KR101756848B1 - Unlawfulness parking and no standing control system and method thereof - Google Patents

Abstract

The present invention relates to an illegal parking management system, which is located at an illegal parking control site, performs a video signal processing on a camera section for photographing a vehicle to be controlled and a video image taken by the camera section, One or more detection servers for outputting intermittent data including vehicle number data and photographing information recognized by the license plate recognizing unit, a control server for transmitting a control signal to the detection server, The control server receives the intermittent data output from the detection server, analyzes the intermittent data received to determine the offending vehicle, and performs follow-up on the illegal parking vehicle by inquiring the vehicle information of the offending vehicle Server, intermittent data of the detection server received from the operation server, A DB server for storing real-time event information processed by the server, a monitoring server for monitoring an image of a vehicle to be intercepted, which is interlocked with the operation server and is photographed by the detection server, a monitor for outputting an image monitored by the monitoring server, . ≪ / RTI >

Description

{UNLAWFULNESS PARKING AND NO STANDING CONTROL SYSTEM AND METHOD THEREOF}

The present invention relates to an illegal parking management system, and more particularly, to an illegal parking management system capable of further improving the vehicle identification number from a captured vehicle image in illegal parking control.

In general, intersection roads or roads with high traffic volume are equipped with prohibited parking areas prohibiting parking or stopping of vehicles, and detention agents who are prohibited from parking or stopping in prohibited parking areas are placed, .

However, in order to manage such prohibited areas, the police officers patrol the forbidden zones individually, and the illegal parking of the vehicles is monitored. Therefore, a lot of manpower costs have been incurred. In the process of charging the fines, There has been a problem in that a problem is caused by imposition.

To solve these problems, illegal parking control systems have been developed to monitor and detect illegal parking vehicles and buses through CCTV.

However, in the illegal parking monitoring method using the conventional CCTV, when the image of the vehicle to be controlled is acquired from the CCTV installed on the road, it may be difficult to accurately recognize the license plate due to the inclination of the license plate attached to the vehicle, .

Also, when the driving position of the vehicle to be intercepted is changed, the geometric shape of the license plate image acquired from the CCTV can be variously changed, and the characters included in the license plate and the rim of the license plate contact each other .

In particular, a smear phenomenon occurs in an image acquired from the CCTV due to the influence of sunlight or a blurring is generated in an image obtained from the CCTV due to the influence of night illumination at a shooting time, The recognition rate of the vehicle number is remarkably low due to the deterioration of the image quality of the photographed image. Therefore, even in the case of the offending vehicle, the vehicle number is unclear and there is a great difficulty in controlling the illegal driving of the illegally parked vehicles.

Korean Patent No. 1403876 (May 29, 2014)

SUMMARY OF THE INVENTION Accordingly, the present invention has been made to solve the above problems, and an object of the present invention is to provide an illegal parking management system capable of accurately recognizing a license plate area included in an image captured from a vehicle to be controlled, I want to.

In addition, the present invention estimates the sharpness of the license plate area included in the photographed image, and selectively sharpens the license plate area according to the estimated sharpness, thereby improving the recognition performance of the license plate having various image qualities. The system has a purpose to provide.

It should be understood, however, that the technical scope of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided an illegal parking management system including a camera unit for photographing a vehicle to be controlled and an image signal processing unit for processing the image captured by the camera unit, And a vehicle license plate recognizing unit for recognizing the vehicle number of the vehicle to be intercepted, wherein the at least one detecting server outputs the intermittent data including the vehicle number data and the photographing information recognized by the license plate recognizing unit, The control server transmits a control signal to output a control signal to photograph the vehicle to be intercepted, receives the intermittent data output from the detection server, analyzes the intermittent data received to determine the infringing vehicle, An operation server for performing follow-up on the parking lot vehicle, A DB server for storing intermittent data of the detection server received by the server and real-time event information processed by the operation server, a monitoring server for monitoring an image of the vehicle to be intercepted, which is interlocked with the operation server and is photographed by the detection server, And a monitor for outputting an image monitored by the server.

The vehicle license plate recognizing unit according to an aspect of the present invention includes an image receiving module for receiving an image from the camera, a binarization module for binarizing an image received from the image receiving module into a colorized binary image, And a license plate area extracting module for extracting the license plate area.

