Ctrl-P to Print
Adaptive Focal Plane Array (AFPA)
Program Manager: Dr. Stuart Horn
The overall goal of the Adaptive Focal Plane Array (AFPA) program is to demonstrate a high-performance focal plane array (FPA) that is widely tunable across the relevant wavebands in the infrared (IR) spectrum (including the short-wave infrared (SWIR), mid-wave infrared (MWIR), and long-wave infrared (LWIR) bands), thus enabling "multispectral imaging on a chip." The FPA will be electrically tunable on a pixel-by-pixel basis, thus enabling the real-time reconfiguration of the array to maximize either spectral coverage or spatial resolution. The goal is for the FPA to not simply be multi-functional, but rather to be adaptable by means of electronic control at each pixel. Thus, the FPA will serve as an intelligent front-end to an optoelectronic microsystem. The primary objective of the AFPA program is, therefore, to develop an electro-optical imaging sensor that benefits from both hyperspectral and forward looking infrared (FLIR) characteristics while avoiding the large mass of hyperspectral and the poor signal-to-clutter ratio (SCR) of FLIR. An infrared FPA is envisioned that employs an active filtering mechanism of some design in each pixel to provide spectral tuning and scene sampling with temporal and spatial simultaneity.
The intent of the AFPA program is to develop technology to demonstrate multispectral imaging (MSI) and possibly hyperspectral imaging (HSI) in a standard FLIR package. Currently, there are no imaging arrays that can be tuned on a pixel-by-pixel basis across the IR. This technology will enable staring FPAs that are capable of adaptively seeing through obscurants, imaging targets in diverse clutter, and performing precise chemical agent identification. Owing to its adaptivity, this technology will provide an essential sensing element to aided target recognition (ATR) systems.
The military has multiple needs for day and night surveillance and reconnaissance. The ultimate goal of an electro-optical imaging sensor is to achieve a high probability of detection while simultaneously maintaining an acceptable low false alarm rate. The most challenging target detection problem is a stationary target in a cluttered environment where the target-to-mean background signal differential is small compared to the fluctuation in the background. The baseline LWIR scenario is characteristic of well-camouflaged targets under foliage. Modern FLIR imaging sensors can achieve high detection, and low false alarm rates through the exploitation of the very high spatial resolution available on current generation large format focal plane arrays. However, such broadband MWIR and LWIR sensors are limited in their ability to search and detect camouflaged targets in deep-hide or in the presence of significant clutter.
A new generation of electro-optical imaging sensors is required, therefore, to fulfill a wide variety of military mission needs in which targets in deep hide are expected. A proven approach is to exploit the spectral content of targets through a multi- or hyper-spectral imaging scheme. The processing of spectral content improves the SCR available by utilizing the differences in spectral energy content of target and background radiation. However, today's hyperspectral imaging systems are limited to scanning mode detection. The ability to perform multi- to hyper-spectral imaging in a staring mode is critical to targeting missions.
To date, hyperspectral data collections have demonstrated SCR gain; however, researchers have not identified any unique bands of interest, making fixed-band multispectral systems undesirable. With continuous spectral tuning, the warfighter may preprogram the AFPA based upon prior scenario hyperspectral data. A limited number, but intelligent, choice of bands would sharply reduce the amount of data to be processed allowing real-time video display of false color multispectral imagery. Hyperspectral imaging systems currently under development are large, complex, power hungry, and computational intensive systems. They also suffer from poor performance when trying to detect targets that are in rapid motion relative to the observer platform. Airborne hyperspectral sensors are massive, about 50Kg systems, which are typically flown on large non-combat platforms. These units require greater sensitivity than typical FLIR sensors in order to overcome the reduced photon count in narrow wavelength bands. Higher sensitivity is achieved by larger apertures, about 0.3m, large pixel cell, about 100mm, and slower operating speeds, between 1 and 10Hz, as well as low detector temperature, about 10K, to counter dark current. Typical LWIR hyperspectral sensors have wavelength band full-width-at-half-maximum (FWHM) equal to 0.05mm, and a spectral range spanning the LWIR atmospheric window. These systems operate in a pushbroom mode that sweeps a linear format over the scene, gathering all spectral information simultaneously.
Typical FLIR avionics employ stabilized optical systems and cryogenically cooled detectors in order to support tactical target recognition ranges of about 5km depending on system F/#. These systems have mass about 10 Kg and are employed in tactical aircraft and other weapons platforms for the purpose of target ID and tracking. The pixel unit cell is between 25 to 50mm and the system NETD is about 20mK for F/2 optics. Frame rate is typically 60Hz to support target-tracking algorithms and detector temperature is 77K. Despite low noise performance of these devices, realistic tactical scenarios exhibit background variation, or clutter, that generate target-like objects that are confused with real targets.
New mission requirements, which include a variety of distributed remote sensor platforms, drive the need for smaller, lighter weight imaging sensors, but with more capability than the current generation. Multiple platforms, including ground sensor networks, robotics and micro-air vehicles will collect data for integration into global information databases. These platforms must obtain information from wide search and difficult to access areas, and address hidden and camouflaged targets. The imaging sensors must also continue to function in difficult environmental and atmospheric conditions.
These new mission requirements form the motivation for new reconfigurable, or adaptive, imaging micro-systems, which will meet the conflicting requirements for large area search coupled with the ability to detect and identify difficult and hidden targets, while staying within the processing volume and size available in small platforms. A capability to detect information in multiple spectral bands will greatly facilitate target discrimination at the focal plane. Thermal contrast reversals, camouflage matched to a particular background, and the variety of environmental conditions worldwide present significant issues to single band sensors. Without a means of multi-spectral band discrimination, there exists significant risk that targets will be missed and threats to both people and vehicles will not be detected. Also, an increased emphasis on focal plane processing, where target discrimination is integrated into the detector / focal plane, can enable very low power imaging systems, which greatly reduce the raw data sent to the analog to digital converter. Therefore, the fundamental objective of the AFPA program is to solve these challenges by developing FPA technology that exploits both FLIR and MSI/HSI capabilities into a single electro-optical imaging microsystem.