基于正交匹配追踪算法的急性运动超分辨率图像重构方法  

Super-resolution Image Reconstruction Methods for Acute Movement Based on Orthogonal Matching Pursuit Method

作　　者：郭瑞芳 

机构地区：[1]鄂尔多斯应用技术学院,鄂尔多斯017000

出　　处：《科学技术与工程》2017年第30期69-73,共5页Science Technology and Engineering

摘　　要：采用当前急性运动中超分辨率图像重构方法得到的重构图像存在全局误差,导致重构图像质量低下,重构效果不佳。为此,提出一种新的急性运动中超分辨率图像重构方法,设计急性运动中超分辨率图像重构模型,将小波稀疏字典作为急性运动中超分辨率图像重构的理论依据。将低分辨率急性运动图像分割成低分辨率图像块,对无噪高分辨率急性运动图像块相应的无噪低分辨率图像块进行分析。通过OMP方法对稀疏系数求解,依据得到的稀疏系数估计出高分辨率急性运动图像块的高频小波系数,将高分辨率小波系数急性运动图像块返回高分辨率小波系数急性运动图像,通过逆小波变换得到最终的高分辨率图像,对全局误差进行修正。实验结果表明,采用所提方法得到的重构图像质量高,重构效果好。The global error of the reconstructed image is obtained by using the super resolution image reconstruction method in the current acute motion,lead to the reconstruction of image quality is low,reconstruction effect is not good.To this end,a new method of super resolution image reconstruction in acute motion is proposed,design of super resolution image reconstruction model in acute motion,the theoretical basis of super resolution image reconstruction in acute motion with wavelet sparse dictionary.Segmentation of the low resolution acute motion image into a low resolution image block,the image block of the image with no noise and high resolution is analyzed.To solve the sparse coefficients by the OMP method,based on the estimated wavelet coefficients of acute exercise high resolution image block sparse coefficients,the acute high resolution wavelet coefficients of image block return to acute exercise high resolution wavelet coefficients of image by inverse wavelet transform to get the final high resolution image,the global error correction.The experimental results show that the proposed method is of high quality and good reconstruction effect.

关 键 词：急性运动 超分辨率图像 重构 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]