Abstract: Following the optical remote sensing imaging mechanism and combining image spatial spectrum, MTF (modulation transfer function) and SNR (signal noise ratio) models together organically, a novel spatial frequency based radiative respond modeling (SFRR) is presented which has definite physical meaning. SFRR has two breakthrough points, one is reducing the comprehensive radiative model’s dependence on the complicated earth object’s spectral and reflectance; the other is establishing the theory basis of radiative and geometric quality correlation modeling. In geometric positioning accuracy experiments, the ZY3 satellite’s forward and backward stereo images are used to produce multilevel MTF degrade and noised simulation images. By automatic registering the forward and backward images, the Root-Mean-Square error of matched points is computed to judge the correlation between SFRR and geometric positioning accuracy, which proves that SFRR model has good correlation with geometric quality. The research will be helpful for optical remote sensing system optimizing, remote sensing image processing and applying.

Key words: remote sensing, imaging model, MTF, SNR, quality evaluation

摘要： 从光学遥感成像机理出发, 突破MTF (系统传递函数)和SNR (信噪比)模型在影像辐射质量评价中的局限性, 将影像频谱、MTF和SNR有机结合, 提出具有明确物理意义的, 基于连续空间频率的辐射响应模型SFRR以及截止空间频率处的SFRR指标。该模型一方面能有效降低综合辐射响应模型对复杂地物光谱、反射率等实测参量的依赖, 另一方面为研究辐射响应与几何质量的相关性提供了理论基础。利用资源三号前视和后视卫星影像及其MTF退化与噪声仿真影像, 进行SFRR模型及特征点像方几何定位精度相关性分析, 验证SFRR模型及其与遥感影像几何质量的定量关系。研究成果可以作为航空航天光学遥感传感器系统设计、遥感影像处理和应用的重要依据与基础保障, 对微波、激光雷达等遥感影像质量评价也有一定的借鉴意义。

关键词: 遥感, 成像模型, 系统传递函数(MTF), 信噪比(SNR), 质量评价