KR101451070B1 - Method to recognize license plates by removing recognition error by shadow and license plate recognition system - Google Patents

Abstract

Disclosed are a method for recognizing a vehicle registration plate by removing a recognition error due to a shadow and a recognition system for a vehicle registration plate. The method for recognizing a vehicle registration plate includes a step of converting frame data of a photographed image into black and white data; a step of setting a reference frame by comparing a change amount between frames included in a black and white image converted into the black and white data; a step of detecting the existence of a shadow from the photographed image by using a first change amount between a current frame and a previous frame and a second change amount between the current frame and the reference frame; and a step of recognizing the vehicle registration plate from the photographed image according to the existence of the shadow. The present invention allows a CCTV to accurately recognize a vehicle registration plate by preventing the CCTV from recognizing a wrong vehicle registration plate due to the shadow of other vehicles driving on an adjacent lane.

Description

TECHNICAL FIELD [0001] The present invention relates to a method of recognizing a car number by removing shadow recognition errors, and a car number recognition system. [0002]

The present invention relates to a method of recognizing a car number by eliminating a shadow recognition error and a car number recognition system. More particularly, the present invention relates to a car number recognition system for recognizing a car number of a car traveling in an adjacent lane, A car number recognition system for recognizing a car number by eliminating a recognition error caused by a shadow to prevent the car number from being erroneously recognized by the car number recognition system.

In recent years, the number of vehicles has rapidly increased, and many problems related to vehicles such as traffic accidents, traffic congestion, and speeding vehicles have become social issues. Accordingly, vehicle management, vehicle tracking, automatic toll payment, parking management, Many studies have been carried out to solve problems through license plate recognition in applications such as vehicle management.

In general, application fields of vehicle identification system are applied to fixed, portable, and mobile speed control system, signal violation control system, toll deduction prevention system, parking management system, etc. Digital cameras, lighting devices, grabbing Board, car number recognition main controller, auto iris lens, and vehicle detection sensor.

An important key technique in such a car number recognition system is to obtain a good image in which the main controller can recognize the car number, and the key to this recognition rate is largely dependent on whether the obtained car image is a good image to be recognized. In addition, the accuracy and reliability of the car number recognition algorithm, and the status of the car license plate are given to the recognition success rate.

Typically, in order to obtain a better image of the license plate, it is necessary to adjust the intensity of the strobe, adopt an automatic iris lens, set the season time-zone illuminance in advance, or feedback control the iris with the brightness of the photographed image However, it failed to satisfy the conditions for the weather condition, the sun moving condition, and the cloud moving condition, so that it did not work.

Korean Patent No. 10-1102572 discloses a license plate recognition method. In the above technique, the vehicle entering the setting area is photographed, vertical and horizontal components are extracted from the photographed image, and boundary information is detected through the extracted vertical and horizontal components to form a boundary image. Then, a region where the pixel brightness value changes frequently is extracted from the boundary image, and the cumulative value of the pixel brightness values in the vertical and horizontal directions of the boundary image is detected to detect the distribution of the vertical and horizontal components of the entire vehicle. The rectangle area formed by the vertical and horizontal components of the entire vehicle is selected according to the size and the aspect ratio of the actual license plate, and the character estimation part is compared with the character size ratio of the actual license plate through binarization and labeling in the rectangular area. After determining the difference, the moment about the position of the character estimating portions is calculated to finally determine that it is the license plate area.

However, when car number detection is performed for one or more lanes, the car number may be misrecognized depending on the position or time, and the shade due to the vehicle on the opposite side or the car may be displayed. That is, when the shadow included in the photographed image is recognized as a vehicle, the vehicle number detection is performed, but the vehicle number can not be extracted because it is not an actual vehicle. Alternatively, the vehicle number of the vehicle passing through the adjacent lane may be mistakenly recognized by the shadow included in the photographed image.

In conclusion, the process of recognizing the car number is different from other character recognition, and it is considered to be strong against noise or deformation because it is affected by external environment. In other words, there is a need to create a condition that accurately identifies the car number by removing elements that affect the car number recognition process, such as a shadow.

