Development of an image-based electro-optical system performance prediction tool for design and optimization

Electro-Optical (EO) systems are designed for purposes such as detection/recognition/identification and tracking of object(s). In order to design the systems in an optimum manner, there are many processes involved in generation of the images of targets and background at the system detector output, and these components should be carefully examined for various conditions. The image chain starts with a ray originating from target space (object space) and after propagating through atmosphere and EO system; the final position of this wave is the focal plane (image plane) where the detector is placed. EO system designs require optimization of many different system parameters for a given task; therefore, there is a need for an end-to-end imaging system simulator which models cascaded image chain blocks from object space to the detector output. An image-based system performance prediction tool has been developed for generating synthetic data in order to be used for estimation and design/optimization of EO system performance. This paper introduces this image-based performance prediction tool/scene generator in a system designer point of view, and demonstrates some properties of the tool which may be useful for system analyzers/designers for optimization. The synthetic scenes can be generated either via parametric models and/or radiometric measurements for EO system, environment, and object signature. Also, this tool has a user-friendly graphical user interface (GUI) which takes either measurement and/or system/environmental/object space parameters as inputs. The user can observe/obtain the output raw images/videos together with various system design parameters as well as image degrading effects such as modulation transfer function (MTF) and noise. In addition, this tool can be used for generating synthetic data via constructing a big data set for traditional and learning based algorithms.