1、 User demand analysis

In modern parking lot management, it involves various aspects of management, among which vehicle management is an important aspect. Especially for special parking lots, courtyards, and residential areas, it is required to strictly manage various vehicles in real time, monitor their entry and exit time, and register and identify all types of vehicles (including internal and external vehicles). In large-scale areas, there are many vehicles entering and exiting, and manual judgment is required for each vehicle, which is time-consuming and not conducive to management and inquiry. Security work is difficult and inefficient. In order to improve this management mode that is not commensurate with modern parking lots, courtyards, and residential areas, it is necessary to achieve automation and intelligence of vehicle management work as soon as possible, and manage it in the form of computer networks to effectively and accurately monitor and manage all vehicles at entrances and exits. Require the system to provide corresponding application software to achieve efficient and intelligent management of the camp.

This system utilizes video streamingLicense Plate Automatic Recognition AlgorithmOr triggered by ground sensing, capturing vehiclesLicense plate recognitionWhen a vehicle enters the entrance of the community, the license plate automatic recognition algorithm automatically captures a photo of the vehicle and recognizes the license plate number. It records the license plate number, color, license plate feature data, entry time information, etc. The vehicle can enter and exit the parking lot without barriers, providing users with a new service mode.

The system automatically recognizes the number and license plate characteristics of vehicles entering the community, verifies the user's legal identity, automatically compares the blacklist database, automatically alarms, and can monitor and manage the entire parking lot, including entrance and exit management, internal management, data collection, storage, and system working status, for administrators to monitor, maintain, count, query, and print reports. Vehicles entering and leaving the community are monitored by the system, which makes the entry and exit, toll collection, anti-theft, and parking management of the community intelligent, automated, and convenient, fast, safe, and reliable

Its main characteristics are as follows:

The dependence of the recognition system on the environment has been reduced to the lowest level, enabling it to work normally 24/7 and maintain a high level of recognition rate.

The LPR recognition system has improved the speed and accuracy of recognition.

The recognizable width of the smallest card is 80 pixels

Adapt to complex climate and lighting conditions, such as cloudy, rainy, and nighttime, to ensure high recognition rates.

Adapting to high-speed and high traffic flow, high recognition rate (>98%) can still be guaranteed at a speed of 20 km/h and a single lane flow of 30 vehicles/minute.

Realize frame by frame processing of video images, trigger video streams, without burying ground sensing coils (dual channels), to avoid damaging the road surface.

The engineering installation is simple, the operation is stable, and it does not interfere with the user's existing system.

Having the processing capability of *, it can recognize and process all images during the vehicle's movement, without relying on a single image, effectively improving the device's adaptability to complex environments

1.1 Adaptability to different lighting conditions

The environment at the construction site is quite complex, such as smoke, rain, snow, sunlight from different angles, car lights, and large billboards, which may interfere with the recognition system. Especially for recognition devices that use external triggering methods, their recognition rate depends heavily on the captured images. At the moment of capture, the vehicle license plate is in the affected position, which can cause misidentification.

my company'sLicense Plate Recognition AlgorithmReal time frame by frame processing of video images. During the movement of vehicles, the angle and lighting are constantly changing, and the license plate will always be clear at certain moments. We will definitely collect some video frames with clear license plates for analysis and recognition. Therefore, our company'sLicense plate recognition deviceResistance to light and climate.

1.2 Adaptability to vehicles passing levels and driving at ultra-low speeds

Due to the use of a high-speed algorithm platform that adapts to a speed of 20 kilometers per hour, vehicles can accurately recognize license plate numbers and capture images when driving at super high speeds (obstacle crossing vehicles) or super low speeds, avoiding the phenomenon of being unable to capture images at intersections where high-speed vehicles pass through.

2、 System structure and technical indicators

2.1 System workflow

Entrance section

The explanation is as follows:

The vehicle reaches the entrance camera recognition area of the community, automatically recognizes the vehicle license plate number, and makes a judgment on the vehicle type.
Internal vehicle: Automatic release/manual release is optional, and the vehicle enters the information and image storage database.

Temporary vehicle: Automatic gate release/manual gate release optional, timed and saved entrance capture pictures to the database.

Unable to confirm vehicle: can be manually released, license plate number can be manually entered, license plate number can be manually modified, and recorded in the database.

Export section

The explanation is as follows:

The vehicle reaches the entrance camera recognition area of the community, automatically recognizes the vehicle license plate number, and makes a judgment on the vehicle type. Internal vehicle: Automatic release/manual release is optional, and the vehicle enters the information and image storage database.

Temporary vehicle: Automatic gate release/manual gate release optional, timed and saved entrance capture pictures to the database. If charged, it will be charged according to the temporary vehicle fee standard, and manual release is generally chosen.

Unable to confirm vehicle: manual release is possible, manually enter license plate number, record database, and generate correct fees.

2.2. Installation diagram of license plate recognition system

Install a camera pole with a height of 2.0-2.8 meters at each entrance and exit. The pole is equipped with a license plate recognition camera, and the camera lens points to the ground of the lane at a height of about 4.0-4.5 meters, aiming at the license plate.

The specific installation location is shown in the following figure.

Equipment Installation Diagram 1

Specific construction requirements will be detailed according to the site conditions. Choose different focal length lenses for different construction sizes.

