JP2024518582A - License plate image processing device for recognizing vehicle registration numbers and license plate image correction method using the same - Google Patents

Abstract

A license plate image processing device for recognizing a vehicle registration number includes a vehicle image acquisition unit that captures an image of a vehicle and includes the vehicle, a license plate detection unit that detects an image of the license plate of the vehicle from the captured image and generates position data for the position of the image of the license plate in the captured image, a database unit that stores image correction data calculated in advance for each position of the image of the license plate in the captured image, a correction value generation unit that loads the image correction data corresponding to the position data from the database unit and generates an image correction value for the license plate based on the loaded image correction data, and a license plate image correction unit that corrects the image of the license plate based on the image correction value for the license plate. [Selected Figure] Figure 2

Description

The present invention relates to a license plate image processing device for recognizing vehicle registration numbers and a license plate image correction method using the same, and more specifically to a license plate image processing device for recognizing vehicle registration numbers that can correct an image of a license plate captured at an angle to its original shape with a relatively small amount of calculations, and a license plate image correction method using the same.

As image recognition technology develops, it has become possible to automatically recognize vehicle registration numbers using images of vehicle license plates, and currently the majority of vehicle access control systems use this method.

The image of the license plate is obtained by capturing an image of the vehicle using an imaging device installed at the entrance to a location such as a parking lot where vehicle entry and exit control is required. Recognition of the vehicle registration number is implemented in software and performed in a computer system, and is composed of a combination of a number plate identification step from the captured image, a number plate verification step, a code (letters, numbers, symbols) separation step, and a code recognition step.

Figure 1 shows a conventional license plate identification process for recognizing vehicle registration numbers, where the license plate is identified by detecting the outline of the license plate.

First, a camera installed at the entrance or exit of the parking lot captures an image of an entering or leaving vehicle (step S11). Then, the contour of the license plate image contained in the captured image of the vehicle is extracted to finally obtain an image of the license plate. (step S12)

On the other hand, if the imaging device is installed diagonally from the direction in which the vehicle enters, or if the vehicle enters the entrance/exit at an angle, the vehicle's license plate can be imaged at an angle.

As shown in Figure 1, the aspect ratio of a license plate captured at an angle may differ from the standard ratio of a typical license plate, and the license plate has a shape that is rotated at a predetermined angle compared to an image of the license plate captured from the front.

In order to accurately recognize the vehicle registration number, images of such slanted license plates are corrected to the original license plate shape, but to perform accurate correction, the license plate outline must be accurately extracted.

In this case, the license plate contour is extracted using the Generalized Hough Transform, but this method requires a large amount of calculations because it relies on pixels.

However, vehicle registration number recognition devices installed at entrances and exits have limited processing power due to installation locations and equipment specifications, making it difficult to perform tasks that involve a large amount of calculations. In particular, if the quality of the vehicle image acquired is poor due to issues such as focus, the number of pixels that make up the boundary of the license plate in the image will decrease, making it even more difficult to accurately extract the outline of the license plate.

In order to solve the above problems, the technical problem that this invention aims to solve is to provide a license plate image processing device for recognizing vehicle registration numbers, which can correct an image of a license plate captured at an angle to its original shape with a relatively small amount of calculation, and a license plate image correction method using the same.

The technical problems that the present invention aims to achieve are not limited to those mentioned above, and other technical problems not mentioned will be clearly understood by those with ordinary skill in the technical field to which the present invention pertains from the description below.

In order to achieve the above technical objectives, one embodiment of the present invention provides a license plate image processing device for recognizing vehicle registration numbers, including a vehicle image acquisition unit that captures an image of a vehicle and captures an image including the vehicle, a license plate detection unit that detects an image of the license plate of the vehicle from the captured image and generates position data for the position of the license plate image in the captured image, a database unit that stores image correction data calculated in advance for each position of the license plate image in the captured image, a correction value generation unit that loads the image correction data corresponding to the position data from the database unit and generates an image correction value for the license plate based on the loaded image correction data, and a license plate image correction unit that corrects the image of the license plate based on the image correction value for the license plate.

