增广拉格朗日双边全变分压缩成像重构算法

Augmented Lagrangain bilateral total variation compression imaging reconstruction algorithm

针对基于全变分压缩成像算法重构的图像存在虚假边界以及边缘信息对比度低的问题,提出了一种基于全变分成像模型的增广拉格朗日双边全变分压缩成像重构算法。在全变分正则化思想基础上引入双边滤波技术,并加入拉格朗日函数算子,将目标函数转化为增广拉格朗日函数,利用交替方向法求解函数模型的最优解。迭代过程中选用最速下降法对梯度进行求解,对算法进行优化,提高算法运行速度。实验结果表明,算法改进后可以更加精确的重构出原始图像,重构图像的峰值信噪比提高2 dB,重构错误率降低10%,结构相似度提高0. 1,并且对噪声具有较好的鲁棒性。

In order to solve the problem that the reconstructed image based on the total variation compression image reconstruction algorithm has a fictitious boundary and low contrast of the edge information, an augmented Lagrangain bilateral total variation compression reconstruction algorithm based on the total variation is proposed. Based on the idea of total variation regularization, the bilateral filtering technique is introduced and the augmented Lagrangain function operator is added. The mininmm value of the Lagrangain function model is solved by the alternating direction method. In the iterative process,the steepest descent method is adopted to solve the gradient, and the algorithm has been optinfized to improve the algorithm running speed. The experimental results show that the improved algorithm can reconstruct the original image more accurately. The peak signal-to-noise ratio of the reconstructed image increases by 2dB,the reconstruction error rate decreases by 10% and the structural similarity increases 0. 1 ,and has the better robustness to noise.