Blue Card Intelligent Parking Lot Management System

Blue Card Intelligent Parking Lot Management SystemAdvantages in Fixed Vehicle Management

1. Resolve the issue of multiple vehicles on one card during card management;
2. Resolve the issue of cards not being carried during card management;
3. Resolve the replacement and replacement of lost cards during card management;
4. Resolve the replacement and replacement of damaged cards during card management;
5. Communicate the installation and placement of cards during the process of "card management".

Blue Card Intelligent Parking Lot Management SystemAdvantages in temporary vehicle management

1. Opportunities for fee collectors to withdraw money when solving "card management" issues;
2. Liberate toll collectors from manual card issuance work and only need to be responsible for collecting money - even if an entrance card issuing machine is installed, it often requires someone to help transfer the "temporary card" between the "entrance card issuing machine" and the "temporary vehicle", which often deviates from the original goal of "saving human resources". Because temporary vehicles often park far away from the card dispenser, this is a normal subconscious reaction of people's fear of hitting the card dispenser and barrier gate;
3. When solving the problem of "card management", there may be temporary cars mixed out of the parking lot, resulting in the loss of temporary cards and parking fees;
4. Resolve the situation where a few fixed vehicles maliciously take away temporary cards when entering, resulting in the continuous loss of temporary cards;
5. Resolve the issue of limited card capacity in the entrance card dispenser;
6. Automatically distinguish between large and small vehicles and implement different charging standards;

Reason analysis of license plate recognition cases with poor practical application results on the market

There are some situations in the market where the actual performance of "license plate recognition" is very poor. The reason is that some parking lot system manufacturers do not have the ability to independently develop "license plate recognition" and can only purchase "license plate recognition" software externally. However, the reason for the poor recognition results of the purchased "license plate recognition" software is unknown, so they are powerless to improve the recognition rate.

These manufacturers have no choice but to use the "license plate recognition" function as an auxiliary function, essentially relying on card management work. They mostly use "license plate recognition" as a false intelligent gimmick. But the user ultimately paid a higher fee for this gimmick. Since 2009, Blue Card has been developing how to put "license plate recognition" technology into practical use. Based on feedback from its own and peer projects, Blue Card has found that the fundamental reason for the generally poor performance of "license plate recognition" in the current market is:

1. Cannot guarantee obtaining appropriate images, such as inappropriate angles, inappropriate lighting of license plates, and unclear license plate images;
2. A set of license plate recognition software cannot adapt to multiple intersection environments. For example, in a project, the image recognition rate of the east gate is high, but the image recognition rate of the west gate is low; The recognition rate is higher when the license plate is illuminated, but lower when it is backlit;
3. Image acquisition relies on cameras, license plate recognition relies on license plate recognition software on computers, and controlling the lifting of barriers relies on a "control board". If there are too many devices involved in the work, they will not work once the network is disconnected;
4. It is necessary to rely on license plate recognition software on the computer, which increases the difficulty of networking and wiring the capture camera, and cannot adapt to some environments and projects that cannot be configured on the computer.

The Blue Card 3D HD license plate recognition non-stop access parking lot management system integrates multiple technologies such as license plate focus exposure, 3D license plate recognition, and license plate distortion calibration, effectively solving the problem of environmental adaptability. The product integrates online video, vehicle capture, license plate recognition, voice playback, and intelligent control. Equipped with barrier gates and Chinese display screens, it can form a complete parking lot management system, which can manage fixed user vehicles and also realize temporary vehicle charging. Fixed vehicles can pass without stopping, and temporary vehicles can also enter without stopping. The system structure is simple, stable and reliable, and easy to install, maintain and use.

1. High definition snapshot

Mainly through high-resolution and high-definition images, the license plate recognition rate is improved. The resolution has been increased from 704 * 288 in standard definition to 1 million, 2 million, and 3 million pixels in high-definition; The field of view can not only see the license plate clearly, but also cover a larger area where the vehicle model can be seen clearly. The lane can be controlled wider, and the entry angle of the vehicle can be accepted larger; The clarity is controlled by the high resolution of Chinese characters, greatly improving the recognition rate of Chinese characters.

2. License Plate Focus Exposure

An automatic exposure algorithm that measures light in the license plate area and performs focus linkage on the license plate area to obtain exposure effects. Automatically adjust exposure parameters such as shutter speed, sensitivity, gain, etc. based on the brightness value analysis of the license plate area. These parameters can be adjusted frame by frame, with fast response and ensuring the exposure effect of the license plate area, thereby improving the license plate recognition rate. Using two sets of exposure parameters, stream and snapshot, automatic switching and adjustment, the stream effect matches the actual visual effect *, and it is also easy to debug. At present, the underexposure problem caused by daytime backlighting and overexposure problem caused by high sunlight have been successfully solved; Overexposure and underexposure issues at night.

3. Stereoscopic license plate recognition

Multiple cameras in master-slave mode are used for stereo recognition, with both cameras capturing and recognizing simultaneously. The main camera is equipped with an intelligent arbitration algorithm to select and control the recognition results. Thus solving the stringent requirements of license plate recognition technology for roads, angles, and other aspects while improving recognition rates. Increasing road width, reducing lane length, and eliminating the requirement for proper distance have improved the adaptability of the product. Clear and recognizable images can be captured when vehicles enter from various angles and directions.

4. License plate distortion calibration

Due to the constraints of the on-site environment, the surface of the license plate and the camera axis cannot be perpendicular, resulting in distortion of the license plate and smaller characters at positions far away from the camera, which poses certain difficulties for recognition. The distortion calibration algorithm uses a multi-color space for license plate background detection, accurately locates the license plate area, obtains the tilt angle and lateral deformation ratio of the license plate, and achieves an accuracy calibration rate of over 99% within a maximum tilt angle of 25 ° and a safety angle of 15 °.

5. Recognition Algorithm - Neural Network Feature Recognition

By using an energy filtering algorithm with feedback for texture detection, the success rate of license plate localization has increased from over 99%. The detection of multi-color space ensures the accurate identification of multiple types of license plates, including blue plates, yellow plates, military vehicles, police vehicles, WJ vehicles, personalized license plates, embassy and consulate license plates, provincial and Hong Kong license plates, etc. The character feature recognition algorithm for license plates ensures a character recognition rate of over 98% (measured on-site) and does not require on-site training; The auxiliary BP neural network recognition algorithm is a good supplement, while ensuring a good Chinese character recognition rate.

6. Fully networked

Unlike traditional camera audio cables, video cables, and other connection methods, the embedded stereo license plate recognition system adopts TCP/IP network mode. One power cable and several network cables can complete the overall system connection, making system wiring, construction, and maintenance easier and more convenient. The system can be connected using existing networks without the need for repeated wiring, greatly saving construction costs. Due to the widespread use of TCP/IP network users, the operating methods are also more easily accepted by management personnel.