基于光照和帧间预测的空间图像曝光控制算法

Exposure Algorithm for Space Image Based on Light and Inter Prediction

为了解决空间图像中因曝光控制不理想产生的曝光过度和拖尾问题,提出了一种基于光照变化和灵活块划分的空间图像曝光控制算法。该方法考虑大气吸收对光照辐射衰减的影响,得出了光照强度与CCD相机曝光时间的关系,推导出相应的空间图像成像模型;对前一幅图像进行分块,分别比较当前块和周围相邻块的灰度均值,判断两者是否进行合并,如果灰度均值差大于阈值,则两块不合并,否则合并。根据块合并后的划分程度预测当前图像的曝光时间。实验结果表明,经过对曝光时间的准确控制,拍摄的天文图像的信噪比(SNR)最高能够提高2.49 dB。有效地解决了空间图像中曝光过度和拖尾问题,并实现了曝光时间的自适应调节。

To solve the problems of overexposure and tailing which were produced by non-ideal exposure control in the space image,an exposure control algorithm based on illumination change and flexible block segmentation was proposed. First,considering the influence of light radiation attenuation bringing by atmospheric absorption,the relationship between light intensity and exposure time of CCD was derived,and the imaging model of the space image was obtained; Secondly,in order to estimate the exposure condition of the last image,the last image was flexibly segmented,and the gray averages of two adjacent blocks were compared for judging whether the two blocks were combined. If the difference of the gray average is bigger than the threshold,the two blocks are combined,others else they are not. The exposure time can be estimated by the two methods. Experimental results indicate that the SNR of the space images increased with 2.49 dB. This algorithm can effectively inhibit the space image producing tailings and greatly reduce the phenomenon of overexposure.