Description: Space-based imaging sensors are important for NASA's mission in both performing scientific measurements and producing literature and documentary cinema. The recent proliferation of high-definition capture devices and displays (HDTV) provide the general public with first-hand human experiences hundreds miles above sea level in brilliant detail. The recent IMAX film "Hubble," which features one of the final space shuttle missions to repair the orbital telescope, is a prime example. The core of current space-based video capture devices consist of digital imaging sensors. Unfortunately, the harsh conditions of space limit the lifespan of all the imaging sensors, in addition to other electronics. Consequently, NASA is seeking innovative technologies for space-based applications to extend the operational life of these systems to three years or more. In this SBIR project, we propose to investigate robust image reconstruction based on novel signal processing techniques in the vein of compressed sensing (CS) to mitigate pixel damage to the point that is imperceptible by the human eye. Specifically, this proposal is a response to the solicitation for radiation-hardened programmable encoding technology as an identified mid-term NASA solution. CS is a recently introduced novel framework that goes against the traditional data acquisition paradigm. CS demonstrates that a sparse, or compressible, signal can be acquired using a low rate acquisition process that projects the signal onto a small set of vectors incoherent with the sparsity basis. This approach is divided into encoder and decoder stages. We propose performing the encoding in-line with acquisition using a low-SWaP, radiation-tolerant FPGA. The robust reconstruction will occur back on Earth where high-performance GPU-accelerated workstations can be used. A benefit of our solution is that it does not require a modification to the original imaging system.

Benefits: In order to harness the potential of this technology in the military arena we plan to team with prime contractors, in particular Lockheed Martin Space Systems, Lockheed Martin Maritime Systems and Sensors and L-3 communications. We have worked with these companies in the past, transitioning SBIR technology into military programs of record. Regardless, the novel algorithms developed in this SBIR can be implemented in alternative low-cost platforms like CPUs and GPUs, enabling the same benefits in post-processing fashion. Once the cost of the technology is low enough, the applications of dead pixel mitigation algorithms are virtually limitless in the commercial imaging arena: TV broadcasting, medical imaging, digital photography, industrial inspection, robotics, surveillance, etc.

Although the primary application is for space missions involving the acquisition of HD video, NASA will be able to leverage our solution as a technique for mitigating the effects of sensor pixel damage in video imaging for space, air, land or maritime sensors. The dead-pixel problem is of particular importance for night-vision devices mounted in high vibration/dust environments, like aircraft, road-going vehicles and man-portable equipment. In some of these applications the bandwidth is available to process the raw data away from the sensor, enabling cost reductions.