Study On Container Recognition And Location Based On Binocular Stereo Vision

When transporting by containers, the degree of development of the loading and unloading operations of container transportation represents the productivity of the shipping industry, and fast and accurate lifting operations in the loading and unloading stage have become the key to the development of container transportation. In order to improve production efficiency, realizing of container lifting operations intelligence proves to be one of the urgent needs of the container shipping industry. At present, direction of container smartphone operating development is mainly based on computer vision to achieve the recognition and location of the container. In computer vision, binocular stereo vision simulating human eyes, has three-dimensional perception of space objects, and has been widely used.This thesis is mainly research the technology of container recognition and location based on binocular stereo vision. To convenient for the motion of slings for lifting loads, it is necessary of recognition for container and3D location to circle lock holes through binocular stereo calibration, image processing and segmentation technique. Generally speaking, the thesis mainly discusses the following contexts:(1)Basing on theory of camera image formation and Zhang plate calibration algorithm, stereo calibration for binocular cameras and achieving parameters of camera calibration by experiment.(2)Gray scale and median filtering for gathering color images, at the same time, taking advantage of Canny algorithm to extract edge of outline, obtaining precision outline information.(3)Aiming at the recognition of target container, template match algorithm basing on edge of outline information realizes setting target recognition in template; about recognition of circle lock holes of container, Hough alternation, which carries through space convert to extract information of circle, helps to get parameters information of circle.(4)Basing on3D reconstruction theory, leading into parameters of cameras and matching points of the centers of circle, it is available to achieve the3D coordinate of points, which means the location of the centers of circle lock holes.