Automatic Number Plate Recognition, or ANPR, is a great asset to law enforcement, parking management, highway monitoring, gated community surveillance, toll management and other uses. However, for professionals and hobbyists alike, there are several issues with ANPR camera setup that impedes its accuracy. If your license plate recognition accuracy rate is at 68 percent rather than 98 percent, then there is hope for you to increase the effectiveness of your system.
In this article, we will share proven insights from our customers on how to improve your ANPR accuracy rate by making a few modifications to your ANPR camera setup.
What is ANPR?
ANPR is automatic number plate recognition, also known as Automatic License Plate Recognition (ALPR) in some countries. It is a system that is capable of reading vehicle number plates from images captured.
As this industry expands, it has become important to note that there are factors that impact its success. When ANPR accuracy rates are suboptimal, it is common to blame the underlying ANPR engine.
3 Important Factors Impacting ANPR Success
There are three fundamental components that make ANPR successful:
- Camera setup
- ANPR engine
Of these three, we see that camera setup holds the most potential for improvement for businesses and consumers to significantly bolster their ANPR results.
We draw this conclusion because we know that our Plate Recognizer ANPR engine has been thoroughly tested and optimized to handle blurry, low-res, and dark images as well as other aspects of the real-world environment. Also, counter to claims from camera manufacturers, we see that a modest, mid-range IP camera performs just as well as a high-end, expensive camera. We have observed how proper camera setup along with a strong ANPR engine can be an effective solution to detect and decode vehicle license plates.
This article offers specific tips and suggestions on how to best set up your camera to maximize your ANPR results. We tried our best to prioritize the list, with the highest potential contributors on top.
Camera Zoom for Best ANPR
For beginners handling license plate recognition, there is a common misconception that a wide-angle shot is the best. Cameras will seek to aim wide across parking lots rather than towards entrance or exits, which allow them to truly pick up plates and get a good read.
This is a simple fix if you know how to fix it. Adjust the camera positioning and width of the focus so that the license plate is picked up in the shot, and you’ll instantly have greater accuracy. Rather than have a wide-angle view of the street with all the homes around it, it is better to focus on just the street itself.
In the photo below, you can find the image with the red box showing the plate. If the camera was zoomed in closer to the gate, the focus would improve. Make sure you are not over-zooming because you’ll need at least 40 pixels of width in order to get a good read on the plate.
Zooming into the vehicle yields stronger ANPR accuracy as compared to taking a wide-angle view of the area.. Source: Plate Recognizer Clients.
Camera Distance for Better ANPR
The maximum advisable distance between the camera and the vehicle is 35 meters. Actually, whenever possible, it is preferred to minimize that distance. Why? Because minimizing the distance between the camera and vehicle helps ensure that the camera can easily focus without the need to zoom in to the target vehicle. This helps reduce image blurriness.
Distance between camera and vehicle should be minimized as much as possible and definitely under 35 meters. Source: Plate Recognizer.
Camera Angle to Improve ANPR
While Plate Recognizer ANPR has been tuned to support a wide variety of license plate angles, it’s always ideal to have the camera set up appropriately. In terms of angle, the setup of the ANPR camera can be positioned in two ways, slope as well as vertically and horizontally. For both cases, it is advisable to have a maximum of 45 degrees for a proper read of the license plate.
Vertical angle of camera on vehicle license plate should be under 45 degrees. Source: Plate Recognizer .
Horizontal angle of camera on vehicle license plate should be under 45 degrees. Source: Plate Recognizer .
Minimum Resolution for Best ANPR
While Plate Recognizer works with any IP camera, we recommend that the license plate itself must have at least 50-100 pixels in width. Otherwise, the ANPR engine may not be able to effectively read the license plate.
This means that depending on your use-case, you may need a camera with a certain number of megapixels (MPs). Here are some of our guidelines or suggestions based on the ANPR use case:
- 2 MP camera is suitable for parking management and toll ANPR projects.
- 4 MP camera is good for logistics or vehicle repair-type projects.
- 8 MP camera is needed for highway or street monitoring.
The above are just rough guidelines. If you’re unsure, then just grab a few frames from your camera or take a few photos with your mobile phone with settings based on camera megapixels. Thereafter, run our ANPR engine through some of those images to see if the plates are correctly detected.
