基于小波变换的红外与X光图像融合方法研究

红外与可见光图像融合只能反映构件表面信息和温度场分布,由于小波变换具有良好的时域和频域多分辨率性,为了获得构件内部发热源,采用了基于小波变换的红外与X光图像融合方法。首先将红外彩色图像做灰度化处理;然后采用MALLAT算法对红外与X光图像作小波分解,用低频系数平均、高频系数取绝对值大的融合规则进行融合,并利用YIQ空间变换将红外图像的彩色温度信息迁移至小波融合图像上;最后采用信息熵、标准差、方差等质量评价参量对融合后的图像进行评价。结果表明,经过小波变换融合的红外与X光图像融合在继承了两幅图像信息基础上进行了折中,既包含了构件的温度场分布,同时也反映了构件内部清晰的结构。

Study on fusion method of infrared and X-ray images based on wavelet transform

An object's surface temperature distribution and surface information can be obtained generally through infrared and visible image fusion. In order to find out internal heat source within an object, an image fusion strategy between infrared and X-ray images based on wavelet transform was proposed because wavelet transform has better multi-resolution properties in time domain and frequency domain. Firstly, the infrared color image was transformed into gray-scale image. Then, Mallat algorithm was employed to decompose the infrared and X-ray images. The fusion image was gotten by the rules of low frequency coefficient averaging and high frequency coefficient absolute value maximizing. The colored temperature information of infrared images was transferred to the fused wavelet images by using YIQ space transform. Finally, the fused image was evaluated quantitatively by image entropy, standard deviation and average gradient. The experiment results show that the fused image integrates infrared and X-ray image information, which reflects not only the component's surface temperature distribution but also its clear inner structure.