Fitting Algorithm for Overlapped Spherical Fruits Based on Geometry and Hough Transform

摘要

At present, the main algorithm used in feature extraction of spherical fruits is tradition Hough algorithm. The algorithm has the disadvantages of complicated operation and high rate of false recognition. In view of this situation, an object fitting method based on the effective edge information of image is proposed in this paper. Firstly, the image noise of the original image is filtered out by bilateral algorithm and the overlapped objectives are recognized by K-means algorithm. Forwards, the watershed algorithm is used to realize the segmentation of overlapped fruits. Then using Canny operator and 8-connected boundary tracking algorithm to obtain the pixel coordinates of outer contour of the single object. Finally, combining geometry with Hough transform to realize the contour reconstruction of overlapped spherical fruits. In order to verify the effectiveness and feasibility of the proposed algorithm, 30 overlapped fruit images contain 109 fruits were used to verify the algorithm in experiment. The results show that in the natural environment, the method can effectively recognize and reconstruct the actual contour of overlapped spherical fruits, and the success rate of recognition is up to 86.2%.