黄瓜采摘机器人的研究：基于图像的黄瓜果实目标识别

For the cucumber harvesting robot , the degree of maturation and recognition of the position in space are the important steps . First of all, segmentation of cucumber image is required. The color of cucumber is the same with its peduncle and leave, but they are different in reflectivity and containing of water, so the depth of their color are recognizable. Based on the difference, abstract the G channel information from the image which cucumber in the nature background in the color model of RGB and binarize the gray image, separate the cucumber from its background. Then, through erosion and dilatation to remove the little and isolated noises, and do area-sign to the image, after that separated the cucumber exactly from image by feature abstract.