The captured image is divided into a plurality of areas, the image correction data is calculated in advance for each area and stored in the database unit, and the correction value generation unit can load the image correction data corresponding to the area to which the position data belongs.

The image correction data may include a first correction value that corrects the scale of the license plate image and a second correction value that rotates and corrects the license plate image.

The database unit further stores a reference value of the aspect ratio calculated in advance for each position of the license plate image in the captured image, and the correction value generation unit can correct at least one of the first correction value and the second correction value based on the difference between the reference value of the aspect ratio corresponding to the position data and the aspect ratio of the license plate image.

The vehicle image acquisition unit captures an image of the vehicle at a predetermined imaging position, the correction value generation unit corrects the position data based on the difference in height between the ground at the position where the vehicle image acquisition unit is installed and the ground at the predetermined imaging position, and the correction value generation unit can load the image correction data corresponding to the corrected position data.

The correction value generation unit can further correct the position data using ground inclination angle data at the predetermined imaging position.

The position data consists of two-dimensional coordinate data including a horizontal axis value and a vertical axis value, and the correction value generation unit can correct the vertical axis value of the position data based on the height difference.

The second correction value can be corrected based on the angle between the imaging direction of the vehicle image acquisition unit and the direction of the approach road in which the vehicle was located at the time of imaging.

The license plate image correction unit can perform rotation correction of the license plate image based on the second correction value, and then correct the scale of the license plate image based on the first correction value.

In order to achieve the above technical objective, one embodiment of the present invention provides a license plate image correction method for recognizing vehicle registration numbers, comprising the steps of: a) capturing an image of a vehicle to obtain an image including the vehicle, and detecting an image of a license plate from the captured image; b) generating position data for the position of the image of the license plate in the captured image; c) loading image correction data corresponding to the position data from a database unit in which image correction data calculated in advance for each position of the image of the license plate in the captured image is stored, and generating an image correction value for the license plate based on the loaded image correction data; and d) correcting the image of the license plate based on the image correction value for the license plate.

The captured image is divided into a plurality of areas, and the image correction data is calculated in advance for each area and stored in the database unit, and in step c), the image correction data corresponding to the area to which the position data belongs can be loaded from the database unit.

The database unit further stores a reference value of the aspect ratio calculated in advance for each position of the license plate image in the captured image, and the c) step can include a step of loading image correction data corresponding to the position data from the database unit, and a step of correcting the image correction data based on the difference between the reference value of the aspect ratio corresponding to the position data and the aspect ratio of the license plate image.

The step c) may include a step of loading image correction data corresponding to the position data from a database unit, and a step of correcting the image correction data based on the angle between the imaging direction of the vehicle image acquisition unit and the direction of the approach road on which the vehicle is located.

The captured image is acquired using a vehicle image acquisition unit that captures an image of the vehicle, and between steps b) and c), the method may further include a step of correcting the position data based on the difference in height between the ground at the position where the vehicle image acquisition unit is installed and the ground at a predetermined imaging position.

The position data can be further corrected based on ground inclination angle data at a predetermined imaging position.

The image correction data includes a first correction value for correcting the scale of the license plate image and a second correction value for rotating and correcting the license plate image, and in step d), after rotating and correcting the license plate image based on the second correction value, the scale of the license plate image can be corrected based on the first correction value.

According to an embodiment of the present invention, the license plate image processing device can correct the license plate image using image correction data that is pre-calculated for each position of the license plate image in the captured image. That is, the license plate image processing device can estimate the shape of the license plate image without the process of extracting a contour line using complex calculations such as the Generalized Hough Transform. Therefore, the license plate image processing device can correct the image of the license plate captured at an angle to the original shape of the license plate with a relatively small amount of calculations.

