KR101298684B1 - Non sensor based vehicle number recognition system and operating method thereof - Google Patents

Abstract

PURPOSE: A non sensor-based vehicle number recognizing system and an operating method thereof are provided to recognize a number and a number plate while performing city management by using an image photographed by a general CCTV. CONSTITUTION: A vehicle number recognizing system receives a video from a CCTV and a still image by frame (S11,S12). The system performs gray scale transformation for images for the recognition of a number plate and performs image correction operation through contrast (S13,S14). The system extracts a vertical/horizontal area from the image (S15). The system extracts a number plate image from the image and recognizes a vehicle number in the number plate image (S16). [Reference numerals] (AA) Extraction failure; (BB) Extraction success; (CC) Repeat; (S11) Input a video; (S12) Input still images by frame; (S13) Perform transformation into gray scale images; (S14) Perform image correction regarding brightness/contrast; (S15) Extract a horizontal/vertical area from the images; (S16) Extract a number plate image

Description

Non-sensor based vehicle number recognition system and its operation method {Non Sensor Based Vehicle Number Recognition System and Operating Method}

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a vehicle number recognition system and an operation method thereof, and more particularly, to a vehicle number recognition system and an operation method capable of recognizing a license plate and a number of a vehicle without using a sensor.

In general, closed-circuit television (CCTV) is installed for the purpose of managing the city, and for securing and responding to evidence against various crimes. The control center can perform surveillance by monitoring by receiving images taken from the CCTV installed in various places in the city and reproducing them.

However, in the conventional case, only the simple monitoring of the images photographed from the CCTV is performed in the control center, and therefore, the obscured area is generated, so that there is a problem that sufficient data can not be provided in the city management, Prevention becomes difficult. In addition, despite the fact that the number of crimes using a vehicle has recently increased, accurate photographing of the vehicle, and the need to accurately photograph the license plate of the vehicle so as to distinguish the vehicle, it is difficult to obtain such data substantially . In order to crack down on speeding violations or shoulder driving violations, there are cases where a crackdown camera is used to photograph the license plate of a vehicle. However, this is expensive equipment and is used only at a certain point of the road. It is difficult to use as a camera for the purpose.

In addition, even if a vehicle used for a crime is accurately photographed through CCTV, this is mainly used for evidence data, etc., and thus it is difficult to use the photographing data in real time in a special situation such as the crime. .

The present invention is to solve the problems as described above, based on the image taken by the general CCTV, city management and vehicle license plate and number recognition, using the recognized vehicle license plate and number using a vehicle, etc. An object of the present invention is to provide a non-sensing sensor-based vehicle number recognition system and a method of operation thereof that can cope with crime.

In order to achieve the above object, a method of operating a vehicle number recognition system according to an embodiment of the present invention, receiving a video from one or more cameras in a first frame per second, and from the received video Extracting a still image corresponding to two frames per second; analyzing the still image to output a vehicle recognition result based on at least one of a motion detection and a driving direction detection result; and a still recognized that the vehicle exists Extracting the license plate through the horizontal and vertical extraction operation from the image, performing the license plate shaping operation according to the geometric shaping technique for the extracted license plate, and detecting and extracting the license plate by extracting individual characters existing in the standardized license plate Vehicle number by performing a number recognition operation on the individual characters And outputting a result of recognizing the call.

Meanwhile, a vehicle number recognition system according to an embodiment of the present invention includes a number recognition server, wherein the number recognition server receives a video from one or more cameras at a first frame per second, and receives a video from the received video. An image processor configured to extract a still image corresponding to two frames per second, and perform image processing on the extracted still image, including image correction operation through gray scale conversion and contrast, and detecting the motion by analyzing the still image And a license plate for outputting the vehicle recognition result based on at least one of the driving direction detection results, and a license plate through horizontal and vertical extraction operations from the still image recognized as the vehicle, and geometrically shaping the extracted license plate. Performs license plate shaping operation according to the technique As by the presence extract individual characters to the detect plate, and is characterized in that comprising a recognition number and outputting a recognition result of the train number, by performing the code recognition process operation for the extracted individual characters.

According to the non-sensing sensor-based vehicle number recognition system of the present invention and its operation method as described above, it is possible to recognize the license plate and the number, and also to recognize the city while performing city management using the image taken by the general CCTV By using the license plate number and the number, the crime can be coped with the vehicle.

