摘要
针对红外与可见光图像进行提升小波变换后低频图像的特点,提出了一种低频分量的融合算法,高频分量采取邻域方差取大为准则进行融合,然后进行提升小波逆变换得到融合图像。通过与传统小波融合方法进行比较,并引入信息熵、清晰度、Xydeas-Petrovic客观性能指标对融合后的图像进行分析。实验结果表明不论从视觉效果还是从客观性能指标上,该算法都优于传统的融合方法。
The lifting wavelet transform is analyzed and a new image algorithm based on the feature of the low-frequency signals is proposed after the IR and visible images are decomposed by lifting wavelet transform. The max neighborhood variance algorithm can be used to the high frequency and then obtained the fusion image by inverse lifting wavelet transform. Compared the traditional wavelet fusion methods with the entropy, the image definition and the performance of the Xydeas-Petrovie are taken as the evaluation criterion. The experiment result shows that regardless of the vision feelings or the evaluation criterion, this algorithm is better than the traditional fusion method.
出处
《光学与光电技术》
2010年第4期56-59,共4页
Optics & Optoelectronic Technology
