Study On A Method Of Real-time Infrared Synthetic Image Generation Based On GPU

Real-time infrared synthetic image generation is the most important part of various infrared simulation systems. Hence, the study of real-time infrared synthetic image generation is of significance not only in theoretical research but also in practical application.A new method of real-time infrared synthetic image generation that based on the modern programmable GPU is proposed in this paper. This method fully utilizes the powerful processing power of modern GPU, and implements the infrared radiation calculation model directly on GPU, and gains high efficiency and high precision. Firstly, the content and the general process flow of real-time infrared synthetic image generation are discussed according to overseas synthetic image generation platforms in this paper, and two important key techniques, namely signature prediction model and radiation calculation model, are presented with emphasis. Secondly, the history of the GPU development and the features of the modern GPU are presented, the principles of using GPU for general computation are given out, and the feasibility of using GPU for real-time infrared synthetic image generation is discussed. Then, a framework of real-time infrared synthetic image generation is designed with reference to an overseas open-source three-dimension graphics engine. Finally, an initial implementation of the framework is carried out, and infrared images with high precision can be generated efficiently.