In this case, the binarization module extracts horizontal boundary points of the text stroke candidates from the received image to generate a horizontal stroke boundary point map, generates a segment image by grouping pixels between the text strokes, The color image is generated so as to have different colors.

The license plate recognizing unit according to an aspect of the present invention may further include a sharpening module for sharpening the character blobs included in the license plate area, and a character recognition module for recognizing characters corresponding to each of the sharpened character blobs .

The character recognition module according to an aspect of the present invention may include a neural network character recognizer and a hidden Markov model recognizer.

The character recognition module recognizes characters corresponding to each of the character blobs using the neural network character recognizer, and recognizes character blobs not recognized by the neural network character recognizer using the hidden markov model recognizer And recognizes the information.

The neural network character recognizer according to an aspect of the present invention may be implemented by a combination of at least one of a region-specific recognizer, a character-specific recognizer, and a numeric-only recognizer.

The sharpening module estimates the sharpness of the character blobs included in the license plate area. If the estimated sharpness is lower than the reference sharpness, the sharpening module applies the sharpening algorithm to the license plate area, And sharpness.

The camera unit according to an aspect of the present invention may be configured as a PTZ (Pan Tilt Zoom) camera.

In this case, when the camera unit is configured at an intersection where two or more roads intersect, the operation server sets a plurality of photographing regions, a photographing order, and a photographing time unit for both sides of each road, generates a control signal according to the set value, It is possible to accumulate intermittent data for each photographing area transmitted from the detection server and judge a vehicle exceeding a preset time as the violating vehicle for the same vehicle in the accumulated intermittent data.

The illegal parking management system according to an embodiment of the present invention estimates the brightness of the license plate area included in the photographed image of the vehicle to be enforced in the illegal parking license control and selectively sharpens the license plate area according to the estimated brightness, So that it is possible to improve the recognition performance of the license plate.

In addition, since the image of the vehicle normally photographed is not sharpened, excessive sharpening can be prevented.

In the present invention, the horizontal boundary points of the text stroke candidates are extracted, and the SBT algorithm is applied based on the extracted horizontal boundary points to group the pixels between the text strokes, so that the text region can be accurately detected even in a blurred or distorted image.

In addition, in the present invention, each grouped text is colored with different colors in the process of text coloring, so that the touching strokes can be more clearly separated, and the text area can be accurately detected.

In the present invention, the neural network character recognizer further includes a region specific recognizer as compared with the conventional art, thereby reducing the load on the character recognition and shortening the character recognition performance time.

According to the present invention, when a character or a number that is not recognized by the neural network character recognizer occurs, only characters or numbers that are not recognized by the neural network character recognizer are recognized using the hidden Markov model recognizer, Can be shortened.

1 is a block diagram schematically showing an illegal parking management system according to an embodiment of the present invention.
2 is a block diagram showing the detection server of FIG.
3 is a block diagram showing a configuration of a license plate recognition unit according to an embodiment of the present invention.
4 is a view for explaining the operation of the binarization module according to an embodiment of the present invention.
5 is a view for explaining operations of a binning module and a license plate area extraction module according to an embodiment of the present invention.
FIG. 6 is a diagram illustrating a change in sharpness estimated as the degree of blurring in a license plate area changes according to an embodiment of the present invention. FIG.
7 is a view for explaining a process of dividing each of characters included in a license plate area according to an embodiment of the present invention.
8 is a diagram illustrating a neural network structure for license plate character recognition according to an embodiment of the present invention.
9 is a block diagram of a hidden markov model (HMM) for license plate character recognition according to an embodiment of the present invention.
10 is a flowchart showing an illegal estimation difference management method according to an embodiment of the present invention.
11 is a diagram showing the operation of the binarization module.
12 and 13 are views showing the operation of the sharpening module.
14 is a diagram showing the operation of the character recognition module.
FIG. 15 is a schematic view for explaining an interruption process at an intersection using a PTZ camera according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings, which will be readily apparent to those skilled in the art. The present invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In order to clearly illustrate the present invention, parts not related to the description are omitted, and similar parts are denoted by like reference characters throughout the specification.

Throughout the specification, when a part is referred to as being "connected" to another part, it includes not only "directly connected" but also "electrically connected" with another part in between . Also, when an element is referred to as "comprising ", it means that it can include other elements as well, without departing from the other elements unless specifically stated otherwise.

