提升静态小波与自适应pcnn相结合的图像融合算法 image fusion using lifting stationary wavelet transform and adaptive pcnn.pdfVIP

第 37 卷第 12 期 光电工程 Vol.37, No.12 2010 年 12 月 Opto-Electronic Engineering Dec, 2010 文章编号：1003-501X(2010)12-0067-08 提升静态小波与自适应 PCNN 相结合的图像融合算法 郭茂耘，李华锋，柴 毅 ( 重庆大学 自动化学院，重庆 400044 ) 摘要：提出了一种新的基于提升静态小波变换与自适应 PCNN 相结合的图像融合算法。该方法定义一种图像单个 像素的清晰度作为 PCNN 的链接强度，使得 PCNN 能根据像素清晰度的变化来自适应地调整链接强度的大小，接 着对图像经提升静态小波分解得到的低频子带系数的改进拉普拉斯能量和及高频子带系数的单个像素的灰度值， 分别作为自适应 PCNN 神经元的外部输入，并根据点火次数来确定图像融合系数。最后由提升静态小波变换的逆 变换得到融合图像。实验表明，该方法在视觉效果和客观评价指标上都优于传统的基于小波变换、提升静态小波 变换、提升静态小波-PCNN 的图像融合算法。 关键词：图像融合；提升静态小波；脉冲耦合神经网络；拉普拉斯能量和 中图分类号：TP391; TN911.73 文献标志码：A doi：10.3969/j.issn.1003-501X.2010.12.013 Image Fusion Using Lifting Stationary Wavelet Transform and Adaptive PCNN GUO Mao-yun，LI Hua-feng ，CHAI Yi ( Department of Automation, Chongqing University, Chongqing 400044, China ) Abstract : A novel image fusion algorithm based on the Stationary Lifting Wavelet Transform (SLWT) and adaptive Pulse Coupled Neural Network (PCNN) is proposed. Compared with the traditional PCNN where the linking strength of each neuron is the same, this adaptive PCNN uses the sharpness of each pixel as its value, so that the linking strength of each pixel can be chosen adaptively. By using a stationary lifting wavelet transform, we can calculate a flexible multiscale and shift-invariant representation of registered images. A Novel Sum-Modified-Laplacian (NSML) in the low frequency subbands and the pixels value of high frequency subbands of SLWT are input into motivate adaptive PCNN, respectively. The