Furthermore, according to an embodiment of the present invention, the license plate image processing device can store a reference value of the aspect ratio for each area of the captured image. The license plate image processing device then corrects the image correction data based on the difference between the reference value of the aspect ratio of the area in which the captured license plate image is located and the aspect ratio of the captured license plate image, so that the captured license plate image can be corrected to more closely resemble the standard shape of the license plate.

Furthermore, according to an embodiment of the present invention, the license plate image processing device corrects the image correction data based on the angle between the imaging direction of the vehicle image acquisition unit and the direction of the approach road where the vehicle being imaged is located, so that the license plate image can be corrected accurately.

Furthermore, according to an embodiment of the present invention, the license plate image processing device corrects the position data based on the difference in height between the ground at the position where the vehicle image acquisition unit is installed and the ground at the predetermined imaging position, so that the license plate image can be corrected accurately.

In addition, according to an embodiment of the present invention, the license plate image processing device further corrects the position data based on the ground inclination angle data at the predetermined imaging position, so that the license plate image can be corrected more accurately.

The effects of the present invention are not limited to those described above, but include all effects that can be inferred from the configuration of the invention described in the description of the present invention or the claims.

FIG. 1 illustrates a conventional license plate identification process for recognizing vehicle registration numbers. 1 is a block diagram showing an outline of a license plate image processing device according to an embodiment of the present invention; FIG. 13 is a diagram for explaining an example of an image of a license plate captured at an angle. FIG. 2 is a detailed block diagram showing a license plate image processing device according to an embodiment of the present invention. 5A and 5B are diagrams illustrating an example of a captured image acquired by a vehicle image acquisition unit according to the embodiment of the present invention. 1A to 1C are diagrams showing the shape of license plate images in each area in a captured image according to an embodiment of the present invention in a stylized form. 11 is a diagram illustrating an image correction data table that is calculated and stored in advance for each area of a captured image according to an embodiment of the present invention. FIG. 11A and 11B are diagrams for explaining an example in which an aspect ratio correction unit according to an embodiment of the present invention corrects image correction data; 11A and 11B are diagrams for explaining an example in which an imaging direction correction unit according to an embodiment of the present invention corrects image correction data. 11A and 11B are diagrams for explaining an example in which a ground height correction unit according to an embodiment of the present invention corrects position data; 1 is a flowchart showing a license plate image correcting method using the license plate image processing device according to an embodiment of the present invention. FIG. 2 is a diagram showing detailed steps of step S220 according to an embodiment of the present invention.

The present invention will now be described with reference to the accompanying drawings. However, the present invention can be embodied in various different forms and is not limited to the embodiments described herein. In order to clearly explain the present invention, parts that are not related to the description have been omitted in the drawings, and similar parts throughout the specification have been given the same reference numerals.

Throughout the specification, when a part is said to be "connected (connected, in contact, bonded)" to another part, this includes not only the case where it is "directly connected" to another part, but also the case where it is "indirectly connected" through an intervening member. In addition, when a part is said to "include" a certain component, this does not exclude other components, and means that it may further include other components, unless otherwise specified.

The terms used in this specification are merely used to describe certain embodiments and are not intended to limit the present invention. The singular expressions include the plural expressions unless the context clearly indicates otherwise. In this specification, the terms "include" and "comprise" mean the presence of features, numbers, steps, operations, components, parts, or combinations thereof described in the specification, and do not preclude the presence or additional possibility of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

The following describes in detail an embodiment of the present invention with reference to the attached drawings.

Figure 2 is a block diagram showing a license plate image processing device 100 according to an embodiment of the present invention. And Figure 3 is a diagram for explaining an example of an image of a license plate captured at an angle.

As shown in FIG. 2, the license plate image processing device 100 can include a vehicle image acquisition unit 110, a license plate detection unit 120, a correction value generation unit 130, a license plate image correction unit 140, and a database unit D.

The vehicle image acquisition unit 110 can capture an image of the vehicle 1 and acquire an image including the vehicle 1. For example, the vehicle image acquisition unit 110 may be a camera installed at the entrance of a location where vehicle entry and exit management is required.

