METHOD FOR HARVESTING HIDDEN VEGETABLES AND FRUITS ON BASIS OF DEPTH ASSOCIATION PERCEPTION ALGORITHM

摘要

A method for harvesting hidden vegetables and fruits on the basis of a depth association perception algorithm, comprising: acquiring RGB channel color images, a D channel depth image and an I channel infrared image; normalizing the resolution of the acquired images; removing an area in an image after resolution normalization that is outside of the working area of a harvesting robot; identifying a matured target, and labeling in a cropped image each pixel of the matured target; determining whether the matured target is hidden; when hidden, restoring the images in the RGBD channels for the whole matured target; when not hidden, directly acquiring the images in the RGBD channels for the whole matured target; and according to the images in the RGBD channels for the whole matured target, extracting the place and posture of the matured target to the harvesting robot. The present method experiences less interference from the environment and works more stably; moreover, full RGBD information of a target may be estimated, and the success rate of robot harvesting may be increased. Further disclosed is a corresponding harvesting system.