空间相机在轨长时间运行，受空间环境影响，相机性能下降，在轨图像中出现黑色条纹，严重影响了成像质量。局部CCD像元响应下降，高能粒子冲击，宇宙尘埃，空间相机本身的附着物等都能引发图像黑色条纹的现象。文章通过对发射前实验室辐射定标数据研究，提出一种基于最小二乘法和辐射亮度反演的在轨图像的校正方法，可以有效地去除黑色条纹。在不同辐照度下，采集所有像元的灰度值，用最小二乘法估计所有像元灰度平均值与辐照度的对应关系，计算出每个像元的校正系数，并研究实验室辐射定标数据，推算出现黑色条纹像元的校正因子,通过图像黑色条纹校正函数矩阵对每个像元的实际灰度值进行校正。校正前黑色条纹像元灰度的偏差超过10%，校正后偏差小于5%。从图像上观察，经过该方法校正的图像，能够有效去除黑色条纹，得到清晰的图像，结果令人满意。

It is known that step-down response of CCDs part pixels, impact of high-energy particles and absorbed cosmic dust in space camera can all induce the black stripes for the camera image. For the camera in orbit for a long time, the image quality is seriously affected. By researching for the pre-launch calibration data, an algorithm based on the least square method and radiance retrieval was proposed to renormalize the image in orbit, which could make the black stripes in the image be effectively eliminated. Firstly, the gray levels of CCD pixels under different irradiation grades were collected, and the relation between the average gray level of all pixels and the corresponding irradiation values was established by the least square method, then the correction index of each pixel was worked out, and thus the correction index of the black stripes corresponding pixels could be obtained by analyzing the laboratory radiometric calibration coefficients. Lastly, the real gray levels of the black strips were computed by the correction index matrix. The experiment results show that for the modified images, the black stripes can be removed effectively and the image quality is satisfactory.