基于压缩感知的多聚焦图像融合方法

A Multi-focus Image Fusion Algorithm based on Compressive Sensing

随着压缩感知技术的发展,基于压缩感知的图像融合技术研究受到越来越多的重视。针对图像傅里叶变换系数特点,提出了一种双星采样模式下基于高低频重要性度量的压缩传感域图像融合算法。该算法首先通过双星采样模式获得测量值,然后计算高低频区域对应的重要性度量作为融合算子,并对测量进行加权融合,最后通过求解最小全变分优化问题重构融合图像。主客观实验结果表明,该算法优于其他基于傅里叶的方案。

With the development of compressive sensing, the compressive domain image fusion technology attracts more and more attention. In considering the DFT coefficient characteristics, a novel image fusion algorithm employing double-star sampling and salience measure of high-low frequency within compression domain is proposed. Firstly, the image is measured by double-star sampling mode; then, the high-low frequency salience measure is utilized as weighted fusion operator; and finally, the fused image is reconstructed through solving the smallest total variation optimization problem. Ecperiment indicates that compared with those existing DFT fusion algorithms, the proposed algorithm is of a much better performance.