The license plate detection unit 120 can detect an image of the license plate of the vehicle 1 from the captured image. In this case, the license plate can be captured obliquely as shown in FIG. 3a. In order to accurately recognize the vehicle registration number from the license plate image, the license plate image Pa captured obliquely can be corrected to a license plate image Pb having the original shape as shown in FIG. 3b. To this end, the license plate detection unit 120 can generate position data for the position of the license plate image in the captured image. Specific examples of the use of the position data will be described in detail later with reference to FIGS. 5 to 7.

The database unit D can store image correction data that is calculated in advance for each position of the license plate image in the captured image. The image correction data can be used to correct the license plate image.

The correction value generating unit 130 can load image correction data corresponding to the position data from the database unit D. That is, the correction value generating unit 130 can load different image correction data according to the position of the license plate image in the captured image. Then, the correction value generating unit 130 can generate an image correction value for the license plate based on the loaded image correction data.

The license plate image correction unit 140 can then use the license plate image correction value to correct the license plate image.

In FIG. 2, an example is shown in which the components of the license plate image processing device 100 are integrated and installed at the entrance/exit of a location where vehicle entry and exit management is required. However, the components of the license plate image processing device 100 may be installed in different locations. For example, the vehicle image acquisition unit 110 is installed at the entrance/exit of the vehicle 1, and the other components are provided in a different location or can be mounted on an external server and communicate with the vehicle image acquisition unit 110.

Figure 4 is a detailed block diagram showing the license plate image processing device 100 according to an embodiment of the present invention.

As shown in FIG. 4, the license plate detection unit 120 can receive a captured image from the vehicle image acquisition unit 110. The captured image may be an image that includes a vehicle.

The license plate detection unit 120 may include a license plate image detection unit 121 and a license plate position determination unit 122.

The license plate image detection unit 121 can detect an image of a license plate in a captured image. In other words, the license plate image detection unit 121 can generate an image of a license plate.

The license plate position determination unit 122 can generate position data indicating the position of the license plate image within the captured image. The position data may be two-dimensional coordinate data with a specific position in the captured image as the reference point. The position data may have a horizontal axis value and a vertical axis value. The position data may be polar coordinates. The position data may be the coordinates of the center point of the license plate image. For example, the center point may be the center of gravity of the license plate image.

On the other hand, the location data may be data indicating a specific area within the captured image.

Specifically, the captured image can be divided into multiple areas. The license plate position determination unit 122 can then determine which area the position of the license plate image (e.g., the position of the center point of the license plate image) belongs to. The license plate position determination unit 122 can then generate information about the area as position data.

The license plate detection unit 120 can then transmit the license plate image and position data to the correction value generation unit 130.

The correction value generation unit 130 can generate the license plate image correction value required to correct the license plate image.

The correction value generation unit 130 can include an image correction data loading unit 131, an aspect ratio correction unit 132, an imaging direction correction unit 133, and a ground height correction unit 134.

The image correction data loading unit 131 can load image correction data corresponding to the position data from the database unit D.

At this time, the database unit D may store image correction data that has been calculated in advance for each position of the license plate image in the captured image. For example, the image correction data may include a first correction value for scale correction of the license plate image and a second correction value for rotation correction of the license plate image. The image correction data may also be data that has been calculated in advance for each area of the captured image.

Then, the correction value generation unit 130 can select the loaded image correction data as the image correction value for the license plate.

On the other hand, the correction value generation unit 130 can correct the loaded image correction data using the aspect ratio correction unit 132, the imaging direction correction unit 133, and the ground height correction unit 134 in order to correct the license plate image more accurately.

This is because the position or orientation of the vehicle when the image correction data is calculated may differ slightly from the position or orientation of the vehicle when the image was captured.

The aspect ratio correction unit 132 can correct the loaded image correction data based on the aspect ratio of the license plate image obtained by imaging. At this time, the database unit D may be in a state where a reference value of the aspect ratio is stored. The reference value of the aspect ratio may be the aspect ratio of the license plate image calculated in advance for each position of the license plate image in the captured image.

