基于空间约束和改进果蝇算法融合的图像分割

Image Segmentation Based on Spatial Constraint and Improved Fruit Fly Algorithm

针对传统FCM算法在图像分割应用中对图像噪声的敏感性问题,提出一种基于空间约束和改进果蝇算法的FCM图像分割算法。该算法在原FCM算法的目标函数中加入一个包含空间邻域信息的约束项,使得整体上相邻像素点趋于同一类时,目标函数最小;并利用改进的果蝇算法选择模糊均值聚类算法的最优初始聚类中心,实现图像分割。最后采用仿真实验测试算法的性能。实验结果表明,相对于传统FCM算法,本文算法在分割正确率、分割速度以及鲁棒性上均更优,具有更广的应用范围。

For the problem of noise sensitiveness when applying conventional fuzzy c-means （ FCM ） clustering algorithm in noise image segmentation, a new FCM image segmentation algorithm based on spatial constraint and improved fruit fly algorithm is proposed. A constraint term which contains the spatial information is added to the target function of original FCM algorithm, which makes the ad- jacent pixels tending to the same class on the whole and minimizes the objective function; and selects the center of FCM with improved fruit fly algorithm to realize image segmentation. Finally it uses simulation experiment to test the performance of the algorithm. The ex- perimental resuhs show that compared with the traditional FCM algorithm, the proposed algorithm has better performance in segmenta- tion accuracy, speed and robustness, and has a wider application range.