Camera Frames May Impact ANPR
The camera frame (and thus your ANPR camera setup) is largely based on the vehicle’s speed. If you’re looking to capture license plates that are still versus license plates that are moving (also called “free flow”) then the specifications for altering your camera will be different.
You will need to calculate the net difference in speed. If your camera is in a fixed position, then the net speed is the speed of the vehicle. If the camera is maintained in your vehicle (e.g. police car), and then you’re driving the same direction of the target vehicle, then the net speed is the difference in your speed and the target vehicle’s speed. Keeping this in mind prevents you from having any unusable images because you didn’t take the time to equip your camera for the proper camera speed.
For example, if the vehicle is traveling at 10 miles per hour (mph), you can get good ANPR results with just 10 to 15 frames per second. With the vehicle at 30 mph, grabbing images at 15 to 25 frames per second should suffice. At 60 mph, you’ll need 30 to 40 frames per second. The suggestions here may differ greatly depending on the camera quality and zoom level, so it is always best to test and refine.
Moreover, the faster the vehicle is traveling, the more images you’d want to send over to the Plate Recognizer engine. This way, our ANPR engine has more opportunities to evaluate and decode the license plate as the vehicle passes a certain point. At 10 mph, 1 or 2 images would be sufficient, starting at 0 and then 0.5 second of when motion was first detected. At 30 mph, you may want to send 3-5 images to the ANPR engine, at intervals of 0.2 to 0.4 seconds. At 60 mph, it’s best to send 5-10 images, at intervals of 0.1 to 0.2 seconds.
Good summary of how vehicle speed impacts FPS, images sent and also time between images. Source: Plate Recognizer.
Camera Lighting for Better ANPR
While Plate Recognizer has greatly improved its algorithms to support dark images, it is nevertheless a good idea to ensure that the camera is set for the right lighting conditions. You can adjust the camera shutter speed accordingly so as to not to over-expose or under-expose the image. For example, if it is bright outside, adjust the shutter speed to capture at 1/5000 of a second. At night time, set the shutter speed to 0.75 to 1 second.
If the camera does not auto-adjust the shutter speed based on the outside lighting conditions, then you can try to have a light fixture directed towards the entering vehicle. This way, there will always be adequate lighting on the vehicle.
Ensure adequate lighting to avoid over-exposed or under-exposed images. Source: Plate Recognizer.
Additional Tips to Improve ANPR
While there’s plenty that you should do, there’s also plenty that you should not do. Let’s face it. Cameras come with a lot of settings, and it’s easy to switch on a bunch of settings that you either don’t know what they are or don’t need and forget that you’ve activated. Here are some to keep in mind:
Automatic gain control (AGC), digital noise reduction (DNR), autofocus, and back light compensation (BLC) are all features you want to keep disabled while enabling ANPR camera setup. Once again, this is because it will give you the best chance of grabbing that license plate number from a moving vehicle using ANPR.
AGC creates issues because the gain itself prompts digital noise and lower recognition in the image. It’s often much simpler just to leave the feature off. DNR is best left alone because it is performed by removing pixels based on comparing two frames. Although this might seem harmless, it’s often not because it can easy remove pixels that could be helpful to you in the future. Next, you can pass on autofocus because adjusting the sharpness often reduces the recognition quality in the image itself. Finally, the BLC can cause issues with the image because it often occurs when a light source enters a frame. When the pixels do not have enough time to properly adjust, the camera will not be able to capture a good image.
And, while this may seem obvious, we see that the best ANPR contains images that are in landscape rather than in portrait view. This makes intuitive sense, since the license plate itself is more landscape than portrait. And, just like watching TV, we (and thus our ANPR engine) is used to seeing the world in a portrait format.
Following the best practices discussed in this article can be the difference between an ANPR accuracy of 55 percent and 99.5 percent. In many cases, it’s just as much about knowing what to do as it is knowing what not to do. Start with camera height, width, and distance and you’ll be able to start updating your ANPR camera setup, so your images are usable and accurate for safety and surveillance purposes.
With Plate Recognizer, we encourage you to use our API Cloud for ANPR so that you can see the images captured right on the Plate Recognizer dashboard. You can use this as a way to fine tune and diagnose your camera settings. And, if needed, feel free to consult with the Plate Recognizer team experts to further optimize.