Specifically, the aspect ratio correction unit 132 can correct the image correction data based on the difference between the aspect ratio reference value corresponding to the position data and the aspect ratio of the license plate image. A specific example of the aspect ratio correction will be described in detail later with reference to FIG. 8.

The imaging direction correction unit 133 can correct the image correction data based on the angle between the imaging direction of the vehicle image acquisition unit 110 and the direction of the approach road in which the vehicle to be imaged is located.

This is because the vehicle image acquisition unit 110 may be installed at an angle to the vehicle's approach road. For example, the imaging direction correction unit 133 can correct the second correction value of the image correction data by adding or subtracting the angle between the imaging direction of the vehicle image acquisition unit 110 and the direction of the approach road on which the vehicle to be imaged is located. A specific example of imaging direction correction will be described in detail later with reference to FIG. 9.

The ground height correction unit 134 can correct the position data using the difference in height between the ground at the position where the vehicle image acquisition unit 110 is installed and the ground at the imaging position.

This is because if the ground level at the location where the vehicle image acquisition unit 110 is installed is different from the ground level at the preset imaging position, the imaging angle of the license plate may differ even if the position of the center point of the license plate is the same. A specific example of how the ground level is corrected will be described in detail later with reference to FIG. 10.

Then, the correction value generation unit 130 can select the final corrected image correction data as the image correction value for the license plate.

The license plate image correction unit 140 can correct the license plate image using the license plate image correction value.

The license plate image correction unit 140 can include a scale correction unit 141 and a rotation correction unit 142.

The scale correction unit 141 can correct the scale of the license plate image using the first correction value. For example, the scale correction unit 141 can expand or contract the license plate image in the left-right or up-down direction. And the rotation correction unit 142 can rotate and correct the license plate image using the second correction value.

Figure 5 is a diagram showing an example of a captured image acquired by the vehicle image acquisition unit 110 according to an embodiment of the present invention.

Figure 5 shows a single captured image including first to third vehicles C1, C2, and C3 to illustrate an example of captured images captured in different environments (positions).

As shown in FIG. 5, the position and shape of the license plate of an imaged vehicle may vary depending on the relative positions of the vehicle and the image acquisition unit 110. That is, the position and shape of the license plate may vary depending on the distance between the vehicle and the vehicle image acquisition unit 110 and the imaging direction.

For example, if the ground at the imaging position is located to the northwest compared to the ground at the installation position of the vehicle image acquisition unit 110, an image P1 of the license plate of the first vehicle C1 is obtained.

For example, if the ground at the installation position of the vehicle image acquisition unit 110 is located directly in front of the ground at the imaging position, an image P2 of the license plate of the second vehicle C2 is obtained.

For example, if the ground at the imaging position is located southeast of the ground at the installation position of the vehicle image acquisition unit 110, an image P3 of the license plate of the third vehicle C3 is obtained.

Figure 6 is a diagram showing the standardized shape of the license plate image in each area of the captured image according to an embodiment of the present invention. And Figure 7 is a diagram for explaining an image correction data table that is pre-calculated for each area of the captured image according to an embodiment of the present invention.

The embodiments of Figures 6 and 7 assume that the captured image is divided into 15 areas of 5 columns and 3 rows, and that image correction data is assigned to each area.

The license plate images shown in each zone in the captured image in Figure 6 show the estimated shape of the license plate image when a license plate is captured in that zone. Because the vehicle entrance/exit road viewed by the license plate image processing device 100 is fixed, it is possible to estimate the shape of the license plate image captured in each zone.

For example, the license plate of the first vehicle C1 in FIG. 5 is located in the first area A1, so it can be assumed that the license plate has the shape S1 of the image of the first license plate. Then, the image correction data loading unit 131 can load the image correction data corresponding to the first area A1 from the database unit D. Referring to FIG. 7, the first area A1 stores the first correction value ALP_RATIO1 and the second correction value ALP_ANGLE1 as image correction data.

