A Real-Time Pineapple Matching System Based on Speeded-Up Robust Features

Abstract

Real-time and accurate image matching is a key to stereo vision of agricultural harvesting robots. This research is part of a pineapple harvesting robot project. In this paper, pineapples were selected as research objects and low cost binocular vision platform was constructed, fruit area of left image was got by real-time image acquisition and rapid segmentation, rapid matching of fruit area in left image with the integral right image was done by employing SURF (Speeded-Up Robust Features) algorithm. The algorithm was programmed and tested in VC++ software and the results showed that, the segmentation cost 0.017s and the real-time matching performed well. Compared with other agricultural harvesting robots' research, area matching and real-time matching were well-realized.