2.3 Identification system technical indicators

Under normal urban license plate cleanliness conditions, the driving license plate is unobstructed, and the average letter and number recognition rate can reach over 99%. By strengthening the template for the system application area, the entire license plate recognition rate (including license plate Chinese characters) can reach over 99%.

Number plate recognition time:<0.2 seconds

Full license plate recognition rate:>98% (Full license plate recognition rate=* number of correct license plates/natural traffic flow)

License plate detection rate:>99.9%

Permitted vehicle speed: 0-20 kilometers per hour

Output image resolution: 720 * 288

Output information: vehicle image, license plate identification number, license plate color, vehicle type, entry and exit time

2.4 Introduction to License Plate Automatic Recognition Device

Product Introduction:

A high-speed DSP embedded device with embedded recognition software that includes modules such as video capture, image preprocessing, license plate detection, license plate segmentation, character recognition, tracking and comparison, image compression, and data transmission. The system has fast recognition speed and high reliability. The device can perform license plate recognition on the two input videos separately. According to different engineering application scenarios, the two videos can be selected as front or rear mode recognition without affecting each other.

Features:

Equipped with independent two video interfaces, it can simultaneously input two different videos for license plate recognition;

Frame by frame recognition of video streams, extracting license plate number information and color information present in the video, and providing corresponding single or multiple frame images, license plate small images, license plate binary images, and other information containing the vehicle;

Provide additional vehicle information output, such as real-time license plate size and location information;

Provide image pre correction and fine correction functions;

Using dynamic feature detection technology to detect and capture unlicensed vehicles;

Provide video surveillance interface to complete real-time video transmission function;

Provide customized output time function for recognition results.

2.5 Characteristics and Parameters

Features:

1.4 million pixel 1/3 'progressive scanning CCD, with the highest resolution of 1360 × 1024 and a frame rate of up to 30 frames per second

Support dual shutter, support simultaneous output of video stream and captured frames, and perform road monitoring while capturing

Supports dual stream and adopts H.264 video compression technology with high compression ratio and flexible processing. It also supports MJPEG encoding and JPEG encoding for captured frames. The image quality can be adjusted

Support for capturing synchronized flash output, with up to 3 channels of flash synchronized output, and support for multiple fill light methods: independent flash, no flash, associated flash, wheel flash, and strobe, etc
Support flash automatic light control function

Support input of traffic and green light signal status, different triggering methods can be set for red and green lights

Support multiple capture methods

Optocoupler isolation input ordinary coil capture

Optocoupler isolation input coil speed measurement capture

RS-485 triggers capture

Video checkpoint virtual coil capture

Support intelligent recognition function

Built in license plate recognition function
Built in vehicle recognition function

Support network triggered continuous capture function

Support image synthesis function, which can combine multiple illegal images into one image, and the synthesis method can be flexibly configured

Supports RS-485 function and can be connected to serial peripheral devices such as vehicle detectors, traffic light signal detectors, and radars

Camera parameter configuration function

Video and snapshot independent control automatic exposure AGC、 white balance

Independently configure character overlay function for videos and captured images

Signal light synchronization control function

Support local video preview, CVBS signal or HD-SDI signal output

Support SD/SDHC local image storage and retrieval, automatic overlay, and automatic upload

Support external USB storage devices for local image storage, automatic overwrite, and automatic upload

*The network function supports multiple image upload methods (FTP, Hikvision SDK deployment, etc.)

Support hardware watchdog

Support importing and exporting configuration files

Technical Specifications

3、 System functions

3.1 Application System Functions

The automatic license plate recognition algorithm can be flexibly integrated into various management systems that require license plate recognition functions. In general, management software has the following functions:

Management system interface

Entry and exit record query report

Fee Record Report and Query

Temporary vehicle fee standard

License plate recognition: The recognition of license plates adopts automatic license plate recognition algorithm technology, which can work 24/7. Using pure video algorithm recognition, it can work without the need for ground sensing

Storage and archiving of vehicle entry and exit data

Automatic comparison function for license plate numbers, which compares and classifies license plate numbers through database queries

Support four types of vehicles: temporary billing vehicles, monthly rental vehicles, stored value vehicles, and free vehicles.

License plate query: including A: query by entry and exit time; B: Search by entry and exit location; C: Query based on various query criteria such as vehicle license plate number

fuzzy query

Traffic flow statistics function: Collect statistics on vehicles entering and exiting in any direction at any time and generate statistical reports

Print data list or query result list

External output voice and LED information display screen (optional)

As the front-end of the application system, the automatic license plate recognition algorithm requires a vehicle management system and database support in the back-end to effectively utilize and reflect the functions of license plate recognition, and to play its huge role in the vehicle management system.

3.2 Central Management System

Parking Lot Management - includes defining the name of the parking lot, managing the number of parking spaces, and assigning parking lot numbers. Can be added or modified.

User Management - includes operator permission management and user ID. It can be added or modified

System password modification, database connection settings, etc

Vehicle management consists of three parts: 1) internal vehicle authorization, 2) internal vehicle recharge extension, and 3) internal vehicle cancellation.

Report query includes

1) Authorization report

2) Car Owner Report

3) Recharge Report

4) Amount report

5) Fee report

6) Entry and exit report

7) Entry and Exit Details Report

8) Parking lot status