改进边缘质量和运动估计的超分辨率图像重构

凸集投影算法（POCS）是一种广泛使用的超分辨率图像重构方法。针对传统的POCS超分辨率图像重构算法出现的边缘模糊及匹配时的局限性问题，首先利用二阶梯度检测出像素周围0°、45°、90°、135°四个边缘。在构造参考帧时采用基于梯度的插值算法，沿边缘方向进行线性插值，沿非边缘方向进行基于一阶梯度的带权插值。在运动估计时，采用SURF匹配算法，提高匹配的鲁棒性和实时性。在修正参考帧时，分别定义中心在四个边缘方向的点扩散函数（PSF）。利用完全参考图像质量评价和无参考图像质量评价分别对仿真实验和实物实验进行了评价，评价结果表明提出的算法较传统POCS算法有明显的改善。 

Improve the Quality of the Edge and the Motion Estimation of Super-resolution Image Reconstruction

The projection onto convex sets algorithm is widely used for super-resolution image reconstruction. To address the fuzzy edge problems arising from conventional POCS super-resolution image reconstruction algorithm , four edge directions of 0 °, 45 °, 90 °, 135 °were detected first respectively with the second order gradient. When forming a reference frame, if an edge existed，adaptive weighted factor was obtained by calculating its one-step gradient，and then the weighted interpolation was implemented．Otherwise，linear interpolation was adopted. In motion estimation, the SURF matching algorithm was used to improve the robustness and real-time.At the same time, the SURF algorithm was utilized to match the rotational images to enhance the matching flexibility.When revising a reference frame, the point spread function was defined respectively if its center was on the four edges. The proposed POCS algorithm was tested both in the simulated and real data, which were evaluated with full reference image quality assessment and no reference image quality assessment respectively, and the results show that our method can significantly improve the quality of the reconstructed image compared with the traditional POCS algorithm.