图像三要素与MTF的分析研究  

Research on the Relationship of Three Image Elements and MTF

作　　者：尤建洁[1] 胡晔[1] 戴奇燕[1] 刘兆军[2] 王怀义[2] 夏德深[1] 

机构地区：[1]南京理工大学计算机学院,南京210094 [2]北京空间机电研究所,北京100076

出　　处：《宇航学报》2006年第B12期89-96,共8页Journal of Astronautics

摘　　要：中低分辨率在轨运行卫星遥感器的调制传递函数（Modulation Transfer Function，MTF）的监测一直是研究的难题。现提出的方法解决了中低分辨率在轨运行卫星遥感器MTF监测的关键技术。考虑到卫星遥感器的最终产物是遥感图像，从图像信息理解的三要素出发，即灰度、纹理和边缘，分析图像三要素与遥感器MTF的关系。在分析过程中，应用主成分分析方法，不仅对多组原始数据进行了简化，而且对图像参数进行了降维。通过与其它方法的比较，进一步验证了使用主成分分析方法降维的有效性；同时实验结果亦表明了主成分分析方法的可靠性。最后建立了图像单个综合参数与MTF的关系模型，拟合结果说明，此模型能较好地表达图像参数与MTF之间的关系，这将为在轨运行条件下的建模提供十分重要的技术支持。Measuring modulation transfer function （MTF） of the remote sensor of the orbiting satellite with middle or low resolution is a difficult problem to be dealt with. The method proposed in this paper solves the key problem of measuring MTF under those conditions. Considering that the remote sensing images are the final outcome of the satellite remote sensor, we start from three image elements ： gray, texture and edge, to analyze the relationship of image elements and MTF. In the process of the analysis, we apply the principal component analysis （PCA） method, not only to reduce original data, but also to reduce image parameters. The comparison between the proposed method and other method shows that the PCA used in this paper can reduce dimensions effectively, at the same time experimental results also show the stability of this method. Finally, we build a relationship model between a single synthesized parameter of the image and MTF. Fitting results show that this model can better express the relationship of image parameters and MTF. This will be an important technical support for building the model under the condition of the satellite orbiting.

关 键 词：图像要素 调制传递函数(MTF) 主成分分析(PCA) 回归分析 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]