目的 为了解决当前图像融合技术中易丢失图像信息，不能较好地保持源图像的边缘与纹理信息，从而降低了图像分辨率与视觉质量，使其不能对目标进行清晰、完整、准确地信息描述等问题。方法 提出一种导向滤波耦合分形维度的图像加权融合方案。首先对源图像进行预处理，通过增强对比度来提高图像的动态范围。通过小波变换将图像分解为低频与高频部分，并引入导向滤波器，对其低频、高频成分进行处理，获取相应的低频、高频权重，较好地保持图像的边缘信息。然后，通过提取局部特征分形维数来获取微小纹理特征。最后，定义一种加权融合方案，根据低频与高频权重进行融合，得到最后融合图像。结果 实验数据表明，与当前常用图像融合算法比较，文中算法具有更好的融合视觉效果，更好地保持了源图像的真实信息；在信息熵、交互信息、平均梯度和标准差等4种定量分析指标方面，所提算法具有更大的优势。结论 所提算法具有良好的融合质量，在图像处理领域具有一定的参考价值。

The work aims to solve the problems that the image information is easily lost in the current image fusion technology, and the edge and texture information of the source image cannot be better preserved, thus reducing the image resolution and visual quality and discouraging it from the clear, complete and accurate information description of targets, etc. The weighted image fusion scheme based on guided filtering coupled fractal dimension was proposed. Firstly, the source image was preprocessed, and the dynamic range of the image was improved by contrast enhancement. Secondly, the image was decomposed into low frequency part and high frequency part by wavelet transform. Then, the steerable filter was introduced to process the low frequency and high frequency components, so as to obtain the corresponding weights of low frequency and high frequency. The edge information of image could be well maintained. Thirdly, the local feature fractal dimension was extracted to obtain the microtexture feature. Finally, a weighted fusion scheme was defined, and the fusion was conducted according to the low-frequency and high-frequency weights to obtain the final fused image. The experimental data showed that, compared with the current commonly used image fusion algorithm, the proposed algorithm had better visual effects of fusion and kept the real information of the source images in a better manner. With respect to four quantitative analysis indicators (IE, MI, AG and STD), the proposed algorithm had more advantages. The proposed algorithm has good fusion quality, which has certain reference value in the field of image processing.