集成特征分量的阴影多尺度分割方法

发布时间：2019-07-20 15:33

【摘要】：针对高分影像阴影检测精度易受水体、深色地物和暗色植被影响等问题,结合GF-1影像自身特点,提出一种结合特征分量构建和多尺度分割面向对象的阴影检测方法。首先,对GF-1影像多光谱数据、全色数据进行正射校正和信息融合,以达到光谱与空间分辨率信息最大化利用。其次,集成特征分量(主成分第一分量PC1、亮度分量V、绿光波段G、归一化植被指数NDVI)以增强阴影信息。最后,对集成后的影像进行多尺度分割,并利用特征分量构建规则集,最终实现阴影信息提取。实验表明,该方法既能准确地检测出GF-1影像中的阴影信息,又能有效削弱水体、深色地物和暗色植被的影响。
[Abstract]:In order to solve the problems that the accuracy of shadow detection in high score images is easily affected by water body, dark ground objects and dark vegetation, combined with the characteristics of GF-1 images, an object-oriented shadow detection method combining feature component construction and multi-scale segmentation is proposed. Firstly, the multispectral data and panchromatic data of GF-1 images are corrected and information fusion is carried out in order to maximize the utilization of spectral and spatial resolution information. Secondly, the feature components (principal component PC1, luminance component V, green light band G, normalized vegetation index NDVI) are integrated to enhance shadow information. Finally, the integrated image is segmented by multi-scale, and the rule set is constructed by using feature components, and finally shadow information extraction is realized. The experimental results show that this method can not only accurately detect the shadow information in GF-1 images, but also effectively weaken the influence of water body, dark ground objects and dark vegetation.
【作者单位】：兰州交通大学测绘与地理信息学院;甘肃省地理国情监测工程实验室;甘肃省遥感重点实验室;
【基金】：甘肃省科技计划资助(1506RJZA070,148RJZA028) 甘肃省高等学校科研项目(2015A-049) 甘肃省遥感重点实验室(寒旱所)开放基金(2015-RC-28)
【分类号】：TP751

【参考文献】