Korean Registered Patent No. 10-1102572 (registered on December 28, 2011) Korean Registered Patent No. 10-1291302 (Registered on July 24, 2013) Korean Registered Patent No. 10-0770157 (registered on October 19, 2007)

Accordingly, it is an object of the present invention to provide a vehicle lane recognition system and a vehicle lane recognition method in which when a vehicle number is recognized for one or more lanes, a shadow area included in the photographed image is not detected A car number recognition system for recognizing a car number when a car number is recognized when a car number is not recognized,

According to another aspect of the present invention, there is provided a method of recognizing a car number by eliminating a shadow recognition error, comprising: converting frame data of a photographed image into monochrome data; A step of setting a reference frame by comparing the amount of change between frames included in an image, a step of setting a reference frame by comparing a first variation amount between a current frame and a previous frame and a second variation amount between the current frame and the reference frame, And recognizing the vehicle number from the photographed image according to whether the shadow exists or not.

The step of setting the reference frame is a step of comparing the amount of change between the frames included in the monochrome image and setting the current frame as the reference frame if it is determined that the current frame and the previous frame are the same for a predetermined time .

Wherein the step of detecting presence or absence of shadow includes determining that a shadow region exists in the photographed image when the first variation amount is smaller than the second variation amount.

The step of recognizing the car number is a step of extracting a frame at a time point where the shadow does not exist in the photographed image and recognizing the car number.

Here, the point of time when the shadow does not exist is a time point at which the second change amount starts to increase sharply.

Wherein the step of detecting presence or absence of shadow includes determining that there is no shadow region in the photographed image when the first variation amount is similar to the second variation amount.

Meanwhile, a car number recognition system for recognizing a car number included in an image photographed by a plurality of CCTVs transmitted through a CCTV network includes a monochrome conversion unit for converting the frame data of the photographed image into monochrome data, A reference frame is set by comparing the amount of change between the frames included in the converted monochrome image, and a reference frame is set by using a first amount of change between the current frame and the previous frame and a second amount of change between the current frame and the reference frame, And a vehicle number recognizing unit for recognizing the vehicle number from the photographed image according to whether the shadow exists or not.

And the car number recognizing unit recognizes the car number by extracting a frame at a time when the shadow does not exist in the photographed image.

According to the method of recognizing the car number and the car number recognition system of the present invention, the reference frame is set for each photographed image, and the change amount between the current frame and the previous frame and the change amount between the current frame and the reference frame It is possible to accurately recognize the vehicle number by accurately detecting the presence or absence of a shadow whose size or position changes according to the sun position.

Also, since the RGB color image is converted into the black and white image and the image signal is processed, the image processing speed can be greatly reduced, so that the car number recognition can be performed quickly.

In addition, since it is not necessary to perform a separate signal processing process such as deleting a shadow area for a photographed image by extracting a frame at a point where a shadow does not exist according to the location of the vehicle and recognizing the car number, The cost of constructing can be reduced.

FIG. 1 is a flowchart illustrating a method of recognizing a car number according to an embodiment of the present invention.
FIGS. 2 to 8 are views for explaining a car number identification method according to an embodiment of the present invention, and FIGS.
9 is a block diagram showing a schematic configuration of a car number recognition system according to an embodiment of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS The above and other features and advantages of the present invention will become more apparent from the following detailed description of the present invention when taken in conjunction with the accompanying drawings, It will be possible. The present invention is capable of various modifications and various forms, and specific embodiments are illustrated in the drawings and described in detail in the text. It is to be understood, however, that the invention is not intended to be limited to the particular forms disclosed, but on the contrary, is intended to cover all modifications, equivalents, and alternatives falling within the spirit and scope of the invention.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprising" or "having ", and the like, are intended to specify the presence of stated features, integers, steps, operations, elements, parts, or combinations thereof, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, parts, or combinations thereof.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, a method of recognizing a car number and a car number recognition system (hereinafter referred to as a car number recognition method and a car number recognition system) by eliminating shadow recognition errors according to a preferred embodiment of the present invention will be described with reference to the accompanying drawings. Will be described in detail.

FIG. 1 is a flowchart for explaining a method of recognizing a car number according to an embodiment of the present invention, and FIGS. 2 to 8 are views for explaining a car number identification method according to an embodiment of the present invention.

Referring to FIG. 1, first, frame data of a photographed image is converted into monochrome data (S200). That is, the image photographed through the CCTV is an RGB color image, and converts an RGB color image into a monochrome image using Equation 1 defined by the National Television System Committee (NTSC) as a technical standard.

When the video signal is processed by the car number recognition system using the converted monochrome data, the processing speed is three times faster than that when the video signal is processed using the RGB color data.