1 is a block diagram schematically showing an illegal parking management system according to an embodiment of the present invention.

Referring to FIG. 1, the illegal parking management system of the present invention may include a detection server 10, an operation server 20, a DB server 30, a monitoring server 40, and a monitor 400.

The detection server 10 is located at a control site and can be fixed or mobile so as to photograph the illegally parked vehicle and recognize the vehicle number of the target vehicle from the captured image.

Then, the detection server 10 may transmit the intermittent data including the intermittent time data, that is, the intermittent time and the photographing date, to the operation server 20 together with the recognized vehicle number data.

Although not shown in the figure, the detection server 10 includes a communication module to perform mutual data communication with the operation server 20 through a wired or wireless data network, and receives a control signal transmitted from the operation server 20 Can be received and controlled.

For example, the detection server 10 may receive a control signal for controlling the physical operation of the camera from the operation server 20, and may transmit and receive the request command and the corresponding response information, which are driven accordingly.

For example, the detection server 10 recognizes the license plate area of the vehicle and analyzes the vehicle number from the license plate area. In the vehicle number analysis process, A sharpening process, a distortion correction process, and the like, so that more accurate intermittent data can be derived.

At least one detection server 10 may be provided for each predetermined zone and may include unique identification information and be identifiable by the operation server 20 by transmitting identification information during transmission of the intermittent data.

The operation server 20 can receive the intermittent data transmitted from the at least one detection server 10 through a wired or wireless communication network.

The operation server 20 identifies the detection server 10 through the identification information of the intermittent data received from the detection server 10, and classifies and manages the detection server 10 according to the area.

Then, the operation server 20 analyzes the intermittent data received from the detection server 10 to determine the offending vehicle, and inquires of the offending vehicle about the offending vehicle in connection with the related agency, for example, the police information server, Can be extracted.

In addition, the operation server 20 can calculate the violation of the traffic regulation and the penalty or penalty corresponding thereto, and notify the violation target borrower together with the video included in the cut-off data and the violation time.

The operation server 20 can generate and output a control signal for controlling one or more detection servers 10 and transmits the control signals to the detection server 10 to physically control the camera of the detection server 10 According to the Illegally Divided Vehicle, the vehicle to be violated is photographed or the vehicle number is recognized from the shot image.

The DB server 30 stores data received from one or more detection servers 10 to the operation server 20, and can classify and / or statistically store the stored data.

The DB server 30 can transmit response information corresponding to the request information of the operation server 20 to the operation server 20 according to the request of the administrator and can transmit event information processed in real time in the operation server 20 Can be stored.

The DB server 30 is not shown in the figure, but may be composed of a plurality of databases.

The monitoring server 40 interlocks with the operation server 20 to detect an image captured by the detection server 10 in real time and outputs the image to the monitor 400 so that the administrator can check the illegal state So that it can be watched and monitored.

In addition, the illegal parking management system according to the present invention is not shown in the drawings, but a speaker or the like for carrying out vehicle guidance and warning announcement may be installed adjacent to the site, that is, the detection server 10.

2 is a block diagram illustrating a detection server according to an embodiment of the present invention.

2, the detection server 10 according to an exemplary embodiment of the present invention may include a camera unit 100 and a license plate recognition unit 200, The vehicle license plate recognition unit 200 analyzes the images and recognizes the vehicle numbers of the respective vehicles 50. The intermittent data including the recognized vehicle number information is transmitted to the operation server 20 through a wired or wireless communication network, Lt; / RTI >

The camera unit 100 is composed of one or more cameras and is shown as a fixed type in FIG. 2, but the present invention is not limited thereto. The camera unit 100 may have mobility in a form mounted on a vehicle or the like.

The camera unit 100 may be, for example, a closed-circuit television (CCTV) camera as one of various image equipment for photographing an image. The camera unit 100 may photograph a vehicle 50 with a license plate 60, .

The camera unit 100 may transmit an image of the photographed vehicle 50 to the license plate recognizing unit 200.

The car number recognizing unit 200 can recognize and analyze the car number from the car number plate 60 of the car 50 by receiving the image of the car 50 photographed from the camera unit 100. [

2, the camera unit 100 and the license plate recognition unit 200 are implemented as a single device for convenience of explanation. However, the camera unit 100 and the license plate recognition unit 200 200 may be spaced apart and located at different locations.

