基于分数阶小波与引导滤波的多聚焦图像融合
首发时间:2022-05-10
摘要:针对多聚焦图像融合中存在易丢失细节信息、图像边缘处产生伪影等问题,提出一种基于离散分数阶小波变换(DFRWT)结合引导滤波的多聚焦图像融合新方法。首先,该方法采用DFRWT将源图像进行多尺度分解,得到低频部分与高频部分。其次,低频部分应用引导滤波加权的融合规则,高频部分采用分数阶空间频率的融合规则。最后,通过DFRWT逆变换获得融合后的图像。新方法与现有的五种算法进行视觉对比实验和定量评估,仿真实验表明,本文方法有效抑制了Gibbs效应和边缘处的伪影效应,获得了较好的视觉效果和较高的客观评价,融合图像的质量优于已有的几类经典算法。
关键词: 应用数学 多聚焦图像融合 离散分数阶小波变换 引导滤波 融合规则
For information in English, please click here
Multi-focus Image Fusion Based on Fractional Wavelet Transform and Guided Filtering
Abstract:There are some problems such as easy to lose details and produce artifacts at image edges in multi-focus image fusion. To address the aforementioned problems, a new multi-focus image fusion method based on discrete fractional wavelet transform (DFRWT) combined with guided filtering was proposed. First, in order to obtain the low-frequency parts and high-frequency parts, the method uses DFRWT to decompose the source image at multiple scales. Then, the fusion rule of guided filtering weighting is applied to the low frequency part and fractional spatial frequency fusion rule is adopted for high frequency part. Finally, the fused image is obtained by inverse DFRWT. The new method is compared with the existing five algorithms for visual comparison experiments and quantitative evaluation. Simulation experiments show that the proposed method can effectively suppress Gibbs effect and edge artifact effect, achieves better visual effect and objective evaluation, the quality of fusion image is better than several classical algorithms.
Keywords: applied mathematics multi-focus image fusion discrete fractional wavelet transform guided filtering fusion rules
引用
No.****
动态公开评议
共计0人参与
勘误表
基于分数阶小波与引导滤波的多聚焦图像融合
评论
全部评论0/1000