At this time, the first correction value ALP_RATIO1 and the second correction value ALP_ANGLE1 in the first area A1 can be correction values for converting the shape S1 of the first license plate image into the shape of the license plate image that has already been set. Here, the shape of the license plate image that has already been set can be the shape of the license plate when it is imaged from the front, that is, the standard shape of the license plate.

By the same principle, since the license plate of the second vehicle C2 is located in the second area A2, it can be estimated that the license plate has the shape S2 of the image of the second license plate. Then, the image correction data loading unit 131 can load the first correction value ALP_RATIO8 and the second correction value ALT_ANGLE8 of the image correction data corresponding to the second area A2 from the database unit D in order to correct the image of the license plate of the second vehicle C2.

In addition, since the license plate of the third vehicle C3 is located in the third area A3, it can be estimated that the license plate has the shape S3 of the image of the third license plate. The image correction data loading unit 131 can load the first correction value ALP_RATIO15 and the second correction value ALT_ANGLE15 of the image correction data corresponding to the third area A3 from the database unit D in order to correct the image of the license plate of the third vehicle C3.

Accordingly, the license plate image processing device 100 can estimate the shape of the license plate image without the process of extracting the contour line using complex calculations such as the Generalized Hough Transform. In other words, the license plate image processing device 100 can correct the image of the license plate captured at an angle to its original shape with a small amount of calculations.

Figure 8 is a diagram for explaining an example of how the aspect ratio correction unit 132 according to an embodiment of the present invention corrects image correction data. The license plate image in Figure 8 is an image of the license plate of a fourth vehicle (not shown) captured in a fourth area A4. In this case, the image of the license plate of the fourth vehicle may have a slightly different aspect ratio than the shape S4 of the image of the fourth license plate in the fourth area A4. This is because, as described above, the position or direction of the vehicle when the image correction data is calculated may be slightly different from the position or direction of the vehicle when it is captured.

For this correction, the database unit D can further store a reference value of the aspect ratio that is pre-calculated for each position of the license plate image in the captured image. Of course, such a reference value of the aspect ratio may be pre-calculated for each area.

Then, the aspect ratio correction unit 132 can correct the image correction data based on the difference between the reference value of the aspect ratio corresponding to the position data and the aspect ratio of the license plate image (e.g., P4w/P4h). Specifically, the aspect ratio correction unit 132 can correct at least one of the first correction value and the second correction value included in the image correction data.

By doing this, the image correction data is corrected based on the aspect ratio of the captured license plate image, so that the captured license plate image can be corrected to more closely resemble the standard license plate shape.

Figure 9 is a diagram illustrating an example in which the imaging direction correction unit 133 according to an embodiment of the present invention corrects image correction data.

As shown in FIG. 9, when the imaging direction W1 of the image acquisition unit 110 and the approach road direction W2 of the vehicle are oblique, the imaging direction correction unit 133 can correct the image correction data based on the angle T between the imaging direction of the vehicle image acquisition unit 110 and the direction of the approach road in which the vehicle was located at the time of imaging. At this time, the corrected image correction data can be a second correction value.

For example, the imaging direction correction unit 133 can correct the second correction value of the image correction data by adding or subtracting the angle between the imaging direction W1 of the vehicle image acquisition unit 110 and the direction W2 of the approach road on which the vehicle to be imaged is located. In this case, the direction W2 of the approach road on which the vehicle is located may of course be replaced with the forward direction of the vehicle 1.

For example, the angle T between the imaging direction of the vehicle image acquisition unit 110 and the direction of the approach road on which the vehicle 1 to be imaged is located may be a value measured when the license plate image processing unit 100 is installed and input to the imaging direction correction unit 133.

As a result, even if the vehicle image acquisition unit 110 is installed so as to capture an image of the license plate at an angle, the captured image of the license plate can be corrected to resemble the standard shape of the license plate.

Figure 10 is a diagram illustrating an example of how the ground height correction unit 134 corrects position data according to an embodiment of the present invention.

