Infrared Image Simulation of Building Scene Based on SVM

Real-time infrared image simulation of building scene based on support vector machine (SVM) is presented. The parameters that influence object surface temperature field are analyzed and set as the input vector of SVM, the temperatures of object surface apartments are got by infrared imaging temperature measurement and set as output target of SVM. After training the learning machines by the training samples which consist of input vector and target set, support vector machines have the ability to forecast the temperature field with simulation parameters. Real-time rendering of building scene infrared images is realized by three-dimension visualization. The result shows this method has a quality of maneuverability, and high simulation precision.