面向对象和GURLS结合的高空间分辨率遥感数据云检测

遥感信息获取过程中云是重要的干扰因素,随着国产高空间分辨率卫星数据的应用,实现数据的准确云检测对有效获取地面信息具有重要意义。本文以高分一号、高分二号多光谱影像为数据源,利用图像分割获取了同质对象,基于对象光谱、纹理和几何8种属性特征建立了规则集,以规则集为输入,利用阈值法和GURLS分类器结合进行了云检测。针对不同时相和场景的高分数据,将该方法与基于像素的最大似然法和SVM法进行了对比,结果表明该方法云提取精度均在95%以上,Kappa系数在0.9以上。

Cloud detection method of high spatial resolution remote sensing data combining object-oriented technique and GURLS classifier

Cloud is an important factor in remote sensing information acquisition. With the application of domestic high spatial resolution satellite data, accurate cloud detection has important significance to the ground information effective acquisition. In the paper, GF-1 and GF-2 multi-spectral images are used as data source to obtain homogenous objects by image segmentation firstly. Then based on spectral features, texture features, and geometrical features-9 features, a rule set is established. With the rule set as input, the GURLS classifier is used to detect cloud combined with threshold method. Applied on high resolution data with different time and scenarios, the method is compared with the pixel-based maximum likelihood method and SVM method. The result shows that the proposed method has a cloud extraction accuracy of over 95% and a Kappa coefficient of over 0.9.