AIA Logo

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:



Machine Vision Takes Driver Safety Seriously

POSTED 11/05/2013

 | By: Winn Hardin, Contributing Editor

More than a million and a quarter people die in road crashes every year. And this number is projected to increase to 2.4 million by 2030, which would make traffic fatalities the fifth leading cause of death. Injuries sustained in crashes are already one of the top three causes of death for people between the ages of 5 and 44.

And yet enforcement of traffic laws that could reduce road crashes remains poor in many countries around the world—indeed dismal, according to a report by the United Nations Road Safety Collaboration. Asked to self-rate their law enforcement performance on a scale of 1 to 10, fewer than 10 percent of countries reported enforcement of speed limits as over 7, only 13 percent reported enforcement of blood alcohol concentration limits as over 7, and so on.

Machine vision companies have developed impressive technologies to help with enforcement of traffic laws, and they continue to roll out new systems with additional capabilities and improved functionalities. Among them: Imaging Solutions Group's (ISG) red light infraction cameras, with 16 MP and 29 MP resolutions and the ability to capture up to 8 lines of traffic at a time; ADLINK’s MXE 5300 Matrix system, with its embedded computing platform and rugged housing designed for temperature endurance, as well as shock and vibration resistance; and VITRONIC’s LIDAR-based PoliScan system.

But technology development isn’t the only challenge that companies face in serving the needs of the traffic law enforcement market. Even as they are introducing technologies that can help to reduce traffic fatalities worldwide, they are navigating a range of other additional concerns.

In the U.S., machine vision’s primary contribution to traffic law enforcement to date has been in the area of red light enforcement. For example, ISG’s red light infraction cameras are placed at intersections and are either hooked into the traffic light or have other sensors to indicate an oncoming vehicle. The cameras continually acquire images at a rate of 3 to 8 frames per second at a very high resolution using a Truesense Imaging KAI-16070 or KAI 29050 CCD. They can also capture an NTSC signal from another camera and sync these low-resolution images to the high-resolution video stream.

Installations of red light infraction cameras continue in the U.S., though the pace has slowed somewhat because of public backlash. Motorists complain that use of the cameras is unconstitutional as it shifts the burden of proof to drivers, who are left to demonstrate somehow that they are innocent. They also argue that these cameras don’t increase safety, but rather serve only to line the pockets of local governments. Mostly, though, they just don’t like the idea of someone being able to watch their every move.

Research has shown that the general public does not like the cameras. But the same studies indicate that people who live in the vicinity of the cameras learn to appreciate the increased safety that comes following their installation.

“Studies show that after a few months of installation, traffic infractions decline as the locals know where they are and make sure they slow down and do not run red lights,” said Kerry Van Iseghem, Co-Founder of Imaging Solutions Group (Fairport, New York). “The tickets issued after these first few months are limited to outsiders who blatantly disregard the traffic signals, and the locals want them stopped and ticketed.”

Questions of privacy have arisen around the capture of license plates generally, with cameras installed on overpasses and in police vehicles, for example. A police officer might be snapping pictures of license plates in oncoming traffic or while perusing a parking lot and running them against a database to see if the car is reported as stolen or if there is a warrant out for the owner.

This has raised concerns about the use of such cameras to track people who haven’t done anything wrong, to keep records of where they are and where they are going. But this isn’t the case, says Bradley Weber, Director of Application Engineering at Datalogic Machine Vision Business Unit, based in Bologna, Italy. Weber thought quite a bit about these things while working with Image Sensing Systems, a company that produces machine vision systems for advanced traffic management.

Weber emphasizes the difference between cameras used primarily for traffic flow analysis and those used to match license plates against “red flag” databases of recalcitrant traffic offenders. License plate numbers are never saved with the former, he notes. And even with the latter, plates not tied to known offenders are discarded.

Law enforcement is “very careful about privacy issues,” Weber says. “They compare the plate number to a database and, if it doesn’t match, it’s gone. There’s no record of where people are at different times.”

Enforcement around the world
The U.S. doesn’t have a monopoly on unruly drivers, of course. The urge to drive like the proverbial bat out of hell is innately human, and the need for traffic law enforcement therefore universal.

Wiesbaden, Germany-based VITRONIC has a growing enforcement presence worldwide, with installations throughout Europe, the U.S., and Australia. In fact most of the company’s enforcement activities are taking place in the Middle East, according to Daniel Scholz, VITRONIC’s Sales Director, Traffic Technology. It has an office in Dubai and serves countries throughout the region, primarily the United Arab Emirates and Saudi Arabia. Here, VITRONIC’s LIDAR-based PoliScan system is used for speed and red light enforcement, and increasingly for average speed enforcement (P2P - point to point).

Why such demand for the technology in the Middle East? “My read is they’ve been shocked by the level of death and injury that’s been occurring on their extremely well-built transportation network,” says Steve Tuttle, Business Development Manager Traffic Technology at VITRONIC, noting the “dangerous mix” of locals zipping around in high-performance sports cars and working people from uneducated driving countries like Bangladesh and India driving like they are still at home.

The region has the money to build outstanding roads, Tuttle says, “but they need to balance this by having some sort of control over driver behavior. Because they really do drive very quickly over there.”

The Middle East is an important market for VITRONIC, but the company is focusing on other regions as well. At the time of publication, they were about to receive official state certification for China, and they were also making inroads into other parts of Asia. Every market is different, though; each one comes with its own set of challenges, which can determine which technologies are implemented and how. In Australia, for example, a front-facing driver photo is admissible in court with speeding and for red-light infringements; however, the image of the driver’s face is not used for identification.

All of which can impact the decisions governments and other customers make when seeking to implement enforcement technology and any sort of program seeking to change driver behavior, says Tuttle. “They have to consider these sorts of things in parallel with what we as manufacturers can bring to the table.”

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