遥感图像云自动检测  

Automatic cloud detection for remote sensing image

作　　者：郁文霞[1] 曹晓光[1] 徐琳[1] Mohamed Bencherkei 

机构地区：[1]北京航空航天大学图像处理中心,北京100083

出　　处：《仪器仪表学报》2006年第z3期2184-2186,共3页Chinese Journal of Scientific Instrument

摘　　要：云层遮挡是影响遥感图像质量的重要因素,为提高信道传输利用率,要求自动检测并剔除云覆盖率大的图像。目前大多数的云检测算法都依赖遥感图像的光谱特征,然而很多图像数据并不具备完整的光谱信息,因此这类算法对图像数据要求较高且算法不具普适性。本文从云的空间纹理特征和统计特征出发,提出适用于多卫星多传感器的云检测算法。经过对186幅图像的初步试验验证,该算法简单快速、准确性高,同时能有效控制云检测过程的检测错误率,达到很好的检测效果。Cloud shelter is an important factor which influences remote sensing image quality. To improve transmission efficiency of satellite data channel, it is necessary to automatically identify and eliminate images, which are highly-covered by cloud. Currently, most of the algorithms existed for cloud detecting are based on the spectrum characteristics of remote-sensing image, but some of the images do not possess intact spectrum information. As a result, this type of algorithms might have more requirements on original image data and could not be applied widely. A cloud detection algorithm is presented in this paper, which based on cloud texture and statistical features and applicable to multi-satellite and multi-sensor situation. Primary testing results of 186 images have showed that this algorithm is convenient, efficient and its running precision is high. Furthermore, the error rate of cloud detection could be effectively controlled so that perfect detection results might be achieved.

关 键 词：云检测 纹理特征 灰度共生矩阵 分形 

分 类 号：TH7-55[机械工程—仪器科学与技术]