Recognition and Location of Persimmons Based on K-Means and Epipolar Constraint SIFT Matching Algorithm

摘要

Aiming at the requirements of intelligent persimmon picking operations, this paper designs a recognition and positioning system for persimmon picking robots. The system uses the H component under the HSV model to preprocess the image, uses K-means clustering algorithm for segmentation, and successfully extracts fruit targets after subsequent morphology and desiccation processing. In order to extract the three-dimensional coordinates of the fruit, the images obtained by the left and right cameras after calibration are matched using the SIFT algorithm. Finally, the distance is solved by the similar triangle principle. After testing, the sample image cluster center is 2, the number of iterations is 200, and the working range is 200mm ~ 800mm. The recognition success rate is about 90%, and the positioning error is ^±15mm, which can fully meet the needs of picking robot operations.