Nature of the Problem

Maintaining accurate situational awareness is a critical function for modern warfighters. The near ubiquitous use of video cameras has brought with it an increased demand for intelligent evaluation of the video imagery. For forward deployed soldiers, access to these rich source of situational information has been severely limited by several important factors:

  • Hardware limitations imposed when trying to access multiple streams of video data. Current portable computers simply lack to horsepower required for anything other than the most rudimentary use of the data.
  • Power requirements. Advanced vision processing capability including multiple image fusion, multispectral analysis, target identification and tracking, and so on requires high end hardware that has enormous power demands. For portable uses, such power demanding hardware would require carrying multiple batteries.

IEM WISE Solution

IEM addressed this need for a high end processing platform that could be adapted for use in a portable form factor, could be used to perform a variety of high end vision processing functions, and would operate at very low power. The Smart Video Module combines the power of multi-processor computer architecture into a small sized, low-weight, low-power package capable of independently processing two parallel video streams. Furthermore, the Smart Video Module may be integrated with a standard PC to provide specialized video analysis that can be offloaded from the PC’s generalized CPU to the specialized hardware thereby reducing bottlenecks in processing typical of generalized processors.

The Smart Video Module employs a standard single board computer form factor that allows it to be expanded in a variety of ways, from direct stacking of multiple devices to enable simulataneous processing of multiple video streams, integration with Windows software, and much more.

To support the Smart Video Module, IEM developed Slipstreams™ an integrated development environment that enables a programmer with reasonable knowledge of C programming and vision processing to generate high level machine code to enhance the processing capability of the module.

Illustrated here is an example of output from the SVM showing both a visible (upper left)and thermal (upper right) image being processed. The lower images show the results of specific image filtering performed on these video streams in real time.