Enhanced Realtime Millimeter Wave Imaging Using Hardware Acceleration

While the need for effective millimeter wave (mmW) imaging systems is clear, there remains a significant gap in the technology needed to realize it, namely the ability to process and enhance mmW imagery in a real-time manner. This is particularly important because raw mmW images appear very fuzzy, are color mapped, and have regions of intensity that are not normally associated with visible and IR images (such as metals being brighter than thermal objects). For this reason, significant processing is necessary in order to (1) improve the resolution of the captured image, (2) reconstruct the image in the case of a sparse aperture system, (3) compensate for motion during imaging (since most mmW imaging systems have relatively long integration times), and (4) provide adequate region screening and visual mapping. To overcome these limitations, we will develop a novel hardware accelerated processor specific for mmW imaging systems. In the Phase I effort we will survey possible techniques and analyze their suitability for being mapped onto a hardware accelerator. Initial implementations will be performed and preliminary results will be obtained on the mmW system currently being built at the University of Delaware.