KR20030051557A - Advanced License Plate Recognition System - Google Patents

Abstract

PURPOSE: A system for an intelligent vehicle number recognition is provided to reduce a mistake at recognizing the vehicle number by discriminating existence of the vehicle intelligently through a video photographed in real time, recognizing and extracting the vehicle number, and selecting the most accurate vehicle number out of photographed images. CONSTITUTION: A vehicle number recognition system(100) having an image photographed in a vehicle number decoding camera(200) judges whether vehicle information exists on a corresponding image. The vehicle number recognition system(100) decodes a vehicle number. An iris(210) of the vehicle number decoding camera(200) is controlled by a combined value of an illuminance of an image received in a subject total view camera(300) and an illuminance of an image received in the vehicle number decoding camera(200). An on-off control of an illumination device(400) is executed by the calculated illuminance.

Description

Intelligent License Plate Recognition System

The present invention relates to a vehicle number recognition system for extracting a vehicle number from a photographed vehicle image. Due to the recent increase in vehicles, many problems related to vehicles, such as traffic accidents, traffic jams, stolen vehicles, and speeding vehicles, are emerging as social issues. Accordingly, many studies have been conducted to solve problems through license plate recognition in application fields such as vehicle management, vehicle tracking, toll automatic charging, parking management, and arrangement vehicle management.

Application fields of vehicle number recognition system are used for speed control system (fixed, portable, mobile), signal violation control system, toll evasion prevention system, parking lot management system, and the system specifications are digital camera (more than 1 million pixels), It consists of a strobe (lighting device), a grab board, a vehicle number recognition main controller, an auto iris lens, and a vehicle detection sensor (Loop Coil, etc.).

An important core technology in such a vehicle identification system is a method of obtaining a good image that can be recognized by the main controller. That is, the key to the recognition rate depends on whether the obtained vehicle image is a good image to recognize. In addition, the accuracy and reliability of the license plate recognition algorithm and the status of the license plate affect the success rate. Companies developing vehicle identification systems are using various methods to obtain better images. Develop or modify a camera, adjust the intensity of the strobe, employ an automatic IRIS lens, control the aperture with the Ambient Light Sensor, use the pre-set Seasonal Illumination, or adjust the aperture with the brightness of the captured image. There have been many attempts to control feed-back.

The basic considerations of the present invention are firstly to improve the image acquisition method in order to obtain a recognizable image. Second, we propose and devise a detection method that can determine the illuminance of the subject with universality. Third, use nighttime lighting without strobe lighting for ease of installation and maintenance.

In consideration of the above, the following intelligent vehicle number recognition system has been developed to build a system that can increase the vehicle number recognition rate.

-. By continuously analyzing the image taken in real time to determine whether the vehicle has entered, the system determines whether the vehicle itself, and extracts the vehicle number from the captured vehicle image when judged as the vehicle. This intelligent vehicle judgment simplifies the configuration and installation of the system by eliminating the use of a separate sensor, and can dramatically improve the accuracy of vehicle number recognition by real-time information acquisition.

-. As a method of finding illuminance of a subject, a method of finding only a photographed image for recognizing a vehicle number is too simple image data, and thus there is a big difference in recognition success rate according to changes in surrounding conditions. For example, the image obtained due to the inability to control the appropriate aperture, for example, the deviation of illuminance due to a dark vehicle and a bright vehicle, tends to have a low recognition success rate. Therefore, by analyzing the image of the subject to obtain the illuminance, and also to obtain an image showing the entire subject and the surrounding environment as an auxiliary to determine the illuminance of the subject in conjunction with the analyzed illuminance can be obtained an accurate image.

-. Lastly, the system needs lighting at night or in bad weather such as fog by using general halogen lighting device with ON-OFF automatic control according to illuminance brightness. The use of strobe lighting, which is required in the existing system, cannot perform the vehicle number recognition task in real time, and it causes an accident and discomfort due to the driver's visual impairment by emitting intense light momentarily. Disadvantages include high cost and maintenance difficulties caused by complex and expensive equipment.

Means for achieving the present invention includes a vehicle image detection module for detecting the entry of the vehicle, a module for decoding the vehicle number from the captured image, a data management module for managing the recognized vehicle number, other systems (parent system, DVR, etc.) It consists of a communication module, a camera lens and a camera shutter speed according to the illuminance in the captured image, a module for on / off control of the nighttime lighting device, and a HMI (Human Machine Interface) function for the administrator to operate.

