Imagsa - News

Tip #8 – Multi Purpose ANPR: speed, video, ...

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.

ANPR cameras are now Traffic Data Collectors

Traditionally, ANPR cameras were conceived as single purposed devices: getting license plates images and performing number plate recognition.

However, new ANPR cameras have the capabilities to deliver much more information. Additionally to number plate recognition, some ANPR cameras can also obtain overview image of the vehicles (Day and Night), deliver a video stream or even the instant speed of vehicles.

Therefore, the single-purpose ANPR camera is transformed to a traffic data collection station.

Based on the capability to capture and analyse more than 250 images per second, Atalaya ANPR camera can track the license plate at high-speed in order to determine the vehicle’s spot speed.

Atalaya ANPR camera can also count vehicles, provide traffic statistics, deliver a video stream and other useful traffic data.

With one single camera investment and installation, is now possible to deliver more traffic data on top of license plate recognition. And it is now possible to deploy additional applications and services with the same infrastructure.

Tip #7 – ANPR Programmability Helps Integration

Integration Check-ins

When designing any type of system, also an ANPR-based one, it should be analysed the integration and implementation times and costs.

ANPR cameras are usually an element of an overall traffic control system. As such, it should be integrated with other equipments and high-level processes. Thus, in addition to assess the performance of ANPR vendors, it is important to analyze the camera connectivity and interaction with other equipment.

Things like the communication protocol (TCP/IP or proprietary), the existence of an interface (API, SDK) to interact with the camera, or even the possibility to run specific and custom software inside the camera, are also relevant.

In order to minimize the installation and integration costs and speed up the time-to-market, Atalaya has been conceived as a stand-alone and ready to install ANPR camera. This is an all-in-one and autonomous ANPR camera.

No additional roadside equipment (PC station, sensors, etc. ) is required. Atalaya ANPR camera integrates a standard ATOM CPU running a Windows XPe O.S. Hence, system integrators can run their software on it, with no additional equipment costs.

Additionally, Imagsa provides the tools (SDK) to interact with the camera. There is no vendor lock-in, and the system integrator can freely develop its own traffic control applications with Atalaya ANPR camera.

Tip #5 – Top ANPR performance II: Recognition

Detection and Recognition Rates

Overall performance of an ANPR camera is computed by multiplying the Detection Rate by the Recognition Rate. Detection rate has been analyzed in Tip #4, and it is measured by the number of vehicles (plates) detected by the ANPR camera over the Total number of vehicles going through the camera’s field of view. Detection Rates range is 50% to 99%, depending on the selected ANPR solution.

Better Recognition with Good Quality images

Recognition rate mainly depends on two factors:

The quality of the license plate image
OCR (Optical Character Recognition) engine.

OCR is a “mature” technology, and top recognition accuracy can be reached if the input image has good quality. Thus, quality of the license plate - in terms of Resolution (Tip #1, Tip #2) and Illumination (Tip #3)- is a prerequisite for best performance of the OCR engine.

So, real improvement in ANPR comes from a good quality image of the license plate.

Imagsa's Atalaya ANPR camera ensures a good quality of the license plate image by using a 2 Megapixel image sensor, and an auto-adaptive Multi Exposure technique at high speed (250 FPS). Therefore, top License Plate Recognition rates can be achieved.

Tip #6 – Flexible ANPR installation

Where to fix the ANPR camera?

Ideal installation conditions for best ANPR performance are: overhead position, 6 or 7 meters height and around 20 meters focal distance to avoid occlusions.

However, sometimes this ideal installation conditions are not possible. Or an alternative installation position is preferable to reduce installation costs. For instance, an installation in a roadside pole (instead of a gantry) can heavily reduce costs of a project.

Thus, an ANPR camera should allow a flexible installation on an overhead position (gantry), at the roadside, in a range of heights and distances.

Setting Up the ANPR camera

Once the installation point has been decided, it is time for setting up the camera. In order to make this process as less painful as possible, ANPR camera manufacturers should provide suitable tools and interfaces for an easy set-up and camera calibration by the field technician.

Atalaya ANPR camera can be installed both overlane and in a roadside position. In order to achieve top ANPR accuracy from a lateral installation, Atalaya ANPR integrates a real-time perspective correction of license plate images.

Therefore, top detection (Tip #4) and recognition (Tip #5) rates are achieved in both types of installations. On the installation side, Imagsa provides an useful and intuitive GUI (Graphical User Interface) to make a trouble-free installation and achieve top ANPR performance from day one.

Tip #4 – Top ANPR performance I: Detection

Definition: Vehicle Detection Rate

Overall performance of an ANPR camera is computed by multiplying the Detection Rate by the Recognition Rate.

Detection rate is measured by the number of vehicles (plates) detected by the ANPR camera over the Total number of vehicles going through the camera’s field of view.

Detection Rates range is 50% to 99%, depending on the selected ANPR solution.

Motion vs Plate Detection

Standard ANPR cameras with video triggering detect a vehicle by means of motion detection techniques. Motion detection consists on analysing the differences between video frames. When a difference occurs, this technique considers that there is a vehicle and it captures an image to read the license plate.

This technique has so main drawbacks:

Illumination: Performance is sensible to illumination variations, and it has false detections caused by shadows of a vehicle travelling in the adjacent lane, or even by shadows of trees next to a road. Shadows create false detections and undermine the performance of the ANPR solution.

Miss Detection: In addition to false detections, some ANPR cameras are affected by miss detections. This is a more critical issue, as it means that not all vehicles will be detected. Miss detections can occur by a variety of factors. But mainly because the motion detection algorithm, running in the ANPR camera, does not perceive a sufficient change on opacity in a sequence of images. Sometimes this affects the overall performance of the ANPR solution, to an extent of just detecting 50% to 80% of vehicles.

Atalaya ANPR cameras implement an innovative license plate detection technique (instead of motion detection).

Basically, it consists on continuously capturing and analysing images to detect a license plate pattern. When there is a license plate pattern it means that a vehicle is in the field.

This approach avoids the false detections that are common in an ANPR camera based on motion detection. And it achieves vehicle detection rates over 99%. Meaning that nearly every vehicle (full traffic) is detected by the Atalaya ANPR camera.