ANDREW OWEN
Infrared imaging systems visualize heat emitted by objects in a viewed scene. Objects stand out because they are warmer or colder than the background they are viewed against. This contrast is the thermal image seen on a display.
When using thermal imaging systems for surveillance, subjects under observation can include people and objects of interest such as aircraft and ground vehicles. Infrared imagers have played an important role in surveillance missions for both military and law-enforcement purposes for more than 20 years. The need for accurate, real-time information, 24 hours a day in degraded environmental conditions is essential to the modern military command decision-making process and can make the difference between mission success and failure.
The use of IR imagers in surveillance has doubled over the last five years. Factors such as the development of IR technologies that have provided the basis for new, lower-cost products; manufacturing advances permitting production of these lower-cost, high-performance imagers; and the acceptance of IR technology on a wide scale in the law-enforcement community in the USA and abroad contribute to this increase.
Range game
Perhaps the most commonly asked question with regard to an IR imager is "What is the effective operating range?" Unfortunately, unless one is willing to use rough approximations, there is no simple answer short of running a performance analysis such as the standardized US Army Forward Looking Infrared 92 and acquire models to predict the actual results with an imager. These models take into account all variables such as the imager performance and design characteristics, atmospheric conditions, and target size and condition—allowing system performance specifications to be properly written. This approach to system performance is based upon the concept that the probability of target recognition is a function of the number of cycles of a target-equivalent bar pattern (a four-bar pattern with 7:1 aspect ratio) that can be resolved across the minimum target dimension. The bar pattern will have the same temperature difference relative to the background target in question. The number of resolution cycles a particular imager is able to put across the target`s critical dimension (typically the square root of the target area) is determined by the imager minimum resolvable temperature difference (MRTD) performance curve and imager spatial resolution (see Fig. 1).
The ability of an operator to see a target is divided into three levels of discrimination: detection, recognition, and identification. Detection occurs when the observer sees something of interest, generally a hot spot generated with only a few imager pixels on the target. Even with only a few pixels on the target, cues such as movement and relative size can help the operator classify the target.
Recognition is the next order of discrimination—the operator is able to recognize a particular target among a group of similar targets. An example of this is when the operator is able to distinguish wheeled vehicles from tracked vehicles that are together in a group. Identification is the highest order of target discrimination and occurs when the operator is able to distinguish between individual targets in a set and specifically identify them.
The effective operating range of an imager is also determined by factors such as display resolution and operator training. If an imager provides 400 active IR lines, the display monitor must provide at least an equivalent number of resolution lines or it will become the limiting factor. High-quality, ruggedized monochrome monitors that provide 750 resolution lines are readily available at reasonable cost. Color monitors with good resolution are also available, but there are fewer models to choose from, and they are more expensive.
Operator training is often overlooked when considering range performance. Operators that have a thorough understanding of specific target thermal signatures will be able to recognize targets at far greater ranges than those lacking that level of training and experience. In some cases, this performance advantage realized through proper training can add an additional 20% in effective range.
An operator engaged in surveillance of a large area must be able to view the entire scene in a timely manner and be able to pick out targets from the background clutter. Fatigue is often a problem as, quite often, the operator views the IR image for long periods of time. To reduce fatigue and eliminate the need to constantly discriminate between the background clutter and possible targets of interest, imager gain and level controls are adjusted to eliminate all but the hottest features in the scene. When adjusted properly, much of the background detail is suppressed in the image. When a figure or vehicle target enters the scene, it will immediately stand out against the mostly dark, uniform background. Once the target is detected, the gain and level can be readjusted to provide more thermal detail in the viewed scene.
Long-range surveillance imager features typically include remote-control capability; a cryogenically cooled detector array; dual field-of-view (FOV) optics, around 8° for target detection and a narrow FOV around 2° for target recognition; portability; a TV-compatible display; good thermal sensitivity over a wide range of operating temperatures, expressed as a net equivalent temperature difference (NETD) to the background of 0.2°C or better for a 25°C background; modest power consumption; and a robust, environmentally sealed design.
Imager characteristics
A wide variety of thermal imagers are available today to satisfy nearly any surveillance-mission requirement. The well-established thermal-imager manufacturing base can provide products for airborne and ground-based applications, whether short- or long-range surveillance.
First-generation thermal imagers have an optical scanning mechanism to scan the target scene across a small number of photoconductive detector elements—up to 180 elements are arranged in a linear or staggered linear pattern. These systems operate in the 3-5- or 8-12-µm spectrum and can be cooled with Joule-Thomson cryostats, Stirling-cycle, or thermoelectric coolers. Thousands of these first-generation systems were produced over the past two decades, and most are still in service. Many of these systems were based on the US Army`s 60-element common-module detector/ Dewar and included thermal antitank missile-sighting systems.