3 is a block diagram showing a configuration of a license plate recognition unit according to an embodiment of the present invention.

3, the license plate recognition unit 200 includes an image receiving module 210, a binarization module 230, a license plate area extraction module 250, a sharpening module 270, and a character recognition module 290, . ≪ / RTI >

As used herein, a module may refer to a functional or structural combination of hardware for performing technical ideas and software for driving the hardware according to an embodiment of the present invention. For example, the module may refer to a logical or functional unit of a predetermined program code and a hardware resource to be executed by the program code, and does not necessarily mean a physically connected program code or a kind of hardware.

The image receiving module 210 receives an image of the vehicle 50 from the camera unit 100 through a wired or wireless communication network and outputs the received image as a grayscale image RI. That is, the value of each pixel in the image RI can represent only gray intensity information without color information.

The binarization module 230 processes the image RI of the vehicle 50 output from the image reception module 210 and outputs the binary image BI according to a result of the processing.

The binarization module 230 according to the present embodiment can binarize the received image RI using the SBT (Stroke Boundary Transform) algorithm and output the colorized segment binary image BI.

4 is a diagram for explaining the operation of the binarization module 230 according to an embodiment of the present invention.

4, the binarization module 230 receives a gray scale image (FIG. 4A) from the image receiving module 210, and uses a DoG (Different of Gaussian) filter to extract the text stroke candidates The horizontal boundary points are extracted to generate a horizontal stroke boundary map (FIG. 4 (b)). Then, the binarization module 230 generates a segment image (FIG. 4C) by grouping the pixels between the text strokes of FIG. 4B by applying the SBT algorithm. More specifically, the binarization module 230 may group pixels between text strokes through vertical profiling of horizontal boundary line connections.

The binarization module 230 may generate a binary image (Fig. 4 (d)) colored so that each of the segments of the segment image (Fig. 4 (c)) has a different color.

At this time, the binarization module 230 allows the text to be colored with a clear color such as a complementary color relationship in the adjacent text. As a result, the boundaries of adjacent texts can be clearly distinguished.

As described above, in the present invention, the horizontal boundary points of the text stroke candidates are extracted and the SBT algorithm is applied based on the extracted horizontal boundary points to group the pixels between the text strokes, so that the text region can be accurately detected even in a blurred or distorted image .

4E shows a case where the text strokes are in a separate point. The binarization module 230 groups the pixels between the text strokes by applying the SBT algorithm, And generates a colorized segment binary image. By coloring each grouped text in different colors in the process of coloring, the touching strokes can be more clearly separated and the text area can be accurately detected.

The license plate region extraction module 250 may extract the license plate area EI from the binary image BI output from the binarization module 230. [

The sharpening module 270 can estimate the sharpness of the character blobs included in the license plate area EI extracted through the license plate area extraction module 250. [ For example, the sharpening module 270 may use a perspective transform to normalize the size of each of the character blobs and correct the rotation.

The sharpening module 270 may output the sharpening plate area SI in which size normalization and rotation correction have been performed.

For example, if the sharpness of the estimated character blobs is lower than the reference sharpness, the sharpening module 270 may sharpen the character blobs by applying a sharpening algorithm. The sharpening algorithm may be an unsharp masking algorithm, but the present invention is not limited thereto.

The sharpening module 270 may also apply an image enhancement algorithm to the license plate area SI. In this case, the image enhancement algorithm may be an algorithm using a sigmoid transformation function, but the present invention is not limited thereto.

The character recognition module 290 can recognize each of the characters corresponding to each of the character blobs from the sharpening plate area SI output from the sharpening module 270. [ At this time, the character recognition module 290 may use a neural network structure to recognize each of the characters, but the present invention is not limited thereto.

5 is a view for explaining operations of a binning module and a license plate area extraction module according to an embodiment of the present invention.

The binarization module 230 may receive the image RI of the vehicle 50 from the image receiving module 210. [ 5A shows an image RI of the vehicle 50 received from the image receiving module 210 as a first image 5-1.

