为了提高不等厚工件数字射线照相(DR)图像的信息量，将小波图像融合技术应用于两个不同X射线能量下的DR图像的融合。首先，利用小波变换对待融合图像进行多分辨率分解，采用加权平均法提取低频系数；然后利用八个方向掩模计算局部梯度，基于局部梯度选择融合高频系数；最后基于小波融合系数逆变换实现图像的融合处理。利用熵、空间频率等客观参数对融合图像进行评价。结果表明，此方法可以得到理想的融合效果。

In order to increase the digital radiography(DR) image information of the object with uneven thickness, a fusion method based on wavelet transform on two DR images of different Xray energy was introduced. Firstly, the input images were decomposed by wavelet transform. Secondly, the weighted average method was used to choose the weight coefficients of low frequency subimage. And the eight directional masks were used to detect the local gradients. On the basis of the local gradients the fused high frequency coefficients were obtained. Finally, through inverse wavelet transform the output image could be obtained. In addition, entropy and spatial frequency were used to evaluate the fused images. Experiments showed that this method was effective.