基于ZY3遥感图像的反立体校正方法研究

Image correction methods to remove false topographic perception Phenomena(FTPP) of ZY3 imagery

北半球资源卫星图像普遍存在的反立体现象,造成理解、判读遥感图像的视觉混淆。文中以胶东半岛栖霞市丘陵区为例,基于地形阴影模型,分别采用HSV、Gram-Schimdt变换及SRM乘法对ZY3遥感图像的反立体现象进行校正。实验结果表明:三种方法均能实现图像正立体化;SRM乘法校正后图像光谱保持能力好,适宜定量遥感分析;HSV融合图像光谱失真明显,但信息丰富、清晰度好且地形正立体感强,适合地貌学研究;GS融合图像正立体感明显,但光谱失真较明显,信息量与清晰度不如HSV融合法,适合通用制图应用。

Remote sensing images were widely used cartography, planning design and resource monitoring, but northern resources satellite images were widespread false topographic perception phenomenon which offen against human visual habits. Taking the hills of Qixia City,Shandong Peninsula as study area and basing the SRM,it uses three methods to correct this false topographic perception phenomenon in ZY3 remote sensing images. The results showed that： three methods can remove the false topography effectively： multiplying the images by SRM can contain a high-resolution of the original image,but poor three-dimensional effect,which was suitable for guantitative analysis of remote sensing;the remote sensing image distortion based on HSV method was obvious but with abundant information,good clarity and three-dimensional effect,which was suitable for geomorphology study; the GS spectral sharpening method had the better three-dimensional effect than HSV method,but had the poorer clarity and information,which was suitable for general map-making.