多源多时相遥感影像相对辐射归一化方法研究

发布时间：2018-03-18 03:24

本文选题：时间序列　切入点：多源影像　出处：《地球信息科学学报》2016年05期 　论文类型：期刊论文

【摘要】：为满足多源、多时相遥感影像定量化信息提取的应用需求,本文发展了一种半自动化的相对辐射归一化方法。将多源遥感影像的相对辐射归一化分为传感器辐射校正与针对光照等外部因素的辐射归一化2个过程。首先,基于晴空影像,采用分类回归的方式获取传感器辐射校正系数;然后,利用样本传递再分类的方法实现多源影像的半自动分类和传感器辐射偏差校正;最后,基于NDVI差值直方图和类别约束的PIFs自动选取方法,实现影像的相对辐射归一化。采用准同期的GF1-WVF1和Landsat8-OLI影像以及多源时序影像对方法进行了验证,结果表明,本文方法可以对传感器间的辐射偏差进行有效纠正,并在整体上获得比传统方法更好的辐射归一化精度;同时,多源时序影像的辐射校正结果也表明,本文方法能够有效地消除时序影像间的辐射特征波动,使植被等地类的季相变化信息得到更准确地表达,为多源时序影像的协同利用提供了借鉴方法。
[Abstract]:In order to meet the requirements of quantitative information extraction from multi-source and multi-temporal remote sensing images, In this paper, a semi-automatic method of relative radiation normalization is developed. The relative radiation normalization of multi-source remote sensing images is divided into two processes: sensor radiation correction and radiation normalization for external factors such as illumination. The sensor radiation correction coefficient is obtained by classifying regression method. Then, the semi-automatic classification of multi-source images and sensor radiometric deviation correction are realized by the method of sample transfer subclassification. Finally, The method of PIFs automatic selection based on NDVI difference histogram and category constraint is used to normalize the relative radiation of images. The method is verified by quasi-synchronous GF1-WVF1 and Landsat8-OLI images and multi-source sequential images. The method in this paper can effectively correct the radiation deviation between sensors, and achieve a better radiometric normalization accuracy than the traditional method. At the same time, the radiation correction results of multi-source sequential images also show that, The method in this paper can effectively eliminate the radiative characteristic fluctuation among temporal images and make the seasonal variation information of vegetation and other land types more accurate. It provides a reference method for the cooperative utilization of multi-source temporal images.
【作者单位】：中国科学院遥感与数字地球研究所遥感科学国家重点实验室;中国科学院大学;广西农业科学院农业科技信息研究所;
【基金】：国家高技术研究发展计划项目(2015AA123901) 广西科学研究与技术开发计划项目(14125008-1-6) 国家高分辨率对地观测系统重大专项(03-Y30B06-9001-13/15-01)
【分类号】：P237

【相似文献】