The binarization module 230 generates a second image 5-2 that is a horizontal stroke boundary point map of the text stroke candidates using the DoG filter for the first image 5-1, The third image 5-3 can be generated by grouping the pixels between the strokes and coloring the grouped pixels (text) with different colors. The third image 5-3 is output to the license plate area extraction module 250.

At this time, the binarization module 230 may generate the fourth image 5-4 by applying median filtering, which is a nonlinear filtering technique, to reduce the noise of the third image 5-3, Operation may be performed in license plate area extraction module 250.

The license plate region extraction module 250 receives the fourth image 5-4 output from the binarization module 230 and extracts the license plate area EI from the received fourth image 5-4. 5E and 5F illustrate a fifth image 5-5 and a sixth image 5-6 as an example of a process of extracting the license plate area EI.

The fifth image 5-5 may include a plurality of character blobs extracted by the license plate area extraction module 250.

Specifically, the license plate area extraction module 250 can extract a plurality of character blobs estimated as characters (or numbers) from the fourth image 5-4 received from the binarization module 230. The license plate area extraction module 250 according to the present embodiment can use the reference character stroke size information to extract a plurality of character blobs. The reference character stroke size information may be stored in the license plate area extraction module 250 or the license plate recognition section 200.

Then, the license plate area extraction module 250 interprets the arrangement form of the plurality of character blobs included in the fifth image 5-5 and extracts the license plate area as shown in the sixth image 5-6 have.

For example, when the license plate layout type is formed by four numbers, the license plate region extraction module 250 extracts an area including four consecutive blobs 60A to 60D of similar size among a plurality of character blobs And can be extracted as a plate plate area (EI).

Also, the license plate area extraction module 250 can output and transmit the extracted license plate area EI to the sharpening module 270. The license plate area extraction module 250 may output the sixth image 6-6 including the extracted license plate area EI according to the embodiment.

5F shows a sixth image 5-6 in which the license plate area EI is displayed on the first image 5-1. However, according to the embodiment, the sixth image 5-6, The license plate area EI may be displayed on the third image 5-3 or the fourth image 5-4.

FIG. 6 is a diagram illustrating a change in sharpness estimated as the degree of blurring in a license plate area changes according to an embodiment of the present invention. FIG.

The sharpening module 270 receives the license plate area EI output from the license plate area extraction module 250 and can estimate the sharpness S of a plurality of character blobs included in the license plate area EI. The plurality of character blobs may be blurred due to illumination imbalance, camera shake 100, inaccurate focus, and / or movement of vehicle 50.

6, examples of the sharpness S estimated according to the blurring degree of the plate plate area EI are shown. It can be seen that the higher the blurring degree of the plate plate area EI is, the lower the estimated sharpness S is have.

The sharpening module 270 may estimate the sharpness S of a plurality of character blobs included in the license plate area EI using a sharpness estimation algorithm.

The sharpness estimation algorithm is based on the difference between the brightness value of the edge pixel and the brightness value of the right pixel of the edge pixel in the intermediate value filtered image of the license plate area EI and the brightness value of the edge pixel and the brightness value of the left pixel of the edge pixel The algorithm may be a method of estimating the sharpness S using a difference in brightness value, but the present invention is not limited thereto.

Since the degree of blurring and the estimated sharpness S shown in FIG. 6 are arbitrary values for convenience of explanation, the value of the sharpness S estimated according to the degree of blurring may vary depending on the kind of sharpness estimation algorithm.

The sharpening module 270 may also apply an algorithm for size normalization and rotation correction to the license plate area EI to normalize the size of each of the plurality of character blobs included in the license plate area EI and to correct the rotation have. The algorithm may be a perspective transform algorithm, but the present invention is not limited thereto.

7 is a view for explaining a process of dividing each of characters included in a license plate area according to an embodiment of the present invention.

As described above, the sharpening module 270 can sharpen the received license plate area EI and output the sharpening license plate area SI therefrom.

The character recognition module 290 may divide the plurality of character blobs included in the sharpening plate area SI and recognize each of the characters included in each of the plurality of divided character blobs.

Referring to FIG. 7A, the sharpening module 270 estimates the sharpness S of the received license plate area EI, and compares the estimated sharpness S with the reference sharpness. For example, assuming that the reference sharpness is 0.4, the sharpness S of the plateau area EI shown in Fig. 7A is 0.67.

