基于多链球型结构宇宙算法的番茄表面缺陷分割

Tomato surface defect segmentation based on multi-chain spheroidal structure universe algorithm

摘要:

为了改善番茄表面缺陷分割的效果，采用多链球型结构宇宙算法。首先建立多链球型宇宙拓扑结构，每个宇宙体只能、最多与周围的3个宇宙体进行信息交流，以保持该宇宙出入流平衡；通过最优阈值分割番茄缺陷图像，使用一个分离元素作为结束图像分割的停止条件；最后给出了算法流程。实验仿真显示，本文算法对番茄图像表面缺陷分割结果清晰，对各类缺陷的正确分割率比较高。

Abstract:

Multi-Chain Spheroidal Structure Universe Algorithm(MCSSUA) is proposed in order to improve the quality of tomato image surface defect segmentation. Firstly, multi chain spheroidal universe structure topology is established, and the universe can only communicate with the three universes around itself to maintain the balance of the universe. Secondly, the optimal threshold is adopted to segment the tomato defect image, and a separate element is taken as the stop condition for ending the image segmentation. Finally, the flow of the proposed algorithm is given. The simulation shows that the multi-chain spheroidal structure universe algorithm can segment the tomato image surface defect with clear results, and higher correct segmentation rate of various defects than other algorithms.