A New Apple Segmentation and Recognition Method Based on Modified Fuzzy C-means and Hough Transform

Abstract

ABSTRACT Apple segmentation and recognition are the important link in automation picking system. Rapidly and accurately recognition directly affects the real-time performance and reliability of picking robot. In order to realize effective segmentation and recognition, we propose a new apple segmentation and recognition method based modified fuzzy c-means and Hough transform. We use the non-local space information to construct image, according to the histogram of the image, it can automatically select the initializing cluster center. This process greatly improves the segmentation effectiveness. After the apple and background are segmented, the apple shape is extracted from the boundary of the apple pixel area. Then we adopt improved Hough transform to extract it. Finally, the experiments results show that this new method has high robustness stability and accuracy, and it can be used for quick apple recognition under complex background conditions.