双能图像融合算法研究

Research of dual⁃energy image fusion algorithm

为解决X射线单能量成像的图像像素点缺失、图像分辨率较低等问题,且为更好地保护图像的细节信息,给出了结合双能和小波技术的图像融合算法。对原始采集的图像进行小波分解,等价于用一组高、低通滤波器进行滤波。通过高能量和低能量下的小波分解,用基于邻域像素的关联性和区域方差提取低频分量。用Sobel算子检测局部细节信息,将梯度最大值作为算子检测的图像边缘输出,从而提取高频分量。采用小波重构算法进行逆变换,形成新的融合图像并进行定量分析。结果表明,这种双能融合方法可以很好地解决X射线单能量成像的局限性,较好地保护图像的细节信息,增强图像成像质量。

In order to solve the problem of missing pixels and low image resolution in X⁃ray single energy imaging,a dual⁃energy image fusion algorithm based on wavelet transform is presented.Wavelet decomposition of the registered source image is equivalent to filtering with a set of high⁃low pass filters.The low⁃frequency components are extracted by using the correlation and regional variance of the domain pixels through wavelet decomposition at high and low energies.Sobel operator is used to detect local details,and the gradient maximum value is used as the image edge output detected by the operator,so as to extract the high⁃frequency components.A new fusion image is formed by inverse transform of wavelet reconstruction algorithm and quantitative analysis is carried out.The results show that the dual⁃energy fusion method can solve the limitation of X⁃ray single⁃energy imaging,better protect the details of the image,and enhance the image quality.