For design, testing and optimization of infrared systems, generation of big physics based synthetic data is very important, since it is impossible to collect data with experiments only. In order to create such a big radiometric data, an end-to-end synthetic simulation approach is so useful. For generating radiometric data in imaging chain from object space to the system detector output considering environmental and system effects; the parameter space in rendering pipeline can be scanned throughout imaging chain. Therefore, target, environment, electro-optical/infrared system parameter space and related radiometric data outputs of the sensor construct the big physics based data all together. Also, relative motion between the observed object and the observer is another source of physical data. In this paper, the main components and parameter space of the radiometric data are described and some example complex background scene outputs which are generated with 3D rendering are demonstrated. In addition, results of a laboratory (experimental) validation effort are discussed, which show the validity of the mathematical approach applied in generation of the radiometric data. These laboratory experiments show that when the inputs are correctly defined, the data can be generated very close to real measurements, i.e. physical reality can be synthetically generated at acceptable levels of error.
|