面向对象的多时相遥感影像标准化方法研究

Research on Standardization Method of Object-oriented Multi-temporal Remote Sensing Images

为保留遥感影像地物特征,在现有相对辐射校正方法的基础上,提出了一种面向对象的影像标准化方法。首先提取地表覆盖中的感兴趣类型,将研究对象从像元变换到地物;然后根据地物类型关注度设置权重,以地物对象类型为单位,计算多时相遥感影像的辐射校正函数,以达到标准化大气状况的目的;最后利用多时相同波段影像辐射差值和影像的统计值来衡量标准化方法的效果。采用4期Landsat8 OLI遥感影像验证该方法的有效性,结果表明该方法能有效保留地物光谱曲线基本特征,降低成像条件不同造成的辐射异常。

In order to preserve the features of remote sensing images,we proposed an image standardization method based on relative radiation correction in this paper.The method firstly extracts the land cover types of interest,then sets different weights according to the attention,and finally replaces the spectra of other temporal images with the spectra of selected reference image type by type,so as to achieve the purpose of standardizing atmospheric conditions.In this study,we used four Landsat8 OLI remote sensing images to verify this method.The result shows that this method can effectively preserve the basic characteristics of spectral curves of ground objects and reduce the radiation anomalies caused by different imaging conditions.