DARPA has awarded Voxtel with a $5.2 million contract for the first phase of its Reconfigurable Imaging (ReImagine) Program. Voxtel is developing a multi-mode fused image sensor that integrates short-wavelength infrared (SWIR) passive imaging and laser detection and radar (LADAR) sensing in the same imager.
Similar to a field-programmable gate array (FPGA) processor, the field-programmable imaging array (FPIA) architecture that Voxtel is developing in the ReImagine program would incorporate dynamically reconfigurable pixel elements, so that separate regions of the focal plane array may run separately at high resolution, at high frame rate, or in active time-of-flight mode, enabling on-focal-plane object detection and extraction, along with pixel-registered sensor fusion. By designing multi-mode passive- and active-imaging algorithms implemented at the pixel level with reconfigurable on-focal-plane processing, the technology would allow autonomous object detection and imager configuration, with efficient extraction of relevant information.
According to Voxtel, their multi-mode fused image sensor would greatly benefit applications that require significant laser power, like LADAR-assisted automotive navigation. By using laser pulse energy only where the scene requires it, the company can make systems that are smaller, less-expensive, more efficient at scene extraction, and require less processing.
DARPA’s ReImagine program seeks to develop a software-reconfigurable, multi-mode imager that can adapt to different conditions in a scene and autonomously reconfigure itself to collect the most valuable information in the scene.
An approximately three-year contract under the auspices of the ReImagine program was awarded to Voxtel in May, 2017. In addition to use in autonomous automobile navigation and driver assistance systems, military applications would include multi-mode imaging for helmet-mounted, weapon-mounted, and unmanned-vehicle systems.