Since the sharpness S is larger than the reference sharpness, the sharpening module 270 may not apply the sharpening algorithm to the edge pixels of the plurality of character blobs included in the license plate area EI.

Referring to FIG. 7B, since the sharpness S of the license plate area EI is 0.27 and is lower than the reference sharpness, the sharpening module 270 applies a sharpening algorithm to the edge pixels to improve the local contrast Thereby sharpening a plurality of character blobs.

The sharpening algorithm according to the present embodiment may be an unsharp masking filtering algorithm capable of high-speed processing for real-time license plate recognition, but the present invention is not limited thereto.

Referring to FIGS. 7A and 7B, the sharpening module 270 may apply an image enhancement algorithm to the license plate area EI to improve the global contrast of the license plate area EI. The image enhancement algorithm may be an image enhancement algorithm using a sigmoid transform function.

The sharpening module 270 may output the sharpening plate area SI to the character recognition module 290. [ The sharpening plate area SI may refer to a plate area to which a sharpening algorithm and / or an image enhancement algorithm is applied to the plate area EI.

Character recognition module 290 may estimate the type of sharpening plate area SI using various types of license plate types.

The character recognition module 290 may then divide each of the plurality of character blobs included in the sharpening plate area SI based on the type of the estimated sharpening plate area SI. 7 (a) and 7 (b) show a form in which a plurality of character blobs are divided.

8 is a diagram illustrating a neural network structure for license plate character recognition according to an embodiment of the present invention.

The character recognition module 290 may use the back propagation neural network character recognizer as shown in FIG. 8 and may be included in each of the plurality of divided character blobs as shown in FIGS. 7 (a) and 7 (b) It is possible to recognize each of the plurality of characters. Backpropagation neural network character recognizers are examples of non-linear statistical data modeling tools. Backpropagation neural network character recognizers can be used to model the complex relationships between inputs and outputs. Training letters can therefore be used to model the statistical data.

In addition, the neural network character recognizer can be implemented as a region-specific recognizer, a character-specific recognizer, a numeric-specific recognizer, and a local-character-number hybrid recognizer.

The character recognition module 290 recognizes a plurality of characters using a region-specific recognizer, a character-specific recognizer, a numeric-specific recognizer, and / or a local-character-number mixer recognizer according to the positional information of the plurality of divided character blobs according to license plate naming rules Respectively.

In the present invention, as compared with a back propagation neural network character recognizer having a conventional character / numeric-only recognizer, a local-specific recognizer is added, thereby reducing load on character recognition and shortening the time required for character recognition.

9 is a block diagram of a hidden markov model (HMM) for license plate character recognition according to an embodiment of the present invention.

The Hidden Markov Model (HMM) recognizer can be performed at the same time that the character / number is divided and the recognition process is not separated. Thus, using HMM-specific features, words can be recognized even under over-segmentation or unde-segmentation of handwritten words.

In addition, the hidden Markov model recognizer can utilize various structural features as well as statistical features of the pattern to be recognized. Because of this feature, the character recognition module 290 selectively uses the hidden Markov model recognizer only when there are characters or numbers that are not recognized as neural network character recognizers.

According to the present invention, when a character or a number that is not recognized as a neural network character recognizer occurs, recognizing a character or a number that is not recognized by using the hidden markov model recognizer, the character or number can be more accurately recognized and the recognition time can be shortened The effect can be expected.

10 to 14 are flowcharts for explaining the operation of the license plate recognizer 200 according to an embodiment of the present invention.

FIG. 10 is a flow chart showing a license plate recognition method according to an embodiment of the present invention, FIG. 11 is a diagram showing the operation of the binarization module, FIGS. 12 and 13 are diagrams showing the operation of the sharpening module, Is a diagram showing the operation of the character recognition module.

10, the image receiving module 210 of the license plate recognizing unit 200 receives the image RI of the vehicle from the camera unit 100 and outputs it to the binarization module 230 (S100) .

The binarization module 230 binarizes and colors the received vehicle image RI and outputs the colored binary image BI to the license plate area extraction module 250 (S200).

More specifically, as shown in FIG. 11, the binarization module 230 extracts horizontal boundary points of the text stroke candidates using the DoG filter to generate a horizontal stroke boundary map (S210).

