融合WorldView-2与机载激光雷达数据估算城市不透水面

发布时间：2018-06-20 03:54

本文选题：WorldView- + 激光雷达　；参考：《遥感信息》2017年02期

【摘要】：针对高空间分辨率遥感影像进行城市不透水面提取时存在的同物异谱、异物同谱及阴影等局限性,提出一种基于WorldView-2高分影像与机载激光雷达数据融合的分层分类估算城市不透水面的方法。该方法首先运用基于雾霾与比值(haze-and-ratio-based,HR)的融合算法对WorldView-2多光谱波段与全色波段进行数据融合;然后依据LiDAR归一化数字表面模型(normalization digital surface model,nDSM)高度阈值分为地面物体与非地面物体,运用像元尺度上分层支持向量机分类算法进行城市不透水面百分比估算;最后结合特征阈值和GIS空间分析法探测阴影区域不透水面。研究结果表明,与传统的高分影像提取城市不透水面方法相比,该方法可以明显改善材质复杂的建筑物屋顶提取不完整,以及高亮裸土与高反照度屋顶相互混淆的现象,并通过阴影校正可以较好地区分阴影区域的植被与不透水面信息,进而提高城市不透水面估算精度。
[Abstract]:In view of the limitations of isomorphism, foreign body isospectral and shadow in extracting urban impermeable surface from high spatial resolution remote sensing image, and so on, A hierarchical classification method based on the fusion of WorldView-2 high score image and airborne lidar data is proposed to estimate the impermeable water surface of the city. The method firstly uses the fusion algorithm based on haze and ratio-based HR) to fuse the multispectral band and panchromatic band of WorldView-2, and then divides the height threshold of normalization digital surface model into ground objects and non-surface objects according to the LiDAR normalized digital surface model normalization digital surface. A hierarchical support vector machine (SVM) classification algorithm on pixel scale is used to estimate the percentage of urban impermeable surface. Finally, feature threshold and GIS spatial analysis are used to detect the impermeable surface of shaded area. The results show that compared with the traditional method of extracting urban impervious surfaces from high-grade images, this method can obviously improve the incomplete extraction of building roofs with complex materials, and the confusion between high-brightness bare soil and high-reflective roofs. The information of vegetation and impermeable surface in shaded area can be well distinguished by shadow correction, and the estimation accuracy of impermeable surface of city can be improved.
【作者单位】：湖南师范大学资源与环境科学学院;中国科学院遥感与数字地球研究所数字地球重点实验室;海南省地球观测重点实验室;山东科技大学测绘科学与工程学院;
【基金】：国家自然科学基金(41201357) 国家重点研发计划“全球变化及应对”重点专项(2016YFA0600302) 三亚市院地科技合作项目(s15ys02132)
【分类号】：TN957.52

【参考文献】