基于高阶交叠组稀疏正则项的图像恢复方法

为进一步提高交叠组稀疏全变分模型的图像恢复效果,通过在现有模型的基础上结合图像的二阶梯度信息，增加二阶梯度正则项的方法对交叠组稀疏正则项进行改进，研究了基于高阶交叠组稀疏正则项的模型建立和算法以及在图像恢复中的效果及各参数的影响。结果表明：在一定水平噪声标准差的情况下，应用该方法对图像进行恢复时，基本上均可获得比其他模型更好的恢复效果。可见在交叠组稀疏全变分模型中考虑二阶梯度信息有助于提高图像的恢复性能。

Research on Image Restoration Method Based on High-order Overlapping Group Sparse Total Variation

In order to further enhance the image restoration effect of overlapping group sparse total variation, combined with image second-order gradient information, the second-order gradient regular term was used on the model to investigate the model building of high-order overlapping group sparse regular term and the corresponding algorithm and then the effect in image restoration and the parameters. The results show that the model is better than the others in image restoration under certain level of noise standard deviation in most occasions. It is concluded that it is helpful for enhancing the performance of image restoration to use the second-order gradient information in overlapping group sparse total variation.