针对普通相机无法拍摄出符合视觉美感的背景虚化图像的问题，提出了一种基于相控高斯核的多聚焦图像景深模拟算法。对多聚焦图像进行热扩散方程的数学建模，提出自适应的初始化图像深度值来优化求解偏微分方程，通过正则化得到修正后的图像深度信息，分离出前景和背景；并提出一种基于相控高斯核的景深模拟算法来模拟大光圈相机得到的背景虚化效果。理论分析和实验表明，与已公开的算法相比，本算法能准确的提取出图像的深度信息，并模拟出真实的景深效果。

The visual beauty bokeh effect cannot be taken by non single-lens reflex camera, therefore a simulation algorithm for depth of field is presented based on multi-focus image depth information extraction. The mathematical model of heat diffusion equation is built for the multi-focus images. The adaptive depth values of initial image are proposed to optimize the solution of partial differential equation. The foreground and background are acquired by splitting the adjusted depth image obtained through regularization. Also, a simulation algorithm for depth of field based on phased Gaussian kernel is proposed to simulate the blur effect in the background caused by a wide aperture camera. Compared with other published algorithms, the results show that this algorithm can not only extract the depth information of the image accurately but also simulate the real depth of field effect.