多智能体遗传二维Otsu法SAR图像变化检测

Change detection in SAR images using multi-agent genetic and two-dimensional Otsu algorithm

提出一种基于多智能体遗传二维最大类间方差法的合成孔径雷达图像变化检测方法。采用阈值分割的思想,利用对数比值法构造差异影像;通过多智能体遗算法寻找变化和非变化类之间距离测度函数最大的全局阈值,得到变化检测结果。实验结果表明,与遗传算法、免疫克隆选择算法、多智能体遗传一维最大类间方差法、二维最大类间方差法相比,该算法可以快速、准确地得到变化检测结果。

A change detection method for synthetic aperture radar （SAR） images based on the multi-agent genetic and two-di-mensional Otsu algorithm was proposed .The method was based on the threshold segmentation idea .The difference image was constructed by using the logarithmic ratio ,and the multi-agent genetic algorithm was used to find the optimal threshold of the largest distance measurement function between the changed and unchanged .Then the change detection results were produced by using the optimal threshold .The experimental results show that the new algorithm can get the change detection results quickly and accurately compared with the genetic algorithm （GA） ,the immune clonal selection algorithm （ICSA） ,the multi-agent ge-netic and one-dimensional Otsu algorithm and the two-dimensional Otsu algorithm .