Research Group Sensor Simulation
Developing realistic models of the image formation and imaging systems for the visible and infrared spectral range is at the heart of our research. Our main focus lies on long-range imaging through the atmosphere. We analyze and model the entire imaging chain – starting with the scene, through the atmosphere, optics, sensor and signal processing to the display and the observer.
Based on our knowledge and experience, we develop realistic simulations to evaluate the performance of electro-optical systems. Considered performance metrics range from laboratory measured quantities such as the signal-to-noise ratio, modulation transfer functions, and minimal temperature difference perceived over range performances to the simulation and evaluation of expected output images. Typical applications for our software are:
- analysis and evaluation of design concepts for electro-optical systems,
- assessment of the influence of environmental conditions and the scene on the electro-optical system performance,
- specification of the performance of electro-optical systems, e.g. by calculated detection, recognition and identification ranges.
Technically, we mainly work with three different approaches:
- analytical models, such as TRM4, which are based on linear system theory and an abstract description of the scene,
- imaging simulations, such as OSIS, which implement the physical effects along the imaging chain by a stepwise modification of a high-resolution input image of the scene,
- methods for the performance evaluation of electro-optical systems based on generated image data.
We also apply latest machine learning methods such as convolutional neural networks for our imaging simulations and approaches to image-based performance assessment.
In cooperation with the experimental research groups in the Optronics department, we check the predictions of our models and simulations against results from laboratory measurements and field tests.