基于成像系统姿态信息的分时积分图像偏移修复

Time-Sharing Integral Image Offset Restoration Based on Imaging System Attitude Information

针对多帧短曝光图像叠加,提出一种融合成像系统自身的角度信息,通过光学系统与相机系统的坐标变换计算来获得图像偏移像素的方法。对于每一帧短曝光图像的角度,利用光学系统几何关系进行亚像素级偏移配准,再通过傅里叶变换的频移特性构建傅里叶频谱滤波器H (u,v),在频域对图像进行亚像素级别的修正,并使用刃边法对图像进行高频响应评估。对比未修复图像,经所提算法修正后图像在50%调制传递函数(MTF50)衰减处的分辨率从20 lp/mm提升到33 lp/mm,且其修复效果优于传统二次插值算法的修复效果;所提算法在40~80 lp/mm分辨率范围内具有较好的修复效果,MTF的峰值提升了4倍。

Aiming at the multi-frame short exposure image overlay,this paper proposes a method that integrates the angle information of the imaging system and obtains the image offset pixels through the coordinate transformation calculation between the optical system and the camera system.For the angle of each frame of a short-exposure image,the geometric relationship of the optical system is used to perform sub-pixel offset registration,and the Fourier spectral filter H(u,v) is constructed through the frequency shift characteristic of the Fourier transform.The image is corrected at the sub-pixel level in the frequency domain,and the high-frequency response of the image is evaluated by the edge method.Compared with that of the unrepaired image,the line pair of the image corrected by the proposed algorithm is increased from20 lp/mm to 33 lp/mm at a modulation transfer function (MTF) attenuation of 50%(MTF50),with its repair effect higher than that of the traditional quadratic interpolation algorithm.The proposed algorithm shows a positive repair effect in the line pair part of 40-80 lp/mm,and the peak value of the MTF is increased by four times.