联合云检测与高通滤波的含云影像融合方法

发布时间：2018-06-19 19:51

  本文选题：资源三号卫星 + 卫星影像　； 参考：《武汉大学学报(信息科学版)》2016年09期

【摘要】：云噪声是光学卫星影像的常见问题,为了衡量云噪声对影像融合带来的影响,本文以高通滤波融合算法为例进行分析,指出云与地物的均值相差越大,云对影像融合的影响越大,并提出了一种针对含云影像的融合方法,即联合云检测与高通滤波的含云影像融合方法。该方法首先利用NIR/R-OTSU云检测算法实时进行云检测,判别出影像中的云覆盖区域;其次采用局部优化策略利用高通滤波融合方法分块对非云区域进行处理,得到融合影像。利用资源三号多光谱和正视全色影像进行融合实验,结果表明,本文算法比高通滤波融合方法、亮度色度饱和度(intensity hue saturation,IHS)变换融合方法、Pansharp融合方法更适用于含云影像的融合处理。
[Abstract]:Cloud noise is a common problem of optical satellite images. In order to measure the influence of cloud noise on image fusion, this paper takes high pass filtering fusion algorithm as an example to show that the greater the difference between the mean value of the cloud and the terrain, the greater the influence of cloud to image fusion, and a fusion method for cloud images, namely, joint cloud detection and detection, is proposed. The method of cloud image fusion with high pass filtering. This method first uses NIR/R-OTSU cloud detection algorithm to detect cloud in real time to distinguish the cloud coverage area in the image. Secondly, the local optimization strategy is used to process the non cloud region by high pass filtering fusion method, and the fusion image is obtained. The multi spectrum and the full view of resource three are used. The color image fusion experiment shows that the algorithm is better than the high pass filtering fusion method, the intensity hue saturation, IHS transform fusion method, and the Pansharp fusion method is more suitable for the fusion processing of the cloud image.
【作者单位】： 武汉大学遥感信息工程学院;
【基金】：国家自然科学基金(41322010,41301519) 国家863计划(2013AA12A401)~~
【分类号】：P237;TP751
，

本文编号：2041063