一种优化的迭代反投影超分辨率重建方法

Super Resolution Reconstruction Method Based on Optimized Iterative Back Projection

针对现有超分辨率重建方法存在重建的高分辨率图像具有边缘结构不清晰、高频信息提升有限的问题,提出了一种结合迭代反投影(iterative back-projection,IBP)与限制对比度自适应直方图均衡化(contrast limited adaptive histogram equalization,CLAHE)的遥感图像超分辨率重建方法。通过结合迭代反投影与限制对比度自适应直方图均衡的超分辨率重建方法重建高分辨率遥感影像,增加重建图像的纹理细节信息,并利用灰度信息与梯度特征的能量优化模型对重建图像进行优化处理。本文以不同地貌类型遥感图像为实验数据。实验结果表明,该重建方法与经典迭代反投影方法相比,纹理细节丰富,边缘结构清晰,与非局部迭代反投影(nonlocal iterative back-projection,NLIBP)相比,客观质量评价指标均有明显改善,验证了该方法具有较好的普适性与鲁棒性。

Aiming at the problems of the existing super resolution reconstruction method,such as the edge structure of reconstructed high-resolution image is not clear,and the high frequency information improved is limited,a super resolution reconstruction method of remote sensing image based on iterative back-projection and contrast limited adaptive histogram equalization is proposed in this paper.Super resolution reconstruction based on iterative back-projection and contrast limited adaptive histogram equalization is utilized to construct the high-resolution remote sensing image,in order to improve the textural details information.Then,the final reconstructed result is acquired through the energy optimization model based on gray information and gradient characteristic.The different landform types of remote sensing image are selected as the experimental data.The experimental results show that compared with the classical iterative back projection reconstruction method,texture detail is rich and edge structure is clear.Compared with the nonlocal iterative back projection(NLIBP),the objective evaluation indexes are improved obviously.It is proved that the super resolution reconstruction method proposed in this paper has good adaptability and robustness.