News

How To Improve The Recognition Rate Of Parking Lot License Plate Recognition System?

License plate recognition technology uses advanced image processing and artificial intelligence to allow access to a parking lot without a parking card, greatly accelerating vehicle traffic, and it is widely used in major parking lots. The development of license plate recognition systems has improved parking lot management efficiency, reduced personnel costs, and become an important part of intelligent transportation.

KBSP & KFC ANPR parking system

Even so, license plate recognition systems sometimes still make errors. How can we improve the recognition rate?

We can optimize the license plate recognition algorithm according to the following five characteristics to raise the recognition rate of the system.

  1. Accuracy: each calculation step of the algorithm has a precise definition.

  2. The input is a specific license plate image.

  3. The output is the result of license plate recognition.

  4. Feasibility is the effect of the license plate recognition.

Beyond internal factors, the factors affecting the recognition rate include external ones such as viewpoint changes, lighting, occlusion, scale, deformation, and complex backgrounds.

The license plate recognition system consists of crossings, vehicle detectors, high-definition license plate cameras, electronic display screens, and a control center. The high-definition plate camera captures plates at the entrance and exit, extracts the before-and-after images for comparison, and decides whether the vehicle may be released. Parking lot license plate recognition places high demands on the entire system, and the external environment greatly influences the recognition rate.

First, ensure that the light in the parking lot environment is bright enough; at night or when light is low, suitable lighting equipment should be added.

Second, ensure that the license plates of vehicles entering and leaving the lot are not damaged, so that the plate number can be identified and captured.