论文部分内容阅读
为解决现行云地检测算法不适用于亚米级全色遥感影像云地检测的问题,提出一种大尺度自适应匹配阈值(LS-AMTH,large-scale adaptive matching threshold)算法。算法构建包含光谱、纹理与边缘特征的特征参量集,利用提升算法对影像子块进行大尺度云地分类;之后对大尺度分类所得云地子块进行阈值的自适应匹配选择,最终实现像素级云地区域检测并统计云地占比。试验表明,针对亚米级全色影像,本文算法准确度达97.3%,在复杂云地混合区域取得良好检测效果。
In order to solve the problem that the current cloud detection algorithm is not suitable for cloud detection in sub-panchromatic remote sensing images, a large-scale adaptive matching threshold (LS-AMTH) algorithm is proposed. The algorithm constructs a feature parameter set including spectral, texture and edge features, and uses the lifting algorithm to classify the image sub-blocks into large-scale clouds. After that, the threshold adaptive matching is selected for the sub-blocks obtained from the large-scale classification. Finally, Cloud area detection and statistics cloud proportion. Experimental results show that the accuracy of this algorithm is 97.3% for sub-panchromatic images, and good detection results are obtained in the mixed cloud areas.