As shown in FIG. 10, the ground height at the position where the vehicle image acquisition unit 110 is installed may differ from the ground height at the imaging position. In this case, the imaging position means the position where the image of the license plate is captured. Such an imaging position can be determined in advance based on the focus and imaging performance of the vehicle image acquisition unit 110.

Figure 10a shows a case where the ground at the imaging position is higher than the ground at the position where the vehicle image acquisition unit 110 is installed. In Figure 10a, the imaging position is R1, and the difference in height between the ground at the position where the vehicle image acquisition unit 110 is installed and the ground at the R1 position is H1.

Then, the ground height correction unit 134 can correct the position data based on the H1 value. For example, if the position data is two-dimensional coordinate data having a horizontal axis value and a vertical axis value, the ground height correction unit 134 can correct the vertical axis value of the position data.

Note that FIG. 10b shows a case where the ground height at the imaging position R2 is equal to the height of the position where the vehicle image acquisition unit 110 is installed. In this case, the ground height correction unit 134 can omit correcting the position data.

Figure 10c shows a case where the ground at the imaging position is lower than the ground at the position where the vehicle image acquisition unit 110 is installed. In Figure 10c, the imaging position is R3, and the difference in height between the ground at the position where the image acquisition unit 110 is installed and the ground at the R3 position is H3.

Then, the ground height correction unit 134 can correct the position data based on the H3 value using the same principle as in Figure 10a.

On the other hand, the ground height correction unit 134 can further correct the position data using the ground inclination angle data at the imaging position. This is because even if the ground height in front of the vehicle where the license plate is located is the same as the ground height at the position where the image acquisition unit 110 is installed, the ground height at the rear of the vehicle may be different.

Figure 11 is a flowchart showing a license plate image correction method using the license plate image processing device 100 according to an embodiment of the present invention.

First, the license plate image processing device 100 captures an image of a vehicle to obtain an image including the vehicle, and can detect an image of the license plate from the captured image (S200).

Then, the license plate image processing device 100 can generate position data indicating the position of the license plate image within the captured image (S210).

Furthermore, the license plate image processing device 100 can load image correction data corresponding to the position data from the database unit D and generate an image correction value for the license plate based on the loaded image correction data (S220).

The license plate image processing device 100 can also correct the license plate image using the license plate image correction value (S230).

In step S230, the image of the license plate can be rotationally corrected based on the second correction value, and then scaled based on the first correction value.

Figure 12 shows detailed steps of step S220 according to an embodiment of the present invention.

As shown in FIG. 12, step S220 may include a step of correcting the position data based on the difference in height between the ground at the position where the image acquisition unit 110 is installed and the ground at a predetermined imaging position (S221).

For example, if the position data is two-dimensional coordinate data having horizontal and vertical axis values, the position data can be corrected using the following Equation 1.

In formula 1, Center X % width may be a value indicating the horizontal position of the center point of the license plate image in the captured image. Center Y % height may be a value indicating the vertical position of the center point of the license plate image in the captured image.

For example, if the captured image is partitioned into 5 columns and 3 rows as shown in FIG. 5, and the horizontal axis position of the center point of the license plate image is within 0-20% and the vertical axis position is within 66.6-100%, ALP_Ratio[Center X% width][Center Y% height] may be the first correction value of the image correction data corresponding to the first area A1.

In this case, if the Entrance Height Diff is calculated as a negative value due to the difference in height between the ground at the position where the image acquisition unit 110 is installed and the ground at the predetermined imaging position, ALP_Ratio[Center X% width][(Center Y + Entrance Height Diff)% height] can be the first correction value of the image correction data corresponding to the area located below the first area A1.

The Alpha value may be a correction ratio coefficient that is preset depending on the installation location of the license plate image processing device 100. W_H_Ratio may be the aspect ratio of the license plate image extracted from the captured image. If W_H_Ratio is calculated as a negative value, the Alpha value may be set to a negative value.