In addition, according to the non-sensing sensor-based vehicle number recognition system of the present invention and its operation method, since the vehicle number is recognized through a relatively inexpensive CCTV, it is possible to minimize the blind spot by expanding the CCTV, according to the recognition result Through the occurrence and propagation of events, it is possible to respond quickly to crimes.

1 is a block diagram illustrating an embodiment of a vehicle number recognition system including a number recognition server according to an embodiment of the present invention.
FIG. 2 is a block diagram illustrating a specific implementation of the image processor of FIG. 1.
3 is a block diagram illustrating a specific implementation of the vehicle recognition unit of FIG. 1.
4A and 4B are block diagrams illustrating exemplary implementations of the number recognition unit of FIG. 1.
5 is a diagram illustrating an example of extracting a license plate according to an embodiment of the present invention.
6 is a diagram illustrating an example of recognizing a vehicle number according to an embodiment of the present invention.
7 is a block diagram illustrating an embodiment of a vehicle number recognition system according to another embodiment of the present invention.
8 is a flowchart illustrating a method of operating a vehicle number recognition system according to an embodiment of the present invention.
9 is a flowchart illustrating a method of operating a spatial information system according to another embodiment of the present invention.

In order to fully understand the present invention, operational advantages of the present invention, and objects achieved by the practice of the present invention, reference should be made to the accompanying drawings and the accompanying drawings which illustrate preferred embodiments of the present invention.

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, the present invention will be described in detail with reference to the preferred embodiments of the present invention with reference to the accompanying drawings. Like reference symbols in the drawings denote like elements.

1 is a block diagram illustrating an embodiment of a vehicle number recognition system including a number recognition server according to an embodiment of the present invention. As shown in FIG. 1, the vehicle number recognition system 100 according to an embodiment of the present invention may include a number recognition server 1000. In FIG. 1, the number recognition server 1000 may further include a plurality of external cameras (eg, CCTVs 110-110 -N) and an image control server 120. The video monitoring server 120 is connected to a plurality of CCTVs 110_1 to 110_N via a wired / wireless communication network, receives video images captured by the plurality of CCTVs 110_1 to 110_N, And can also control a photographing operation of a plurality of CCTVs 110_1 to 110_N by outputting a control signal through an internal control means (not shown).

The plurality of CCTVs 110_1-110_N may be installed in an area where a management service provided by the vehicle number recognition system 100 is provided. (For example, a traveling image of a vehicle) from each CCTV is provided to the video control server 120 and is also provided to a plurality of CCTVs 110_1 to 110_N through the operation of a user who operates the video control server 120. [ Such as panning, tilting, and zooming (PTZ). Alternatively, when occurrence of a specific situation (event) is detected, information for responding thereto may be provided to the video control server 120, and a plurality of CCTVs 110_1 to 110_N ) Can be automatically controlled.

According to the exemplary embodiment of the present invention, the number recognition server 1000 may receive a driving image of the vehicle photographed from the plurality of CCTVs 110_1 to 110_N through the image control server 120. The number recognition server 1000 controls the overall operation of the server. The number recognition server 1000 analyzes the traveling image photographed by the CCTV according to the embodiment of the present invention and recognizes the vehicle, (1100). The number recognition server 1000 may further include an image processing unit 1200, a vehicle recognition unit 1300, a number recognition unit 1400, and a database unit 1500.

The image processing unit 1200 receives the traveling image from the video monitoring server 120 and performs image processing on the received image to generate a converted image for vehicle and car number recognition. As an example, the image processing unit 1200 receives a plurality of frames per second (for example, 30 frames per second) as a running image and extracts some of the still image images for recognition. For example, it is possible to extract 6 to 10 still image images per second, which can be adjusted according to the hardware performance of the number recognition server 1000. As an extraction method, it is possible to sequentially extract frames received sequentially. For example, when extracting 10 still images per second, one frame may be selected for every three frames of the traveling image to extract a still image have.

In addition, the image processing unit 1200 performs image conversion processing on the extracted image. For example, it is possible to perform an image correction operation on the extracted image through contrast, and this can be performed through a processing operation such as applying a histogram equalization technique. In addition, it is possible to perform still image degaussing by applying a morphology technique such as erosion and expansion.

