新型云区域检测算法  

New cloud detection approach of remote sensing image

作　　者：徐慧[1,2] 梁栋[1,2] 夏云[1,2] 杭丹萍[1,2] 

机构地区：[1]安徽大学计算智能与信号处理教育部重点实验室,安徽合肥230039 [2]安徽大学电子科学与技术学院,安徽合肥230039

出　　处：《计算机工程与设计》2011年第11期3764-3767,共4页Computer Engineering and Design

基　　金：国家自然科学基金项目(60772121);安徽省教育厅重点科研计划资助基金项目(KJ2010A021);安徽大学‘211工程’学术创新团队基金项目

摘　　要：为了更加精确地检测出遥感图像中云区域的边界及细节信息,提出了将最小交叉熵和形态学相结合的方法来对遥感图像进行云区域检测。从遥感图像的灰度特征出发,通过最小交叉熵准则选取最优的阈值来检测图像中的云区域,再通过形态学的开运算,消除与云区域不相连或者被误判的小的光亮的地物信息,最后在彩色遥感图像上勾勒出云区域的边界。实验结果表明,该算法简单快速,能够很好地区分出云区域和下垫面,并且能够准确地对云区域边界细节信息做出判断。In order to detect the cloud region boundaries and details of remote sensing image more exactly, minimum cross-entropy combined with morphology is applied for cloud detection of remote sensing image. Firstly, based on the grayscale feature of the remote sensing image, through the optimal threshold value which is selected by the smallest cross-entropy criteria to detect cloud area, and then the morphology opening operation is used to exclude smaller bright pixels of land cover not connected to the cloud area, which are mis- taken as the cloud pixels. At last, the boundary of the cloud region on the colored remote sensing image is outlined. The experimental results show that this method can judge detailed information of the cloud area and distinguish the cloud area from the underlying surface clearly.

关 键 词：遥感图像 云区域检测 阈值法 最小交叉熵 形态学 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]