论文部分内容阅读
针对传统蚁群算法及模糊C-均值聚类算法在合成孔径雷达遥感图像分割中精度低下和收敛速度较慢的问题,该文提出了一种改进的自适应阈值的蚁群及模糊C-均值聚类算法,实现对复杂合成孔径雷达图像进行分割。针对不同的合成孔径雷达图像,首先利用最大类间方差法获取最优阈值,通过最优阈值干预避免蚁群算法陷入局部最优解;再将自适应阈值蚁群算法得到的聚类中心和聚类类别数输入模糊C-均值聚类算法中,最终实现图像分割。实验结果证明,该算法在时间和误分率上较传统方法有显著的改进。
Aiming at the low accuracy and slow convergence speed of traditional ant colony algorithm and fuzzy C-means clustering algorithm in synthetic aperture radar remote sensing image segmentation, this paper proposes an improved adaptive threshold ant colony and fuzzy C-means Clustering algorithm to achieve the complex synthetic aperture radar image segmentation. According to the different synthetic aperture radar images, the optimal threshold is obtained by the method of maximum inter-class variance firstly, and the ant colony algorithm is avoided to get into the local optimal solution through the optimal threshold intervention. Then, the clustering centers and poly Class number input fuzzy C-means clustering algorithm, the final image segmentation. Experimental results show that the proposed algorithm has significantly improved the time and error fraction ratio compared with the traditional method.