January 12th, 2012

whats needed for a machine vision camera to conduct license plate recognition?

Machine Vision Analysis

in Traffic Technology International January 2012

 

It's not every day that something moves out of a niche and enters the mainstream traffic market, yet that’s exactly what machine vision cameras have done. And the best thing, writes Louise Smyth in her Machine Vision Analysis, is that this might only be the start of things to come. In this article Adimec assesses whats needed for a machine vision camera to conduct license plate recognition.

 

Advances in digital camera technology enable more sophisticated intelligent traffic management and enforcement systems to better meet governments’ objectives to improve road safety, increase traffic flow, and enforce traffic laws with fair payments. The latest camera technology improves system performance to: minimize detection time in case of accidents, increase the detection rates in cases of violations, and minimize overall costs from police time and traffic jams. These systems also need to maintain reasonable costs and prove their efficacy to gain adoption by government decision makers.

 

License plate recognition (LPR) is one of the more common needs in traffic systems. There are several performance

and system parameters to consider which can be easily solved through the appropriate camera considerations.

LPR requires a crisp, clear image of a license plate to withstand dispute upon issuance of any penalties. This means

obtaining an image whenever a car appears, regardless of lighting conditions, weather conditions, the color

of the license plate, or the car speed. An IR flash light can be used to illuminate the license plate since it is not visible for people and the license plate material is reflective at 800nm. Cameras that have significant sensitivity at this wavelength allow for the desired image quality without a complicated and costly work around. 

Blooming and smear are challenges with outdoor vision systems. Bright spots can originate from headlights, reflections off license plates, the sun at certain times of the year, or sun reflecting on the tarmac. Image processing in the system cannot correct these defects so blooming and smear must be managed in the camera to ensure that the license plate is not obscured in the original image data.

 

Typically at least two good images need to be obtained on a car traveling up to 250km/h (155 mph). With a controlled test, an existing two megapixel traffic camera with a frame rate of 60 fps could still obtain a usable image for a car traveling at 360km/h. The fast frame rate needs to be also combined with high resolution. Highresolution cameras mean that fewer cameras can be used to see more of the road which simplifies the system and reduces the costs. Typically at least one megapixel resolution is required to limit a maximum of one camera per lane.

 

There are special considerations for the camera connectivity with regards to optical fit, electrical fit, and functionality, regardless of whether it’s a system upgrade or a brand new system.

Ghost images can appear if IR lighting is used in combination with a visible light block filter. The simplest way to prevent ghost images and lens artifacts from interfering with the system performance is to utilize a supplier that also has the expertise to properly integrate the filter and lens with the camera.

 

It is desirable to have the large, multilane processing part of the imaging system in a different location from the camera, such as next to the road rather than close to the cameras above each lane. This requires the use of rugged cables that can transmit large amounts of data over long distances. CoaXPress is the preferred interface that meets all of the requirements and allows for use of existing analog coax infrastructure. The specific functionality required depends on the system design. External triggering is important as images are required exactly when a car drives by, which is unpredictable. Since there is no second chance when the car is gone, effective automatic exposure control is often vital to have properly exposed images.

Adequate color processing with incorporated auto white balance ensures accurate representation of the colors, important for regions where license plates are different colors.

 

The needs of intelligent transportation systems are different from those of traditional machine vision, and therefore both markets have significantly different camera requirements. Utilizing cameras that are optimized for traffic applications can simplify the overall system design and increase performance while also reducing costs.

 

back to overview

 

How-to-compare-cameras