The Guide “10 Tips For Best Free Flow ANPR Deployment” aims to provide a useful list of characteristics and features of ANPR cameras to be used in free-flow, auto-triggered and high-speed traffic control applications.

Capturing the vehicle and an image of the license plate in an exact and deterministic point of the road is essential to guarantee the target resolution of 2,7 pixels/cm for the license plate, even when a megapixel sensor is used (see Tip #1).

A great number of auto-triggered ANPR cameras use image sensors at low to medium frame rate speeds: capturing between 15 to 30 images per second (or FPS, frames per second) in a field of view of around 5 to 10 meters (virtual detection loop). This has at least one important consequence for the ANPR performance: the resolution of the license plate image decreases as the vehicle travels away from the camera. If the vehicle’s plate is captured at the shortest distance of the camera’s field of view, then maximum resolution of the license plate image will be attained. On the contrary, if the vehicle is detected when it is at the limit of the virtual loop (longest distance from the camera), then the resolution of the plate will be lower (less pixels per centimetre).

An OCR engine will underperform on a plate image with low resolution (as explained in Tip #1).

Furthermore, a non precise auto-triggered ANPR camera will impact the precision of a free-flow travel time or average speed application. There would be no certainty on where (position, km point) the vehicle is captured and thus it would create an error on the speed measurement.

Instead of working on a virtual detection loop in a range of 5 to 10 meters, Imagsa's ANPR cameras define a virtual loop in the lowest part of the overview image (a region of interest of 1,5m - see video). This is the closest point to the camera. In this region, 250 images per second are captured and analysed in real-time thanks to the power and intelligence of the Mercury’Traffic ultra-fast smart camera. Thus, the vehicles’ plates are always detected at the same point, with the same target resolution of 2,7 pix/cm. An optimum plate resolution and a good illumination of the plate image ensure that the OCR engine always achieves high recognition rates.

Additionally, Atalaya’s highly precise auto-trigger is key to achieve great precision on a free-flow (no externally triggered) average speed enforcement solution.

The Guide “10 Tips For Best Free Flow ANPR Deployment” aims to provide a useful list of characteristics and features of ANPR cameras to be used in free-flow, auto-triggered and high-speed traffic control applications.

OCR providers usually suggest a minimum of 3 pixels per 1 cm to achieve high license plate recognition rates (above 95%). Considering an European standard car or truck license plate with a character stroke of 1cm, to fully cover 4 meters of the road (3,5 meters lane width plus 0,5m overlapping) it would be required an image sensor with minimum resolution of 400cm x 3pixel/cm = 1.200 pixel.

On the opposite, if an ANPR camera with lower resolution is used (for instance 752x480), the following takes place:

Poor Recognition Rate: To cover 4 meters of the road, the resolution of the license plate image will be: 752pixel / 400cm = 1,88 pixel/cm. With this resolution it will not be possible to reach a decent license plate recognition rate (OCR software providers suggest to use a minimum resolution of 2,7 pix/cm to achieve 95% recognition rates).

ANPR on 2 meters: To achieve a 2,7 pix/cm resolution with an image sensor of 752x480 pixels, the lane width that is covered is: 752pix / 2,7pix/cm = 250 cm. Therefore, only the central part of the lane will be covered with the target resolution of 2,7 pix/cm. Vehicles travelling on the sides will be missed.

Sensitivity in Non-ideal operating conditions: License plate detection and recognition can underperform when a 752x480 image sensor is used. Furthermore, ANPR in non ideal operation conditions can suddenly decline. Non ideal operation conditions include dust in the lens, challenging meteorological conditions such as rain, snow, fog, etc. These circumstances can be overcome if a better resolution (megapixel) is used, because there is more information, more pixels per centimetre.

Imagsa’s Atalaya ANPR cameras use sensors of 2048x1024 pixels resolution. Hence, one Atalaya camera fully covers 7,5 meters of road width attaining the target resolution of 2,7 pix/cm. With one Atalaya ANPR camera, one installation and one investment is possible to simultaneously detect multiple vehicles on 2 Lanes of traffic.