在显微成像过程中，受系统景深限制，沿光轴方向不同层面之间聚焦位置存在显著差异，同时不同层面的显微图像存在部分聚焦区域重叠，现有多聚焦融合算法往往无法并行提取和融合多幅显微图像中最清晰的聚焦部分。提出一种多聚焦显微图像融合算法，首先构造了一种类高斯四邻域梯度算子并结合快速引导滤波，实现高频聚焦信息的提取；同时针对大视场显微图像序列中存在聚焦信息重叠、像素数量大的情况，引入了一种小区域聚焦度量方法，提高了对聚焦清晰区域高频信息提取的能力，实现了多图最佳聚焦点的融合。拍摄3组包括4 mm和2 mm对角线视场的多聚焦显微图像序列进行测试，相较5种常用多聚焦图像融合算法，所提算法的峰值信噪比平均提高了2.4772，结构相似性指数达0.9400以上，对聚焦清晰区域有更好的融合效果，融合图像细节丰富且清晰度高，能够满足大视场多聚焦显微图像融合的准确性要求。

In the process of microscopic imaging, the system's depth-of-field limitation results in considerable differences in the focused positions across various planes along the axial direction. This leads to partial overlap of focused regions among microscopic images from different planes. Current multifocus fusion algorithms often struggle to simultaneously extract and merge the sharpest focused parts from multiple microscopic images. Hence, this article proposes a multifocus microscopic image fusion algorithm. First, a Gaussian-like four-neighborhood gradient operator was constructed and combined with fast guided filtering to extract high-frequency focus information. Additionally, a small region focus measurement method was introduced to enhance the extraction for high-frequency focus information from sharply focused regions considering the overlap of focus information and the substantial number of pixels in wide-field microscopic image sequences. This method effectively fuses the best focus points from multiple images. Through experiments, three sets of microscopic multifocus image sequences covering the diagonal fields-of-view of 4 mm and 2 mm were captured and testing was conducted. Through comparative analysis against five commonly used multifocus image fusion algorithms, our algorithm yields an average improvement in the peak signal-to-noise ratio of 2.4772 and surpasses a structural similarity index of 0.9400. These results exhibit superior fusion effects obtained by the proposed algorithm in the focused regions, rendering fused images enriched in details and high clarity. This algorithm meets the accuracy requirements for multifocus image fusion in applications involving large field-of-view microscopic images.