The vehicle recognition unit 1300 may analyze the image-processed image to perform a vehicle recognition operation. For example, an optical flow technique can be applied to detect the vehicle motion to recognize the vehicle, and also to detect the running direction of the vehicle.

On the other hand, the number recognition unit 1400 can perform a process for recognizing a license plate and a number recognition according to a recognition result from the vehicle recognition unit 1300. For example, a process for license plate recognition, It is possible to perform only the image in which the vehicle motion is detected among the images. The number recognizing unit 1400 can estimate the position of the license plate by applying a binarization and contour extraction technique using the OTSU algorithm, extract individual characters from the position of the estimated license plate, recognize the extracted individual characters The number of the vehicle can be recognized.

The database 1500 may store various types of information in the number recognition server 1000. For example, when the images in which the vehicle motion to be used for the number recognition is detected are stored in the database unit 1500, or the images of the license plate detected from the recognized vehicle and the characters extracted from the license plate, And the like may be stored in the database unit 1500. In addition, various data for driving the number recognition server 1000 may be stored together.

According to the embodiment of the present invention described above, an image having a plurality of frames per second taken from the CCTV is subjected to image processing, and then a vehicle recognition operation is performed. Then, a vehicle license plate and number recognition By performing the operation, it is possible to improve the recognition rate of the vehicle and the number by using the general CCTV, and to improve the surveillance ability by minimizing the monitoring blind spot.

FIG. 2 is a block diagram illustrating a specific implementation of the image processor of FIG. 1. As illustrated in FIG. 2, the image processor 1200 may include a driving image receiver 1210, a still image extractor 1220, and an image converter 1230.

The driving image receiver 1210 may receive the driving image photographed by the CCTVs 110_1 to 110_N and decoded by the image control server 120. In addition, the still image extracting unit 1220 extracts a still image from the received traveling image having a predetermined frame, and extracts a still image from the traveling image received at 30 frames per second according to a predetermined extraction method (e.g., sequential extraction) And extracts a still image (e.g., one frame per 3 frames to 10 still images per second).

The image conversion unit 1230 performs an image conversion process on the extracted image. For example, a gray scale conversion operation may be performed on the extracted image, and an image correction operation using contrast may be performed by a processing operation such as applying a histogram equalization technique to the extracted image. The converted image may be used for vehicle recognition and vehicle number recognition, and the image converted image may be stored in a database inside the number recognition server 1000.

3 is a block diagram illustrating a specific implementation of the vehicle recognition unit of FIG. 1. As illustrated in FIG. 3, the vehicle recognizer 1300 may include an image extractor 1310, a motion detector 1320, and a driving direction detector 1330.

The image extracting unit 1310 may use a technique of extracting a different image to detect whether a vehicle exists in the transformed image. For example, the difference between the previous frame and the current frame can be extracted, and the recognition rate can be improved by distinguishing between the vehicle image and the background image.

The motion detection unit 1320 and the traveling direction detection unit 1330 can detect the motion sensing operation and the traveling direction of the vehicle based on the image extraction result. Although the motion detection unit 1320 and the traveling direction detection unit 1330 are illustrated as different functional blocks, the detection and detection operations may be performed by the same functional blocks. The vehicle recognition result may be generated depending on the distinction between the vehicle image and the background image, and whether motion is detected based on the extracted difference between the previous frame and the current frame. In addition, the optical flow method extracts the feature points of the image, finds the motion of the motion vector, tracks the direction, and extracts the direction values of the previous frame and the current frame to detect the running direction of the vehicle have. The vehicle recognition result may be derived based on the image processing operation of the vehicle recognition unit 1300, and the vehicle number recognition process may be performed using the recognition result.

4A and 4B are block diagrams illustrating exemplary implementations of the number recognition unit of FIG. 1. As shown in FIGS. 4A and 4B, the number recognition unit 1400 may include a license plate detector 1410 and a number recognition processor 1420. The license plate detector 1410 includes an image corrector 1411, a license plate position estimator 1412, a license plate extractor 1413, a shaping unit 1414, a character extractor 1415, and a license plate classifier 1416. It may include. The number recognition processing unit 1420 may include a character recognition unit 1421, a filtering unit 1422, and a result output unit 1423.

