On-tree fruit automatic identification method based on color information and three-dimensional contour information

• Main Authors: ZHU QIBING, GUO YA, HUANG MIN, WU GANG
• Format: Patent
• Language: Chinese; English
• Published: 09.08.2019
• Subjects: CALCULATING; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention discloses an on-tree fruit automatic identification method based on color information and three-dimensional contour information, relates to the technical field of machine vision, the method comprises the following steps: removing a background by using color information of three-dimensional point cloud data to obtain fruit point cloud data of a fruit area; clustering the fruit point cloud data to be the point cloud cluster through the Euclidean clustering algorithm fusing the distance difference and the color difference. The distance information and the three-dimensional contour information of the fruits are used for achieving one-by-one segmentation of all the fruits in the point cloud cluster in the three-dimensional space, the fruit recognition accuracy can be improved, andthe missing detection rate in the fruit detection process is reduced. 本发明公开了一种基于颜色信息和三维轮廓信息的树上水果自动识别方法，涉及机器视觉技术领域，该方法利用三维点云数据的颜色信息去除背景得到果实区域的水果点云数据，通过融合距离差异和颜色差异的欧式聚类算法聚类为点云团，在三维空间，利用距离信息和水果的三维轮廓信息实现对点云团中每个水果的逐一分割，可以提高水果识别的准