Description
Automatic License Plate Recognition (ALPR) is the combination of a vision camera and an OCR (Optical Character Recognition) software. OCR software providers suggest that above 95% recognition rates can be achieved. This is only true if the OCR software process a very good quality (in terms of resolution, sensitivity, etc.) image of the license plate.
Here is where Mercury'Traffic provides a great value: Mercury'Traffic is an ultra-compact sensor that delivers the best quality image of the license plate in real-time. With the license plate images provided by Mercury' Traffic, the OCR software can really achieve top recognition rates.
There are many different challenging situations - such as sunlight reflections on the license plates, high-speed vehicles, shadows of vehicles, etc. - that difficult the correct detection of vehicles in road traffic applications. Mercury'Traffic overcomes these problems. It reliably detects the vehicles' license plates and delivers a very good quality image of the license plate. Mercury'Traffic combines a megapixel high-speed image sensor, able to capture up to 1,000 frames per second, with a high-performance hardware performing image processing algorithms at high speed. Following, the two basic steps to perform a license plate detection:
Capture images: The image sensor captures hundreds of images per second, varying different caption parameters such as the exposition time.
Process images: The next step is to process this large amount of images, in order to get the best quality image of the license plate (in terms of resolution and contrast). This is achieved by means of a high-performance image processing hardware, which uses parallel computing to process in real-time all the images. In order to perform the same amount of calculations of Mercury'Traffic with conventional CPUs, you would approximately need three INTEL III 2GHz CPUs .
The following video shows the results of an ALPR using Mercury'Traffic.