The IRTV-445L LORIS, a long-range infrared system manufactured by Inframetrics (North Billerica, MA), embodies several improvements over the original imagers: it has only 15-W power consumption, a galvanometer-based scanning system for standard TV output, and electro-optical zoom capability that allows the operator to increase system thermal sensitivity while magnifying the image two or four times.
In late 1995, the US Immigration and Naturalization Service contracted Inframetrics to supply 500 IRTV-445L systems in a broad effort to crack down on illegal immigration. These imagers are installed on mobile and fixed platforms to detect individuals illegally crossing the border at ranges in excess of several miles. Early detection of border incursions facilitates rapid apprehension of illegal entrants and enhances the safety of the agents. Using a long-range LORIS, border-patrol agents can determine if the suspect individuals are carrying objects and distinguish between males, females, and children within a group (see photo at top of this page).
The technology adopted by the US Army uses cryogenically cooled, scanned photovoltaic mercury cadmium telluride (HgCdTe) focal-plane arrays (FPAs) operating in the long-wavelength IR band. These standard FPA/Dewar configurations use 480 × 4 FPA readout architectures. Although these systems use an electro-optical scanner, they provide very high thermal sensitivities by using time-delay-integration readouts and produce high-resolution thermal imagery much im proved over first- generation scanned systems.
The devices are designed primarily for high-end reconnaissance and weapons platforms such as the Bradley Fighting Vehicle. Scaled-down versions are used in programs such as the shoulder-fired antitank Javelin missile system. In the future, these systems will also be mounted on fixed- and rotary-wing aircraft.
There are systems that no longer need an electro-optical scanner—a two-dimensional, staring FPA does the imaging. Cryogenically cooled systems that operate in the 3-5-µm range using platinum silicide (PtSi), HgCdTe, or indium antimonide (InSb) FPAs have been available for several years now. These systems are usually smaller, lighter, and less expensive than second-generation scanned systems, yet still provide high-end performance. The FPAs are typically composed of 256 × 256 or 320 × 240 pixels and exhibit a high degree of uniformity and operability.
As FPA manufacturing technology has improved, readout and correction circuitry has become smaller and requires less power. Cameras are being built with FPA uniformity gain and offset correction factors stored in permanent memory. And field nonuniformity correction is accomplished by simply pressing a button on the camera for several seconds. This single-point update corrects for drifts in offset and improves spatial performance against changing backgrounds.
Variety in end products
Although many camera manufacturers start with the same size FPA, there can be several important differences in the end product, including FPA fill factor, or active pixel density—higher being better. The camera should have NETD stability over a range of operating temperatures. A selection of lenses will allow both short and long stand-off ranges. And robust packaging, light weight, and low power consumption enhance hand-held operator interface options.
For example, a follow-on to the hand-held 256 × 256 PtSi-based InfraCAM introduced in 1993 is the Inframetrics ruggedized MilCAM. Qualified to military environmental specifications, the camera is available in PtSi and InSb versions. It consumes less than 6 W of battery power and features an integral viewfinder, a family of sealed lenses, programmable reticle, and video overlay allowing RS-232 inputs from devices such as laser rangefinders and satellite-positioning units. The MilCAM uses a modular design and can be outfitted with a wide-angle lens for hand-held short-to-medium-range surveillance applications or a dual FOV lens for long-range, remotely controlled operations.
In addition to cryogenically cooled staring FPA cameras, designs utilizing "uncooled" FPAs of 320 × 240 pixels operating in the long-IR band are now emerging. These cameras use a solid-state thermoelectric cooler to maintain the FPA at or near room temperature. The devices are actually thermal detectors and experience a temperature change in response to incident infrared radiation. For this reason, their operating temperature must be precisely maintained to accurately measure their infrared-induced temperature rise.
Elimination of the cryogenic cooler provides the basis for dramatic cost reductions and a corresponding increase in system reliability. Such cameras provide decent sensitivities in normal operating environments and are useful for short-range surveillance missions. Much work is still underway to improve sensitivity, in crease FPA fill factor, and reduce pixel size. The current cameras operate with f/0.75 optics and lens focal lengths around 50 mm, so they are easily hand-held for surveillance out to 500 m.
Surveillance organizations in the market for thermal imagers thus have a wide range of products to choose from. The first decision should be which imager characteristics are most important for specific applications. Selection of an imager with the proper range performance, ability to operate over a wide range of temperatures, design modularity, and field worthiness are keys to mission success.
Andrew Owen is manager of ground surveillance systems at Inframetrics, 16 Esquire Road, North Billerica, MA 01862-2598.