一种基于图像统计量的相对辐射纠正算法

线阵推扫式传感器，由于在信号传输和转换以及量化过程中产生的奇偶效应，和不同CCD探元的响应差异，导致原始图像出现明显的纵向条纹，相对辐射纠正的目的就是降低或者消除这些影响，以恢复图像原始信息。针对宽幅线阵推扫式传感器的特点，提出了一种基于图像自身统计量的相对辐射纠正算法——相邻列均衡法，消除了原始图像中明显的纵向条带效应。和常规的基于图像统计量的相对辐射纠正算法相比，这种算法的特点在于求解相对辐射定标系数时不依赖于均匀场景，并在处理北京一号小卫星图像中得到了验证。

A New Method of Relative Radioactive Correction Based on Image Statistic

Vertical striping is an obvious phenomenon in pushbroom CCD image because of the different response among CCD detectors and the odd/even effect raised in signal transferring. Image relative radioactive correction is a process to reduce or eliminate these factors and resume the true radioactive information. In this paper, a new method is brought forward aiming to destripping the strips in broad swath CCD image. The method is based on the column average and standard deviation of raw image. Compared with classical methods based on image statistic, this method obtains relative radioactive calibration coefficients independent of uniform image, which is validated in the relative radioactive correction for raw image of Beijing-1 microsatellite.