The number recognition unit 1400 may perform a process for recognizing the vehicle number when the vehicle recognition result is generated from the vehicle recognition unit 1300. [ That is, a process for recognizing the car number can be selectively performed when motion is detected, such as when a vehicle exists from an image subjected to image conversion processing or when a vehicle moves. The license plate detecting unit 1410 may perform an operation of detecting a license plate of the vehicle from the image in response to the vehicle recognition result. To this end, the converted image from the image processing unit 1200 may be provided to the license plate detecting unit 1410 have. In addition, the number recognition processor 1420 can recognize the individual characters described in the sensed license plate and output the number recognition result by filtering the final car number.

A detailed implementation example of the license plate detector 1410 and the number recognition processor 1420 will be described with reference to FIG. 4B.

The image correction unit 1411 can perform image correction through contrast by applying a histogram equalization technique or the like to the received image, and can remove erroneous images from the still image by applying erosion, expansion morphology, have.

The license plate position estimating unit 1412 estimates the position of the license plate included in the image from the image-processed image. In performing the position estimation operation, the position of the license plate is estimated from the images recognized as vehicles by applying the binarization and contour extraction technique using the OTSU algorithm. The license plate extracting unit 1413 extracts an actual license plate by referring to the estimated position of the license plate. For example, the license plate extracting unit 1413 corrects the license plate slope for license plate extraction. The license plate extracting unit 1413 applies a trigonometric function, And applying the sharpening technique and the median filter to improve the image of the license plate. In addition, binary and contour extraction techniques using the OTSU algorithm can be applied to extract the actual plate.

The shaping unit 1414 can shape the license plate by applying a geometric shaping technique using extraction information (or license plate position information) of the license plate. Further, the character extracting unit 1415 extracts individual characters based on the number plate formatting result, and the license plate classifying unit 1416 shapes character positions based on the extracted individual characters and performs sorting operation by the pattern, (E.g., old license plate, new license plate, etc.). Based on the above operation, the license plate detection result is generated and provided to the number recognition processor 1420.

On the other hand, the number recognition processor 1420 carries out a number recognition processing operation to determine the actual vehicle number based on the license plate detection result, and outputs the number recognition result accordingly. The character recognition unit 1421 recognizes the individual characters of the extracted license plate, and character recognition can be performed by applying, for example, a binarization and a neural network technique. The filtering unit 1422 filters and outputs the final vehicle number by applying the recognized character and the formatted character position, and the result output unit 1423 outputs the recognition result of the vehicle number based on the filtered information do.

Each of the functional blocks illustrated in FIGS. 4A and 4B may be implemented in hardware or software, and the functional blocks illustrated in FIGS. 4A and 4B may be used to separately describe the operation according to an embodiment of the present disclosure. Although shown separately from one another, operations performed in at least two functional blocks may be performed in series in one functional block or one software.

5 is a diagram illustrating an example of extracting a license plate according to an embodiment of the present invention.

As shown in (a) of FIG. 5, the vehicle recognition and image extraction may be performed. For example, the vehicle and the vehicle number may be recognized in consideration of the photographing area passing time of the vehicle by recognizing the motion of the vehicle and using the motion recognition result. Allow frames for recognition. A part of the still image for recognition may be extracted based on the result of the motion recognition from a plurality of frames in which the movement of the vehicle is photographed. As shown in FIG. Allows extraction of images.

On the other hand, as shown in (b) of FIG. 5, it is possible to improve the recognition probability of the vehicle number by performing an image processing operation on the recognition still image. For example, in addition to the gray scale conversion operation on the extracted image, a bright / contrast improvement processing operation on the gray scale converted image may be performed.

In addition, as shown in (c) of FIG. 5, in order to extract the license plate (or license plate image) from the still image for recognition, the license plate may be extracted through horizontal and vertical operations of the image image, for example, After extraction, the license plate can be extracted by serially extracting the vertical regions.

6 is a diagram illustrating an example of recognizing a vehicle number according to an embodiment of the present invention.

As shown in (a) of FIG. 6, an unnecessary portion of an image and a correction of a text size may be performed for character recognition. First, by removing the license plate noise, noise components that may be unnecessary or obstructive to the recognition of characters may be removed, and then, a process for correcting the character size and separating operation for each character may be performed.

