为了拓展红外图像获取方式，提出了一种基于可见光衍生生成红外图像的方法。利用脉冲耦合神经网络与人工微调相结合的方式，将可见光图像分割成不同区域并标定为不同材质，通过材质热特征建模预测不同材质在不同环境中的温度值，并计算其热辐射以及大气传输效应，最终通过细节调制的方法增加了仿真结果的真实性。实验表明，通过可见光图像生成红外图像有其可行性，并且图像分割结果直接影响了仿真结果的好坏，分割越精细仿真结果越好，利用细节调制可增强仿真结果的真实性。

To expand the infrared image acquisition method, a method of infrared image generation based on visible light image is proposed in this paper. Visible light image is segmented into different regions, which are calibrated into different materials, by combining the method of pulse coupled neural networks and artificial tuning. The temperature of different materials is obtained by the model establishing for thermal characteristics in different environments, then we calculate the heat radiation and atmospheric transmission effect. The authenticity of the simulation is increased through the method of the details modulation. Results show that it is feasible to generate infrared images by visible light images and the simulation result is directly affected by the segmentation of image. The finer the segmentation, the better the simulation results by using the details of modulation we can enhance the authenticity of simulation results.