The potential application of a new compressive sensing ( CS) theory in the image fusion field was investigated, and a new image fusion method based on the CS principle was put forward to overcome the common defects in fused images such as too high brightness, contrast inconsistency to source images, and emergence of striped noise. The method uses a novel self-adaptive weighted average fusion operator presented in the study based on the standard deviation (SD ) of compressive measurements to fuse source images in the special domain, and then the total variation optimization algorithm is employed to reconstruct fused results. A series of simulation experiments on fusing multiple images from different kinds of sensors were carried out to validate the effectiveness of the proposed method. Both the subjective visual effect and the objective evaluation indicate that the presented algorithm can achieve a high level of fusion quality. It can prevent from the emergence of stripes effectively and enhance the definition of fused images greatly by extracting more useful informaton from source images.%研究了新兴的压缩感知(CS)理论在图像融合中的潜在应用,针对已有图像融合方法容易导致融合图像的亮度过高、与源图像的对比度不能保持一致,且伴有明显条纹出现等的缺点,提出了一种新的基于CS原理的图像融合方法,该方法用基于压缩测量值标准偏差(SD)的自适应加权平均融合算子对各待融合图像在CS域的投影测量值进行融合,再经总变分优化算法对融合测量值进行重构而得到融合图像.为验证该方法的有效性,进行了多组不同类型传感器所获图像的融合实验,主观视觉分析和客观评价指标的统计结果均表明,该算法在有效抑制图像中条纹现象发生的同时从原始图像提取了更多的有用信息,显著增强了图像的清晰度,取得了更好的视觉效果,从而展示了CS技术在图像融合方面的突出效果和应用潜力.
展开▼