Then, the binarization module 230 generates a segment image by grouping pixels between text strokes by applying the SBT algorithm (S220). More specifically, the binarization module 230 may group pixels between text strokes through vertical profiling of horizontal boundary line connections.

The binarization module 230 generates a colorized binary image such that each text of the segment images has a different color (S230).

The license plate region extraction module 250 may extract the license plate area EI from the binary image BI received by the binarization module 230 at step S300.

The license plate area extraction module 250 can detect a plurality of character blob candidates from the binary image BI and extract the license plate area EI using the arrangement form of the detected plurality of character blob candidates.

The sharpening module 270 may estimate the sharpness of a plurality of character blobs included in the extracted license plate area EI (S400).

12, the sharpening module 250 estimates the sharpness of a plurality of character blobs included in the extracted license plate area EI (S410), and the sharpening module 270 Normalizing the size of the plurality of character blobs and correcting the rotation (S420)

The sharpening module 270 compares the estimated sharpness with the reference sharpness (S500). If the estimated sharpness is lower than the reference sharpness, the sharpening algorithm is applied to the plate area EI to sharpen the plurality of character blobs (S600).

13, the sharpening module 270 applies an unsharp masking algorithm to the license plate area EI to sharpen a plurality of character blobs, and according to the sharpening result, (SI) (S610). According to another embodiment, the sharpening module 270 may apply the image enhancement algorithm to the license plate area EI to generate the sharpening license plate area SI (S620)

The image enhancement algorithm may be an image enhancement algorithm through global contrast enhancement using a sigmoid transform function.

The character recognition module 290 can recognize the characters corresponding to each of the plurality of character blobs included in the sharpening plate area SI (S700).

14, the character recognition module 290 estimates the type of the sharpening plate area SI using a plurality of license plate types (S710), and the character recognition module 290 estimates And divides the plurality of character blobs included in the sharpening plate area SI according to the selected type (S720).

Then, the character recognition module 290 recognizes characters corresponding to each of the plurality of divided character blobs using the neural network character recognizer (S730).

At this time, the character blobs not recognized as the neural network character recognizer are recognized using the hidden Markov model recognizer (S740).

FIG. 15 is a schematic view for explaining an interruption process at an intersection using a PTZ camera according to an embodiment of the present invention.

In general, the illegal parking system is composed of a rotary camera for intermittent shooting and a fixed camera for distinguishing the objects to be intercepted. In a fixed camera, when a specific vehicle does not move more than a certain distance, The way to control the vehicle was applied. However, this type of system is composed of different kinds of cameras and at least two cameras must be provided. This increases the installation cost and increases the maintenance cost.

Therefore, the present invention solves the above-mentioned problem by reducing the installation and maintenance costs by allowing the photographing and interception of various directions to be performed while using only one rotary camera.

The present embodiment can be applied to a case where the camera unit is formed at an intersection where two or more roads intersect, wherein the camera unit of the detection server may be a PTZ (Pan Tilt Zoom) camera and is preferably located at the center of the intersection So that photographing can be performed on various surfaces. Such a PTZ (Pan Tilt Zoom) camera can perform upward and downward movement and zooming as well as left and right rotation by a control signal.

The operation server 20 sets a plurality of photographing areas, a photographing order, and a photographing time unit so that both sides of each road can be photographed sequentially at an intersection, generates a control signal according to the set values, Lt; / RTI >

For example, as shown in FIG. 15, the operating server 20 can set up eight shooting regions in total at each of the four road directions in which the camera unit 100 is configured at the intersection intersection, That is, the coordinate control signal can be generated and transmitted to the detection server 10.

In addition, the operation server 20 can set a photographing order and a photographing time unit together with a coordinate control signal. For example, it is possible to set a plurality of photographing regions to photograph in a clockwise or counterclockwise direction in units of one minute, And transmit the generated signal to the detection server 10.

The detection server 10 receiving the control signal controls the driving of the camera unit 100 in accordance with the photographing order and the photographing time unit with respect to the plurality of photographing regions, To the operation server (20).

At this time, the intermittent data may include photographing information photographed in each photographing area and vehicle number information recognized from the target vehicle when the vehicle to be intercepted exists.

The operation server 20 accumulates the intermittent data for each photographing area transmitted from the detection server 10 and analyzes cumulative intermittent data to determine when the same vehicle is repeatedly recognized in each photographing area, It can be designated as a vehicle.

