Member Since 1984


AIA - Advancing Vision + Imaging has transformed into the Association for Advancing Automation, the leading global automation trade association of the vision + imaging, robotics, motion control, and industrial AI industries.

Content Filed Under:

Law Enforcement Law Enforcement


New Machine Vision Tech Proves Its Value at the Border

POSTED 05/11/2016

 | By: Winn Hardin, Contributing Editor

The challenge: secure borders with low-/no-fail technology that is also cost-effective for cash-conscious government agencies so it can be deployed at thousands of points, including airports, ocean terminals, border crossings, and rail junctions.

In short, law enforcement agencies need technologies to build “comprehensive situational awareness,” according to Xenics nv (Leuven, Belgium), a supplier of high-resolution infrared sensors and cameras. And, because challenges vary and potential threats change quickly, integrators need to develop systems “much faster and provide innovative solutions” for better public safety.

Machine vision technology is well-suited for these mission-critical roles. Pyramid Imaging (Tampa, Florida), for example, supplies and assists its customers developing systems that rely on machine vision to serve multiple border-security applications. One such system inspects rail cars for illegal riders and suspicious packages. Another reads all five fingerprints on an individual’s hand from a distance of 6 feet. A third imaging system supports wide-area surveillance to detect rocket fire or rifle fire and pinpoint its origin quickly.

“All of these use traditional machine vision hardware,” says Dr. Rex Lee, Pyramid Imaging’s president and CEO. “It really demonstrates how machine vision has become such a broad term.”

Dr. Lee also says that technology for border security purposes needs to be low cost and able to be deployed across “every port of ingress and egress.”

To help meet demand, industry standards are particularly important. Dr. Lee says that security customers acknowledge efforts to develop standards like GigE Vision and USB3 Vision. “All customers are concerned with supply chain logistics, so having multiple sources is important to them,” he says. Pyramid Imaging maintains multiple suppliers for cameras, lighting, and other components. If one supplier cannot keep up with demand, then a second firm can step in and deliver industry-standard equipment.

Another problem that Dr. Lee identifies is that some suppliers use lower-level staffers or even interns to compile a product data sheet. Integrators who use these materials for guidance often need to read them “with a grain of salt. We still encounter things like that all the time,” Dr. Lee says. The solution is for companies to work with knowledgeable integrators and distributors who know the product lines and can offer expert guidance.

Simplifying Design Cycles
Xenics says customers are asking for faster and more flexible technology innovation for border security and critical infrastructure applications. The company says one key to meeting demand is to shorten and simplify design cycles, and adopt modularity across specifications and functionalities. Doing so may better enable system designers to select from a range of components to meet application goals and create solutions.

To illustrate this strategy, Xenics developed a line that includes high-resolution infrared OEM modules. These IR cameras, called XenicsCores, are built on a modular, common FPGA platform. They can be tuned to achieve desired power dissipation levels and may be combined with sensors for short- and long-wave IR, a variety of lenses, and data interfaces such as GigE Vision, CameraLink, and CoaXPress, as well as analog video.

The company recently introduced an image-enhancement feature on its thermal-imaging products. Long wavelength infrared (LWIR) cores support image-enhancement filters that can be used along with histogram equalization. On the medium wavelength infrared (MWIR) core, the enhancement filter’s algorithm compresses an image’s dynamic range while also retaining small amplitude for details and a natural look. Details can be seen in images that have bright and dark areas and in thermal scenes that exhibit temperature differences.

The company says these embedded features relieve systems designers of the task of developing their own software-based specs to obtain good image quality. “Getting the visual information best suited to the application environment straight from the camera makes a newly configured system directly operable by security personnel,” Xenics says in a statement.

Focus on SWIR
The shortwave infrared (SWIR) part of the spectrum is of particular interest to Gooch & Housego (Ilminster, UK), manufacturers of high-quality precision optical components and systems, including precision lenses. The company has identified a market niche opportunity in SWIR applications, says Trevor Cook, European Business Development Manager.

SWIR imaging is effective in border security because it allows people to see through harsh conditions, such as rain, fog, dust, and darkness. Using SWIR, for example, an unmanned aerial vehicle (UAV or drone) can be guided remotely to land in bad weather or at night. The downside of SWIR sensors is that photons are scarce in this spectrum at dusk and in starlight conditions. Thanks to back thinning of the indium gallium arsenide (InGaAs) sensors, SWIR cameras can now detect photons from the red portion of the visible through the near infrared (NIR) and into the SWIR spectrum. Extending the sensor’s sensitivity into the visible gives the sensor better dynamic range thanks to improved signal-to-noise ratios.

According to Cook, the traditional approach to developing a lens that works from the visible through the SWIR spectrum is to re-coat and adapt existing visible lens designs for the SWIR band. This only works, however, for low-resolution cameras, he says. However, InGaAs focal plane arrays (FPAs) also are advancing in resolution while prices fall thanks to maturing processes and increasing volume sales. These factors mean that high-definition SWIR imaging is possible, but until now, optics development has lagged, meaning SWIR image quality is more likely to be limited by the lens and not the camera.

“Because the high-definition large-area sensors are more like 12-micron pixels, we have to design a lens that is 40-50% MTF at 50 line pairs per millimeter,” Cook says. “And moreover, we are trying to do that at the corners of the detector, so that’s a wide field of view.”

Cook says that lens coatings are critical. He points out that the visible band ranges from 400 to 700 nm, making it “fairly easy” to design a coating that has high transmission over the full bandwidth. But with the new SWIR sensors, lenses need to function across a much broader spectrum: 500 to 1700 nm. To solve the problem, Gooch & Housego focused on modifying the coatings.

“You can’t design a 99% coating from 500 to 1800 nm,” Cook says. “That’s impossible. The detector has maximum gain between 1000 and 1600 nm. So we tailored the coating to maximize the transmission from 500 to 850 nm to maximize transmission in that range. But should the customer need the optics optimized around 1500 nm — the eye-safe IR laser range for active illumination systems — then we can optimize for that part of the spectrum. It just depends on what the customer needs. That’s a common request because by timing the camera shutter to take a snapshot while an active illumination pulse is at a given distance, you can eliminate backscatter and get a much better image through fog, smoke, etc. Cameras optimized for 850 to 1100 nm can make a figure hiding in a camouflage suit stand out like a bright white light.”

Gooch & Housego’s new Barle 25-mm lenses for SWIR cameras also provide better performance over a wide temperature range and in harsh mechanical vibration/shock conditions, together with immunity to total immersion in water, in order to meet the needs of defense and border-security applications.

Unfortunately for world peace, the need for border protection isn’t going away. What seems more likely is that demand will remain strong for solutions aimed at enhancing border security. A big challenge is how to solve the problem in a cost-effective manner. “We’re probably not going to see fully open borders any longer,” says Pyramid Imaging’s Dr. Lee. “I think those days are gone.”

Embedded Vision This content is part of the Embedded Vision curated collection. To learn more about Embedded Vision, click here.