correction. The experiment results show that the method has high accuracy regardless of the calibration plate or the correction of the physical map (potato).(4) In this paper, we used the method of region matching in the process of matching, and study the principle of two classical stereo matching algorithms (BM algorithm, SGBM algorithm).Through several experiments, the SGBM algorithm is used to get the best depth image.(5) At the end of this paper,we use the depth map and the re - projection matrix to find the three - dimensional point cloud coordinates of the potato in the world coordinate system. The experiment lost 340,000 three-dimensional coordinates of the coordinates. In order to image the three-dimensional points displayed in the graphics, this article used OpenGL to draw point cloud images. The experiment results show that the 3D point cloud has a good three-dimensional effect and authenticity.