红外场景实时仿真研究

对目标和背景组成的场景的红外特性进行了实时仿真。为实现红外场景实时仿真，以场景的红外特性数据库为基础，将机器学习算法引入到仿真之中。利用遗传算法优化权值的神经网络建立了场景温度模型，结合三维可视化显示技术，模拟出不同设定条件下的场景红外图像。仿真结果显示，采用该模型的场景温度计算值和实测结果相符，并能够满足仿真对实时性的要求。

Real-time simulation of the IR scene

An IR scene which consists of targets and backgrounds is simulated in real-time. Based on the infrared feature database of the scene, a machine learning algorithm is applied to realize the IR scene real-time simulation. The temperature model of the scene is built up using the neural network whose weights are optimized via genetic algorithms. Combined with the 3D visual technique,the infrared scene images under different given circumstances are simulated. The simulation results show that the temperature values predicted by the scene using this model accords with the measured temperatures, and also fulfils the real-time requirement.