That is, a commonly used true color is formed by adding 8 bits of each color of RGB, so 24 bits of data are used per pixel. On the other hand, in the case of black and white, when the gray level is divided into 16 levels, 4 bits of data are used per pixel. As a result, the number of bits used in the image signal processing can be greatly reduced by converting the RGB color data into monochrome data, thereby greatly improving the processing speed.

Then, a reference frame is set by comparing the amount of change in brightness between frames included in the monochrome image (hereinafter referred to as a 'change amount') (S210). That is, the car number recognition system compares the amount of change between frames included in the monochrome image converted into monochrome data, and sets the current frame as a reference frame when it is determined that the current frame is the same as the previous frame for a preset time.

When the vehicle does not travel, the screen illustrated in FIG. 2 is continuously input. Therefore, when comparing the amount of change between frames in the car number recognition system, it is determined that the previous frame and the current frame are the same frame. Therefore, in the car number recognition system, the current frame shown in Fig. 2 is set as the reference frame.

Then, a first brightness variation (hereinafter referred to as 'first variation amount') between the current frame and the previous frame and a second brightness variation amount between the current frame and the reference frame (hereinafter referred to as 'second variation amount') are compared (S220) . If the first variation amount is similar to the second variation amount (S230-Y), it is determined that there is no shadow area in the photographed image (S240). The car number recognition system operates to recognize the car number at this time (S250).

If it is determined in step S230 that the first variation is not similar to the second variation (S230-N), it is determined that a shadow area exists in the photographed image (S240). In this case, the car number recognition system does not perform the process of recognizing the car number with respect to the photographed image, but operates to recognize the car number when the first variation amount and the second variation amount are similar to each other.

For example, when only the shadow of the vehicle is photographed, the first variation amount between the current frame and the previous frame of the photographed image is analyzed to be smaller than the second variation amount between the current frame and the reference frame. The difference in the amount of change is shown in FIG. 4. The horizontal axis in the graphs described below represents time, and the vertical axis represents the amount of change between frames. In addition, yellow is a line showing a first variation amount between the current frame and the previous frame, and blue is a line showing a second variation amount between the current frame and the reference frame.

When only the shadow exists, as shown in the graph of FIG. 4, both the first variation amount and the second variation amount are low, and the first variation amount has a characteristic smaller overall than the second variation amount. That is, since the shadows are compared with the reference frame, only the difference in brightness of the portion in which the shadow is displayed occurs. Therefore, when the comparison is performed frame by frame, the amount of change between frames is not large and the amount of change between frames is small. .

On the other hand, if there is no shadow area in the current frame, the first variation amount between the current frame and the previous frame is distributed similarly to the second variation amount between the current frame and the reference frame. For example, if the vehicle and the shadow of the vehicle are photographed together but the shadow of the vehicle is hardly displayed due to the position of the sun, the first variation amount and the second variation amount are similarly distributed. This characteristic is shown in the graph of FIG.

Referring to the graph of FIG. 6, a yellow line representing the first variation amount and a blue line representing the second variation amount are closely distributed, and this characteristic is exhibited when there is no shadow region in the photographed image. Therefore, when the characteristic of the graph shown in FIG. 6 is detected as a result of analyzing the image signal in the car number recognition system, it is determined that there is no shadow in the photographed image.

7, the first variation amount between the current frame and the previous frame is smaller than the second variation amount between the current frame and the reference frame even when the shadow of the vehicle and the vehicle are displayed together on the screen. That is, as shown in the graph of FIG. 8, the first variation amount in the shadow region S and the vehicle region C is smaller than the second variation amount.

Here, if the shadow region S disappears and the time P at which only the vehicle region C is displayed is seen, it can be seen that the first variation amount and the second variation amount are closely distributed. That is, the shadow region S disappears from the image photographed from this point P and only the vehicle region C is displayed. Therefore, the car number recognition system recognizes the car number by judging that the shadow region does not exist in the photographed image at this point P (P).

9 is a block diagram showing a schematic configuration of a car number recognition system according to an embodiment of the present invention.

9, the car number identification system 120 according to the present invention receives an image photographed from a plurality of CCTVs 100-1, 100-2, ..., 100-n through the CCTV network 110 The car number is recognized. The car number recognition system 120 includes a network interface unit 121, a photographed image storage unit 122, a monochrome conversion unit 123, a variation comparison unit 124, a car number recognition unit 125, a video display unit 126 ), A map DB 127, and a central control unit 128.