In addition, as shown in FIGS. 6B and 6C, the vehicle number recognition operation may be performed by applying the directional vector together with the pattern recognition and processing operations. For example, the vehicle number recognition operation can be performed by performing a pattern matching process with an existing character and performing a pattern matching character recognition process as shown in FIG. 6B, and also as a false recognition rate as shown in FIG. 6C. Redirection processing can be performed by applying a directional vector to this high character. For example, the recognizing process may be applied to the main misrecognition character, and the vehicle number recognition operation may be performed through the process of reanalyzing the directional vector.

7 is a block diagram illustrating an embodiment of a vehicle number recognition system according to another embodiment of the present invention. As shown in FIG. 7, the vehicle number recognition system 200 according to another embodiment of the present invention may include a number recognition server 230. In FIG. 7, the vehicle number recognition system 200 includes a plurality of external cameras (for example, CCTVs 210_1-210_N), a video control server 220, an intelligent search server 240, and an information provision server in addition to the number recognition server 230. An example further includes 250, a context propagation server 260, and a GIS 270. The video control server 220 is connected to a plurality of CCTVs 210_1-210_N through wired and wireless communication networks, etc., and can receive and decode images captured by the plurality of CCTVs 210_1-210_N, and also internal control means. A control signal may be output through the control signal to control the photographing operation of the plurality of CCTVs 210_1 to 210_N.

The number recognition server 230 may perform a vehicle recognition and a vehicle number recognition operation in the same or similar manner as in the above-described embodiment. In an embodiment of the present invention, the number recognition server 230 receives the driving image of the actual vehicle photographed by CCTV as a predetermined frame per second, and the vehicle number through an image processing operation for some frames of the received driving image Output the recognition result.

The vehicle number recognition result from the number recognition server 230 may then be used as information for city management or as information for immediately responding to a crime using a vehicle. To this end, the intelligent search server 240 receives the vehicle number recognition result from the number recognition server 230, receives various database information provided from the information providing server 250, and is present in the captured image. A search operation for determining whether the vehicle corresponds to a wanted vehicle or a delinquent vehicle may be performed based on a result of the number recognition of the vehicle. The information providing server 250 may include various types of servers, such as a police agency server that provides information for real-time inquiry of a wanted vehicle and a delinquent information server that provides information for real-time inquiry of a delinquent vehicle. The intelligent search server 240 receives a search operation by comparing the recognized vehicle number with information stored in the problem vehicle database, and outputs an event indicating this when an event such as the same vehicle number exists in the problem vehicle database. Provided to the propagation server 260.

The situation propagation server 260 propagates the situation through an approval procedure for the received event. In addition, information for predicting a moving path of the problem vehicle may be preset and constructed, and may operate in conjunction with the GIS 270 to display the moving path of the problem vehicle. The user who manages the situation propagation server 260 may perform an operation of approving the situation propagation upon receiving the event, and in response to the situation propagation approval, the situation propagation server 260 may be configured to track the target according to a preset event scenario. Vehicle information is shared with other servers. For example, the situation propagation server 260 may deliver the event to another integrated video management server (not shown) as another server, and together with various images indicating a vehicle image, a vehicle number, a violation status, and a recognition place in a shared database. By storing the information, it is possible to refer to the information for immediate response when an event occurs in another server. The integrated video management server may deliver the event back to the video control server 220 (or the situation propagation server 260 may deliver the event directly), and the GIS as the event is delivered to the GIS 270. 270 may display the current location of the problem vehicle, the moving route, and the predicted predicted route on the map by referring to the information stored in the shared database.

8 is a flowchart illustrating a method of operating a vehicle number recognition system according to an embodiment of the present invention. 8 shows a process of recognizing a vehicle license plate from a driving image photographed from CCTV.

First, according to an embodiment of the present invention, without performing a sensor-based license plate detection operation (or vehicle recognition operation), the vehicle license plate is detected from the driving image captured from the commonly used CCTV. As a moving image having a predetermined frame photographed from CCTV is input (S11), and some still images are selected according to a predetermined extraction method (for example, sequential extraction) from the driving image, still images by frame as image recognition plates. The image is input as a means for performing a later process (S12).