1 is a block diagram of an intelligent vehicle number recognition system

(Figure 2) is the intelligent vehicle number recognition system software configuration

3 is a flowchart of vehicle image identification

(Figure 4) is a vehicle number analysis flow chart

5 is a table for information management in the intelligent vehicle number recognition system

The present invention for achieving the above object will be described with reference to the accompanying drawings for the detailed configuration and operation.

First (Fig. 1) is a diagram of a vehicle number recognition system, and the vehicle number recognition system 100 receiving the image captured by the vehicle number decoding camera 200 determines whether there is vehicle information in the corresponding image, If determined, the car number is decoded. In this case, the control of the aperture 210 of the camera for decoding the vehicle number is based on a combination of the illuminance of the image received from the camera 300 for viewing the entire subject and the illuminance of the image obtained from the camera number decoding camera 200. The aperture 210 of the lens is controlled. The on-off control of the lighting device 400 is performed by the illuminance calculated here.

The second (Fig. 2) is the software module configuration of the vehicle number vehicle identification system, which adjusts pan-tilt, aperture, zoom, shutter speed, etc. to obtain the desired image in real time. A vehicle image identification module 200 for acquiring a vehicle image with a camera control communication module 500 to determine whether the image is a vehicle image, a vehicle number analysis module 300 for recognizing a vehicle number from the obtained vehicle image, and a recognized vehicle. Data management module 400 for managing the number and the image of the decoding vehicle image and the image for detecting the illumination in the database, the illumination analysis module 600 for detecting the illumination around the subject, the collected vehicle number is interlocked with the external system It is a software 100 consisting of an external system supporting the communication module 700 and the HMI (Human Machine Interface) module for managing the intelligent vehicle number recognition system.

The third (Fig. 3) shows the processing flow diagram of the vehicle image identification module 200 to obtain a vehicle image in Figure 2 and determine whether the image is a vehicle image.

The fourth (Fig. 4) shows a process flow diagram of the vehicle number analysis module 300 for recognizing the vehicle number in the vehicle image obtained in (Fig. 2).

Fifth (Fig. 5) is a data structure handled by the data management module 400 for managing the vehicle number recognized in (Fig. 2), the image of the decoded vehicle image, the image of illuminance detection, the image capturing time, etc. in the database.

Instead of acquiring the vehicle image by the external trigger signal, the system intelligently identifies whether there is a vehicle in the continuously captured image in real time, and recognizes / extracts the vehicle number when it is determined by the system as the image of the vehicle. By selecting the vehicle number with the highest accuracy in the image, the probability of the recognized vehicle number becoming unrecognized and misrecognized is greatly reduced. In addition, it does not use external trigger signal and strobe lighting device, which makes the driver's safety, installation and maintenance easy.

In addition, by controlling the aperture control and shutter speed by linking the brightness of the subject with the brightness around the subject, it is possible to obtain a good quality image. That is, it is an important management factor to improve the vehicle number recognition rate.

In this way, by recognizing the wrong number when decoding the license plate image and reducing the probability of unrecognition, it is efficient and intelligent in applications such as vehicle management, vehicle tracking, automatic toll collection, parking management, and arrangement of vehicles. It is possible to build the system, reduce expenses, prevent traffic accidents, and enable efficient manpower operation.

Claims (4)

Intelligent vehicle number recognition system, characterized in that for determining the vehicle number from the image taken in real time, and the process for recognizing the vehicle number without a signal using a separate vehicle detection sensor in the vehicle number recognition. In order to obtain a vehicle image that is easy to recognize in vehicle number recognition, in order to accurately grasp the illuminance of the subject, the image of the vehicle number interpretation camera is analyzed by linking and analyzing the image for recognizing the vehicle number and the image that can be viewed as a whole. Intelligent vehicle number recognition system, characterized in that the automatic control of the speed and aperture. In the vehicle number recognition, the vehicle number recognition system, characterized in that the on-off control of the constant lighting device according to the illuminance value obtained in claim 2 without using the strobe illumination of the ambient brightness at night. An intelligent vehicle number recognition system according to claim 1, wherein the vehicle number recognition system selects and recognizes the most accurate vehicle number by comparing and analyzing a plurality of vehicle numbers obtained from continuous vehicle images in real time.