基于微位移的超分辨重建技术研究

CCD图像传感器广泛用于图像采集,受其内部结构和外部条件的影响,采集到的图像质量不能满足人们的需求,通过硬件改进提高图像质量面临经济与技术两方面的难题。为此本文利用多幅微位移图像间的冗余信息重建出高分辨率图像,既低成本又易实现。针对CCD与目标物间有相对微位移来获取序列低分辨率图像的情况,采用一种解线性方程组的方法对采集到的4幅微位移图像进行重建,并对这种算法进行优化。实验结果图可清晰地看到重建图像可分辨更多线对,高频信息量增加,算法具有较好的效果。

Research on super-resolution reconstruction based on micro-displacement

CCD image sensor is widely used in image acquisition.The quality of images can not be content with people’s requirement by the influence of internal structure and external conditions.We will be up against two big puzzles if we want to enhance the image quality through improving the hardware.Therefore this paper uses the redundant information among multiple micro-displacement images to obtain high-quality image which is lowcost and easy to realize.This paper aim at that the low-resolution images are obtained by the relative micro-displacement between CCD and target.The solving linear equations algorithm which is employed to reconstruct the four micro-displacement images are suggested and optimized in this paper.Experiment shows that the images been reconstructed can resolve more pair and the high-frequency information have increased.Moreover,the algorithm has better effect.