基于稀疏支撑集先验的压缩感知图像序列重建算法

Compressed Sensing Image Sequence Reconstruction Algorithm Based on Sparse Support Prior

针对现有压缩感知图像序列重建算法重建精度不高、模型参数设置较多的问题,提出了一种结合稀疏支撑集先验和残差补偿的算法。在已知前一帧图像重建结果的基础上,通过求解1个最小化加权l1范数问题得到当前帧图像的初始估计。通过对估计残差进行压缩感知重建并对初始估计加以补偿,得到当前帧图像的最终重建结果。与其他同类算法相比,该算法减少了阈值参数的设置。实验结果表明,在相同的测量值数目下,该算法重建图像的相对误差、峰值信噪比和结构相似度指标均优于同类比较算法。

Aiming at the problems of low accuracy and more model parameters of traditional compressed sensing image sequence reconstruction algorithms, a novel algorithm combining sparse support prior and residual compensation is proposed. The initial estimation of the current image is obtained by solving a weighted lI norm minimization problem based on knowing the reconstruction of the previous image. The final estimation of the current image is generated by the compressed sensing reconstruction of the estimation error and the compensation of the original estimation. Compared with other similar algorithms, the proposed algorithm reduces the number of threshold parameters. Experimental results show that the proposed algorithm is superior to other similar algorithms in terms of relative error, peak signal to noise radio and structural similarity of reconstructed images with same number of measured values.