Thereafter, the operation server 20 analyzes the accumulated intermittent data to calculate a total cumulative time of the specified intermittent vehicle, compares the accumulated cumulative time with a preset first time, and if the cumulative time exceeds the first time, Can be judged as a violating vehicle.

Here, the first time may be the time corresponding to the parking or stop violation time specified in accordance with the relevant laws and regulations.

When there is a vehicle to be controlled that is judged to be a violation vehicle in connection with the enforcement process of the present embodiment, the operation server 20 notifies the related agency, for example, the information about the police server information server, For example, by contacting a contact network of the mobile clerk who controls the adjacent region, and may provide the clerk with a text message (SMS).

One embodiment of the present invention may also be embodied in the form of a recording medium including instructions executable by a computer, such as program modules, being executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. In addition, the computer-readable medium may include both computer storage media and communication media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Communication media typically includes any information delivery media, including computer readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave, or other transport mechanism.

It will be understood by those skilled in the art that the foregoing description of the present invention is for illustrative purposes only and that those of ordinary skill in the art can readily understand that various changes and modifications may be made without departing from the spirit or essential characteristics of the present invention. will be. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description and all changes or modifications derived from the meaning and scope of the claims and their equivalents are to be construed as being included within the scope of the present invention do.

10: detection server
20: Operational server
30: DB server
40: Monitoring Server
100:
200: license plate recognition unit
210: image receiving module
230: Binarization module
250: License plate area extraction module
270: sharpening module
290: Character recognition module

Claims (8)

And a vehicle license plate recognition unit which is located at an illegal parking control site and recognizes a vehicle number of a vehicle to be controlled by performing a video signal processing on a camera unit for photographing a vehicle to be controlled and an image photographed by the camera unit, One or more detection servers for outputting intermittent data including vehicle number data and photographing information recognized through the license plate recognition unit,
The control server transmits a control signal to the detection server to output a control signal for capturing the vehicle to be intercepted, receives the intermittent data output from the detection server, analyzes the intermittent data received to determine the infringing vehicle, An operational server that performs follow-up on illegal parking vehicles through inquiries,
A DB server for storing intermittent data of the detection server received by the operation server and real-time event information processed by the operation server,
A monitoring server interlocked with the operation server and monitoring an image of a vehicle to be intercepted photographed by the detection server,
And a monitor for outputting an image monitored by the monitoring server,
The vehicle license plate recognizing unit,
An image receiving module for receiving an image from the camera unit,
A binarization module for binarizing an image received from the image receiving module into a colorized binary image,
A license plate area extracting module for extracting license plate area of the vehicle from the binary image,
A sharpening module for sharpening the character blobs included in the license plate area, and
Further comprising a character recognition module for recognizing characters corresponding to each of the sharpened character blobs,
Wherein the binarization module extracts horizontal boundary points of the text stroke candidates from the received image to generate a horizontal stroke boundary point map, generates a segment image by grouping pixels between the text strokes, The binary image is colorized so as to have a different color, and the text is colored with a color whose color classification is more clear for the adjacent text,
Wherein the character recognition module recognizes characters corresponding to each of the character blobs using a neural network character recognizer, recognizes character blobs not recognized as the neural network character recognizer using a hidden Markov model recognizer,
The neural network character recognizer may be implemented by a combination of at least one of a region-specific recognizer, a character-specific recognizer, and a numeric-
A loudspeaker for car navigation and warning announcement can be provided to the vehicle to be controlled,
When an intersection where two or more roads intersect is constructed
The camera unit includes:
And a PTZ (Pan Tilt Zoom) camera installed at the center of the intersection and capable of photographing both sides of each road,
The operating server,
A plurality of photographing areas, a photographing order and a photographing time unit are set so that the PTZ camera rotates in any one direction and photographs both sides of each road sequentially, and a control for controlling the movement and photographing of the PTZ camera according to a set value And transmits the generated signal to the detection server. When the same vehicle is repeatedly recognized in each photographing area, the control server accumulates the intermittent data for each photographing area transmitted from the detection server and analyzes the cumulative intermittent data, And when the specified vehicle to be intercepted exceeds a preset time, the vehicle is judged to be a vehicle for violation. delete delete delete delete delete delete delete

Images