改进径向基函数插值法的多聚焦图像滤波融合
Multi Focus Image Filtering and Fusion Based on Improved Radial Basis Function Interpolation
尚宇辉 1孟伟 1房健 1王雪峰2
作者信息
- 1. 河北民族师范学院数学与计算机科学学院,河北 承德 067000
- 2. 伊犁师范大学电子与信息工程学院,新疆 伊宁 835000
- 折叠
摘要
由于受到拍摄设备、光学镜头景深以及噪声等不同因素的影响,多聚焦图像往往存在边缘弱化问题.为了使多聚焦图像更加清晰,提出基于改进径向基函数插值法的多聚焦图像滤波融合方法.通过能量函数计算图像边缘细节估计误差,建立空间同性误差准则,去除图像边缘噪声,实行滤波保护.利用径向基函数各层网络建立多聚焦图像抑噪模型,减少图像的噪声污染.通过非线性函数恢复像素,在图像空间坐标下分离图像纹理、清晰程度和图像局部特征,提取图像高频信息.利用径向基函数插值引导高频信息,在决策引导下实现图像滤波融合.经过实验证明,所提方法的多聚焦图像融合效果较好,减少了残余噪声,能够实现图像去噪和边缘特征保护,清晰度较高,图像结构性相似度高.
Abstract
Due to various factors such as shooting equipment,optical lens depth of field,and noise,multi focus im-ages often suffer from edge weakening.In order to make multi focus images clearer,a multi focus image filtering and fusion method based on improved radial basis function interpolation is proposed.Firstly,the estimation error of image edge detail was calculated by energy functions,and then the spatial homogeneity error criterion was established to re-move edge noise.Meanwhile,filtering protection was implemented.After that,a model of multi-focus image noise suppression was built by all radial basis function networks,so that the noise pollution for images could be reduced.Then pixels were recovered by nonlinear functions.Moreover,the image texture,sharpness and local characteristics of the image were separated under image space coordinates,and the high-frequency information of the image was extrac-ted as well.Finally,high-frequency information was guided by the radial based functions interpolation.Thus,the im-age filtering and fusion was achieved.The experiment results show that the proposed method has better multi-focus image fusion,reducing the residual noise.In addition,this method can achieve the image denoising and edge feature protection,with high definition and similarity of image.
关键词
径向基函数/多聚焦图像/图像融合/图像去噪/图像特征Key words
Radial basis function/Multi-focus image/Image fusion/Image denoising/Image feature引用本文复制引用
基金项目
河北省高等学校科学技术研究项目(Z2020253)
出版年
2024
NETL
NSTL
万方数据