Research On Fruit And Vegetable Picking Robot Vision Technology And The System Construction

Binocular stereo vision technology is an important branch of computer visiontechnology, along with the development of the market demand and technology, binocularstereo vision technology is more and more widely used in agricultural production. Thistopic main research mature binocular stereo vision system for greenhouse tomatoidentification and positioning, to obtain the maximum transverse diameter of a targetobject and the three dimensional position information, for picking operations to provideeffective information harvesting robot.This paper expounds the application of binocular stereo vision technology in pickingrobots at home and abroad, points out the problems existing in the research at home andabroad. Combined with project design goal to build appropriate binocular stereo visionsystem, the system hardware mainly include bumblebee binocular cameras,1394acquisition card,1394connections, and a PC computer.The technologies of computer vision and digital image process were applied torecognize ripe tomatoes. Pairs of stereo images were obtained stereo cameras, which wereRGB images. The chromatism value of ripe tomato was greatly different from itsbackground. A fixed threshold value was determined to separate the ripe fruit from itsbackground. Threshold segmentation method was used to remove noises from recognitionimages.The algorithm based on centroid Feature matching method was designed to get the3-D locations information of ripe tomatoes. The centroid feature of the target object wasextracted, and the principle of triangulation was used to obtain the depth value of the targetobject according to different disparities in stereo images. The error of depth is quite largewhen distance is more than l000mm, while the error of depth is ranged within±20mmwhen distance is less than1000mm...