In addition, a conversion process is performed on the images for license plate recognition, for example, a gray scale conversion operation may be performed on the images for license plate recognition (S13), and contrast may be achieved by a processing operation such as applying a histogram equalization technique. An image correction operation may be performed (S14). On the basis of the image converted image as described above, it can be detected whether the vehicle exists in the image, and the process of selectively identifying the license plate for the image detected by the vehicle is performed. The detection operation of whether the vehicle exists in the corresponding image may be performed by detecting the motion and driving direction of the vehicle by applying an image extraction operation for distinguishing the vehicle image from the background image, and an optical flow technique. .

For example, in order to extract the license plate, an operation of extracting a horizontal / vertical region from the image is performed (S15). For example, after extracting a horizontal region, a vertical region is sequentially extracted, and thus a license plate image existing in the image is extracted. It may be (S16).

9 is a flowchart illustrating a method of operating a spatial information system according to another embodiment of the present invention. 9 shows a process of recognizing a vehicle number from the extracted vehicle license plate image.

As shown in FIG. 9, the extracted license plate image is input to a function block for number recognition (S21). In order to recognize the vehicle number, first, the received license plate image may be subjected to histogram equalization technique to correct an image through contrast, or to remove noise from a still image image by applying erosion or expansion morphology technique (S22). ).

Subsequently, the position of the license plate included in the image is estimated from the image processed image, the actual license plate can be extracted by applying OTSU binarization and contour extraction technique, and if the actual license plate is extracted, the license plate is applied by applying a geometric shaping technique. Individual characters in the license plate may be divided through the standardizing operation (S23).

The vehicle number is recognized based on the processing result as described above (S24), for example, a character recognition operation is performed by applying a neural network method of binarization and a mesh network method, thereby providing data related to the recognized vehicle number to the outside. Together, the captured image, the extracted license plate image, and the vehicle number information may be stored in an internal database or stored in a common database available by an external server (S25).

While the present invention has been described with reference to exemplary embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims. Accordingly, the true scope of the present invention should be determined by the technical idea of the appended claims.

Claims (7)

Receiving video from one or more cameras in a first frame per second;
Extracting a still image corresponding to a second frame per second from the received video;
Analyzing the still image to output a vehicle recognition result based on at least one of a motion detection and a driving direction detection result;
The license plate is extracted through horizontal and vertical extraction operations from a still image recognized as a vehicle, and the license plate shaping is performed according to the geometric shaping technique, and the individual characters existing in the standard license plate are extracted. Detecting the license plate; And
Outputting a recognition result of the vehicle number by performing a number recognition processing operation on the extracted individual characters,
Detecting the license plate,
Performing image correction including noise cancellation on the still image;
Estimating the position of the license plate by applying a contour extraction technique;
Extracting a license plate by performing a slope correction with respect to a contour line;
Shaping the license plate by applying the geometric shaping technique using the extraction result of the license plate; And
And extracting the individual characters in the license plate based on the license plate shaping result. The method of claim 1,
And an image processing step of correcting the image through the gray scale conversion and the contrast with respect to the extracted still image. The method of claim 1,
Detecting the license plate and outputting a recognition result of the vehicle number are selectively performed on a still image recognized as the vehicle exists. delete A vehicle number recognition system comprising a number recognition server,
The number recognition server comprises:
Receives a video from one or more cameras at a first frame per second, extracts a still image corresponding to a second frame per second from the received video, and images through gray scale conversion and contrast for the extracted still image An image processing unit which performs image processing including a correction operation;
A vehicle recognition unit configured to analyze the still image and output a vehicle recognition result based on at least one of a motion detection and a driving direction detection result; And
The license plate is extracted through horizontal and vertical extraction operations from a still image recognized as a vehicle, and the license plate shaping is performed according to the geometric shaping technique, and the individual characters existing in the standard license plate are extracted. And a number recognition unit for detecting the license plate and outputting a recognition result of the vehicle number by performing a number recognition processing operation on the extracted individual characters,
The number recognition unit,
The position of the license plate is estimated by applying the contour extraction technique, the license plate is extracted by performing the slope correction on the basis of the contour, and the license plate is shaped by applying the geometric shaping technique using the extraction result of the license plate, and the license plate Vehicle number recognition system, characterized in that for extracting the individual characters in the license plate based on the results of the standardization.
delete delete

Images