光学显微数码成像系统中，光学退化是影响图像质量的重要因素。结合物镜测试，提出了一种结合分视场点扩散函数估计的梯度约束显微图像复原方法，并应用于显微镜以形成质量提升系统。通过物镜调制传递函数测试，分视场计算获取点扩散函数；引入零阶范数梯度约束作为正则化项，设计基于该约束的快速复原方法；并采用渐变加权拼接方法实现分视场复原结果的无缝拼接。对不同物镜下的显微图像进行了复原测试，并采用评价方法进行评估。结果表明，本文方法给出的显微图像复原效果好、运行速度快，对尺寸为2048 pixel×1536 pixel的图像只需1.62 s，适用于显微成像系统。

Optical imaging degradation is one of the most important factors for image quality in digital optical microscopic imaging. Combined with the digital microscopic imaging system and optical objective lens measurement, a fast gradient constrained restoration method is proposed based on point spread function estimation under different fields. By measuring the modulation transfer function of objective lens, the point spread function is obtained with different fields. A novel restoration approach is designed with the L0 gradient constraint as a regularized term. Meanwhile, the weight image stitching method is designed for the restoration of multiple fields. The image data with different objective lens are tested, and the evaluation methods are used for assessing the results. The results show that the gradient constrained microscopic imaging quality improvement method with the objective lens measurement is suitable for microscopic imaging system for its good restored effect and high computational efficiency, it takes only 1.62 s for the image size of 2048 pixel×1536 pixel.