Proceedings Article | 24 April 2010
KEYWORDS: Sensors, Distortion, Image registration, Surveillance, Cameras, Data modeling, Staring arrays, Computer simulations, Image sensors, Image processing
With the advances in focal plane, electronics and memory storage technologies, wide area and persistence
surveillance capabilities have become a reality in airborne ISR. A WAS system offers many benefits in comparison with
the traditional airborne image capturing systems that provide little data overlap, both in terms of space and time. Unlike
a fix-mount surveillance camera, a persistence WAS system can be deployed anywhere as desired, although the platform
typically has to be in motion, say circling above an area of interest. Therefore, WAS is a perfect choice for surveillance
that can provide near real time capabilities such as change detection and target tracking. However, the performance of a
WAS system is still limited by the available technologies: the optics that control the field-of-view, the electronics and
mechanical subsystems that control the scanning, the focal plane data throughput, and the dynamics of the platform all
play key roles in the success of the system. It is therefore beneficial to develop a simulated version that can capture the
essence of the system, in order to help provide insights into the design of an optimized system. We describe an approach
to the simulation of a generic WAS system that allows focal plane layouts, scanning patterns, flight paths and platform
dynamics to be defined by a user. The system generates simulated image data of the area ground coverage from
reference databases (e.g. aerial imagery, and elevation data), based on the sensor model. The simulated data provides a
basis for further algorithm development, such as image stitching/mosaic, registration, and geolocation. We also discuss
an algorithm to extract the terrain elevation from the simulated data, and to compare that with the original DEM data.
