Inside the tiny footprint of Tech Square’s HPERC lives the power of Intel® Xeon® processor and optional GPGPU parallel processing engine based on a NVIDIA Quadro MXM module. Dual removable secure erase RAID-0 SSDs provide 12Gb/s throughput and security for deployment in hostile environments.
Easy configuration and expansion allows for fast integration of custom rugged embedded applications. Uniquely-keyed MIL-DTL-38999 connectors provide a wide array of fast IO. The HPERC reserves connector pins for expansion interfaces. The GPGPU sits on a 16-lane 3rd generation PCI Express interface to enable image processing and other applications driven by AI and Deep Learning technologies. Built with proven Ampro (acquired by Tech Square in 2008) reliability, the rugged capabilities of the HPERC are ideal for ground, air and sea deployments in military and aerospace. The HPERC systems survive shock, resist corrosion and galvanic oxidation, and perform in the humid jungle, the heat of the desert, and the bitter cold of frozen mountain passes.
In addition to Intel’s constantly evolving roadmap, NVIDIA’s many-core GPUs have transformed some areas of sensor processing, including radar, wide-area surveillance, and hyperspectral imaging, due the inherent suitability of parallel processing of those data sets. While this parallel processing paradigm has rapidly increased the processing capability within a given SWaP envelope, the real gains are yet to be exploited – the advent of Deep Learning will likely power the next wave of systems capabilities.
Image processing methods perform operations to enhance an image and/or extract useful information from it. Applications include pattern of life analysis, surveillance, reconnaissance, target identification, and geolocation in GPS-denied areas. For example, to collect tens of thousands of hours of aerial surveillance video and other theaters of operation. Turning what are often grainy, shaky videos taken by drones or satellites into clean image streams suitable for analysis — especially real-time analysis — poses a massive computational burden. According to NVIDIA and Motion DSP, GPUs can process such video seven times faster than CPU-only approaches. Coupled with NVIDIA’s CUDA graphics processing and the OpenGL API, Tech Square’s HPERC, provides solution engineers with a powerful COTS (commercial off-the-shelf) sub-system with which they can design the next generation of image processing systems. More applications, including:
● ISR platforms (intelligence, surveillance and reconnaissance)
● Radar processing
● Software-defined radio (SDR)
● Sonar processing
● Forward looking infrared radiometry (FLIR)