基于改进的人工蚁群的图像分割算法

蚁群算法具有良好的离散型、并行性、正反馈性和鲁棒性,非常适合用于图像分割.基本蚁群算法存在着收敛速度慢、图像边缘的细节信息保留不完全等不足.改进了蚁群算法的启发信息,提高了蚁群算法的收敛速度,同时更有效地保留图像边缘的细节信息.图像的奇异值中只包含了少量的细节信息,大量细节信息体现在图像矩阵的2个正交矩阵中.通过利用奇异值分解作为启发信息,与信息素共同指导蚂蚁的行为.通过对蚂蚁行走路径上的信息素分布进行更新,使得分布在目标路径上的信息素逐渐增大,逐渐向分割图像收敛,根据信息素分布提取分割结果.仿真实验表明,对图像得到了理想的分割结果.

Image Segmentation Based on Improved Ant Colony Algorithm

The ant colony algorithm, with good discretion, parallel, robustness and positive feedback, is suitable for image segmentation. The basic ant colony algorithm has such shortages as slow convergence speed and incomplete image edge detail information. By improving the heuristic information on ant colony algorithm, the convergence speed of ant colony algorithm can be increased and the detail of the image edge information is preserved more effectively. The image of the singular value only contains a small amount of detail information, a lot of detail information embodied in two orthogonal matrix of image matrix. The core idea is to use the singular value as enlightening information, which guides the ant colony with the pheromone together. By updating the distribution of pheromones on the routes that the ants have passed, the pheromones on the target route are increased, and the searching routes will converge on the segmented image progressively. Finally, the segmentation results can be extracted according to the intensity of the pheromones. The simulations and experiment results show that the image can be segmented well by using this method.