基于层次反应型agent的可见光遥感图像阴影提取  

Shadow extraction based on layered reactive agents in remote sensing optical images

作　　者：陈克[1] 尹东[1] 马婧[1] 张开华 

机构地区：[1]中国科学技术大学电子工程与信息科学系,安徽合肥230027

出　　处：《中国科学技术大学学报》2009年第8期825-829,共5页JUSTC

摘　　要：根据可见光遥感图像中阴影区域的特点,提出一种新的利用层次反应型agent的阴影区域提取方法.层次反应型agent分为底层阴影搜索agent和高层管理agent两部分.底层分布式搜索agent感知图像局部信息,通过定居、发展、迁徙、标记和死亡行为搜索可见光遥感图像中的局部阴影区域;高层管理agent感知全局信息,对底层分布式搜索agent的搜索行为进行约束和指导.实验结果表明,该方法效率较高,并且可以有效地提取可见光遥感图像中的阴影区域.To extract the shadow region in the remote sensing optical images effectively, a novel approach, which used layered reactive agents according to the feature of shadow region, was represented. The layered reactive agents can be divided into two parts： shadow searching agents and monitoring agent. The searching agents detected the local shadow region by the behaviors such as inhabitation, breeding, diffusion, labeling and death in the local image environment; the monitor agent understands the global environment of the image, and then instructed the searching agents. The experimental results on remote sensing optical images show that the approach is rapid and efficient.

关 键 词：反应型agent 搜索agent 管理AGENT 阴影提取 可见光遥感图像 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]