The network interface unit 121 is connected to the CCTV network 110 and receives images transmitted from the plurality of CCTVs 100-1, 100-2, ..., 100-n. Here, the images photographed by the plurality of CCTVs (100-1, 100-2, ..., 100-n) are RGB color images.

The photographed image received through the network interface unit 121 is stored in the photographed image storage unit 122 as an RGB color image, and is deleted when a preset period of time elapses.

The monochrome conversion unit 123 converts the RGB color image stored in the photographed image storage unit 122 into monochrome. That is, the frame data of the photographed image is converted into monochrome data.

The change amount comparing unit 124 compares the amount of change between frames included in the monochrome image converted into monochrome data to generate a reference frame. That is, the variation comparing unit 124 compares the amount of change between the frames included in the monochrome image, and sets the current frame as the reference frame when it is determined that the current frame is the same as the previous frame for a predetermined time.

The variation comparing unit 124 detects the presence or absence of shadow from the captured image using the first variation between the current frame and the previous frame and the second variation between the current frame and the reference frame. That is, when the first variation amount is smaller than the second variation amount, the variation amount comparison unit 124 determines that a shadow region exists in the photographed image. If the first variation amount is similar to the second variation amount, It is determined that the area does not exist.

The car number recognizing unit 125 recognizes the car number from the photographed image according to the presence or absence of a shadow. That is, the vehicle number recognizing unit 125 recognizes the number of the vehicle included in the photographed image at the time when the shadow area does not exist. In addition, the car number recognizing unit 125 recognizes the car number from the photographed image by judging the time point at which the amount of change between the current frame and the reference frame starts to increase sharply.

The image display unit 126 displays images photographed through a plurality of CCTVs. When a car number is recognized, the car number and the car number are displayed together.

The map DB 127 stores the map of the area where the plurality of CCTVs 100-1, 100-2, ..., 100-n are installed and the location of the plurality of CCTVs 100-1, 100-2, ..., Is stored.

The central control unit 128 controls the operation of the plurality of CCTVs 100-1, 100-2, ..., 100-n through the network interface unit 121 and controls the operation of the plurality of CCTVs 100-1, 100-2, ..., and 100-n, and controls the monochrome conversion unit 123, the change amount comparison unit 124, and the vehicle number recognition unit 125 so as to recognize the vehicle number.

The vehicle number recognition system 120 as described above can accurately recognize the vehicle number by eliminating shadow recognition errors.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit and scope of the invention as defined in the appended claims. It can be understood that it is possible.

100-1, 100-2, ..., 100-n: a plurality of CCTV
110: CCTV network 120: car number recognition system
121: Network interface unit 122: Photographed image storage unit
123: monochrome conversion unit 124: change amount comparison unit
125: car number recognition unit 126: video display unit
127: Map DB

Claims (8)

Converting the frame data of the photographed image into monochrome data;
Setting a reference frame by comparing a brightness change amount between frames included in the monochrome image converted into the monochrome data;
Detecting the presence or absence of shadow from the photographed image using a first lightness change amount between a current frame and a previous frame and a second lightness change amount between the current frame and the reference frame; And
And recognizing the vehicle number from the photographed image according to whether the shadow exists or not,
The step of recognizing the vehicle number may include:
And recognizing the vehicle number by extracting a frame at a point in time when the shadow does not exist in the photographed image. The method according to claim 1,
Wherein the setting of the reference frame comprises:
Comparing the brightness variation between the frames included in the monochrome image and setting the current frame as the reference frame if it is determined that the current frame is the same as the previous frame for a preset time period, How to recognize the car number by removing the error. delete delete delete delete A car number recognition system for recognizing a car number included in a video captured by a plurality of CCTVs transmitted through a CCTV network,
A monochrome conversion unit for converting the frame data of the photographed image into monochrome data;
A reference brightness is set by comparing the amount of brightness change between the frames included in the monochrome image converted into the monochrome data to determine a first brightness change amount between the current frame and the previous frame and a second brightness change amount between the current frame and the reference frame A change amount comparing unit for determining whether a shadow exists from the photographed image; And
And a car number recognition unit for recognizing the car number from the photographed image according to whether the shadow exists or not,
Wherein the vehicle number recognizing unit comprises:
And recognizing the vehicle number by extracting a frame at a time point at which the shadow does not exist in the photographed image, wherein the vehicle number is recognized by removing shadow recognition errors. delete

Images