Video processing on the fly
Video – in common with other sensor-derived data – plays an increasing role in today’s military. Its proliferation at all levels of command reflects a deep hunger on the part of decision makers for remotely collected imagery that helps them see threats and, if necessary, deploy soldiers and weapons. This data is all the more important in maximizing the effectiveness and minimizing the vulnerability of forces as troop levels decrease in theaters of war and, in the future, as the overall footprint changes. As such, the demand for image processing solutions is widely expected to increase.
This market development is all the more likely as potential uses for video multiply in number and expand in scope. The National Aeronautics and Space Administration (NASA), for example, already uses captured video to monitor wildfires and hurricanes. The Department of Homeland Security (DHS) uses it for border surveillance. And local jurisdictions want to use vehicle-captured video in law enforcement and public safety roles.
Increasingly, military vehicles are being equipped with high-definition video cameras and sometimes with multiple sensors and sensor types per vehicle. Imagers on legacy platforms, meanwhile, are being upgraded to provide higher-resolution data. But how can users handle this torrent of image data pouring into control centers hundreds or thousands of miles away via satellite or other links that may be much more limited in bandwidth than the data source?
In addition to the volume of data to be uploaded, compressed, manipulated, stored, or transmitted, the remoteness of video capture platforms puts a premium on timeliness. To address threats and other intelligence within tactical timelines, information must be presented with minimal latencies. This means the image processor needs to be flexible enough to accommodate multiple data streams and data formats simultaneously and adjust output bandwidth on the fly.
Technology to the rescue
Fortunately, image processing technology has evolved to keep pace with data volume. Small-format boards can handle multiple data streams, bandwidths and storage, or transmission tasks. They also can format the data according to multiple standards, convert formats on the fly, and vary compression ratios according to the limits of particular data links. Reconfiguration can be performed in real time via radio frequency link or remote commands to the host processing boards. This type of flexibility is key to military customers. Further, while multiple vendors supply image processing hardware, the most desirable solutions optimize flexibility in all areas of performance, from packaging to remote configurability and data compression.
An image processor’s ability to modify compression parameters on the fly enables users to maximize video quality within the limits of available bandwidth. In critical phases of a mission, higher-capacity, line-of-sight data links can be used for vehicle safety. But when the vehicle is in theater within an area of interest, operators can limit its output bandwidth to suit the chosen data link. The image processor would then maximize video quality within those constraints by intelligently managing the data compression ratio.
One means of satisfying the hunger for optimized, manifold video processing is to utilize video streaming, compression, and capture boards built to provide a diverse range of video processing capability. One example is the GE Intelligent Platforms ICS-8580, a video processing board in an XMC form factor with high-speed PCI Express connectors that can be plugged into a single board computer in a larger system or function as a ruggedized stand-alone product (Figure 1). Such a board’s architecture can accommodate multiple data streams simultaneously. One card can process: two channels of high-definition video (1,920 x 1,080 pixels); one channel of high-definition and two channels of standard-definition (720 x 480 pixels) video; or four standard-definition video channels. Data compression ratios are configurable and adaptable but range as high as 150:1 with 75 ms latency from input to output for full high-definition video. By comparison, JPEG 2000 compresses data at a 30:1 ratio, which could overload a narrowband data link.
Video processing: Flexible technology is key
Whether it’s a question of the packaging, multidata stream capability, compression rate adaptability, or support for multiple camera input and output types, the video processing platform’s flexibility is the key to image processing performance. This ensures not only that the disparate needs of military users can be met, but that new video technologies can be adopted as they become mainstream.