摘要
充分利用小波变换的多分辨率分析特性,文章提出了一种基于显著特征的小波变换图像融合方法。首先对待融合图像进行小波分解,得到低频系数和高频系数;然后对低频系数提出了一种基于显著特征——空间频率的融合规则,对高频分量使用系数绝对值选大的规则进行融合;最后对融合后的低频系数和高频系数进行小波反变换得到融合图像。同时探讨了多聚焦图像中存在的虚假轮廓现象并提出了一种简便的定位方法。仿真实验表明,文中算法得到的融合图像在视觉效果上较传统算法及其一些改进算法有所提高,同时熵、标准差和互信息等客观评价指标值也得到了提高。
The introduction of the full paper,after discussing Refs.1 through 6,proposes a new and more effective image fusion algorithm which combines the wavelet transform with salient feature.Section 1 explains our new algorithm;its core consists of： we decompose the images to be fused and obtain their low-frequency coefficients and high-frequency coefficients by using the wavelet transform;for the low-frequency coefficients,we propose a new fusion rule based on the salient feature,which is obtained by computing spatial frequencies with eqs.（1）,（2） and（3）;the maximum high-frequency coefficients are selected by using maximum absolute values;thus we obtain the fused image by using the inverse wavelet transform of the fused low-frequency coefficients and high-frequency coefficients;we also give the procedural steps of our image fusion algorithm.With the help of Fig.1,section 2 discusses the illusive contours of multi-focus images and proposes a simple method for locating them.The simulation results,presented in Figs.2 and 3 and Table 1,and their analysis show preliminarily that our image fusion algorithm is more effective than both the traditional and some improved methods based on visual effect and objective evaluation values such as entropy,standard deviation and mutual information.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2010年第4期486-490,共5页
Journal of Northwestern Polytechnical University
基金
航空科学基金(20090153003)
陕西省自然科学基金(2010JQ8015)
陕西省科技计划项目(2010k06-23)
西北工业大学"翱翔之星"人才计划项目资助
关键词
图像处理
小波变换
算法
图像融合
显著特征
image processing
wavelet transforms
algorithms
image fusion
salient feature