On the other hand, the predetermined imaging position may be a position that is predetermined based on the focus and imaging performance of the vehicle image acquisition unit 110. Also, if the Entrance Height Diff is 0, that is, if there is no difference in height between the ground surfaces, step S221 may be omitted.

Next, a step can be performed in which image correction data corresponding to the position data is loaded from database unit D (S222).

Then, a step can be performed of correcting the image correction data based on the difference between the reference value of the aspect ratio corresponding to the position data and the aspect ratio of the license plate image (S223).

For example, the image correction data may include a first correction value for a scale correction of the license plate image and a second correction value for a rotation correction of the license plate image. Then, at least one of the first correction value and the second correction value may be corrected in step S223.

For example, the image correction data can be corrected such that the first correction value is corrected by the following formula 2, and the second correction value is corrected by the following formula 3.

In formula 2, Center X% width may be a value indicating the horizontal position of the center point of the license plate image in the captured image. Center Y% height may be a value indicating the vertical position of the center point of the license plate image in the captured image.

For example, if the captured image is partitioned into 5 columns and 3 rows as shown in FIG. 5, and the horizontal axis position of the center point of the license plate image is within 0-20% and the vertical axis position is within 66.6-100%, ALP_Ratio[Center X% width][Center Y% height] may be the first correction value of the image correction data corresponding to the first area A1. The Alpha value may be a correction ratio coefficient that is preset depending on the installation location of the license plate image processing device 100. W_H_Ratio may be the aspect ratio of the license plate image extracted from the captured image. If W_H_Ratio is calculated as a negative value, the Alpha value can be set to a negative value.

In formula 3, Center X% width may be a value indicating the horizontal position of the center point of the license plate image in the captured image. Center Y% height may be a value indicating the vertical position of the center point of the license plate image in the captured image.

For example, if the captured image is partitioned into 5 columns and 3 rows as shown in FIG. 5, and the horizontal axis position of the center point of the license plate image is within 0-20% and the vertical axis position is within 66.6-100%, then ALP_Angle[Center X% width][Center Y% height] may be the second correction value of the image correction data corresponding to the first area A1. And, the Beta value may be a correction ratio coefficient that is preset depending on the installation location of the license plate image processing device 100. W_H_Ratio may be the aspect ratio of the license plate image extracted from the captured image. If W_H_Ratio is calculated as a negative value, the Beta value can be set to a negative value.

Then, a step can be performed to correct the image correction data based on the angle between the imaging direction of the vehicle image acquisition unit 110 and the direction of the approach road on which the vehicle being imaged is located (S224).

For example, the image correction data can be corrected to a second correction value using the following equation 4:

Equation 4 differs from equation 3 in that the second correction value is corrected using LP_CAR_Photo_Angle, which indicates the angle between the imaging direction of the vehicle image acquisition unit 110 and the direction of the approach road on which the vehicle being imaged is located.

The above description of the present invention is for illustrative purposes only, and a person having ordinary skill in the art will understand that the present invention can be easily modified into other specific forms without changing the technical concept or essential features of the present invention. Therefore, the above-described embodiments should be understood to be illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and similarly, each component described as distributed may be implemented in a combined form.

The scope of the present invention is defined by the claims set forth below, and all modifications and variations that fall within the meaning and scope of the claims, as well as equivalent concepts, are included in the scope of the present invention.

The license plate image processing device for recognizing vehicle registration numbers and the license plate image correction method using the same according to an embodiment of the present invention can store a reference value of the aspect ratio for each area of a captured image, and correct the image correction data based on the difference between the reference value of the aspect ratio of the area in which the captured license plate image is located and the aspect ratio of the captured license plate image, so that the license plate image can be corrected to its original shape with a relatively small amount of calculations, and can be performed under various conditions and industrial sites.

Claims (14)

a vehicle image acquisition unit that captures an image of a vehicle and acquires an image including the vehicle;
a license plate detection unit that detects an image of a license plate of the vehicle from the captured image and generates position data for a position of the image of the license plate in the captured image;
A database unit in which image correction data calculated in advance for each position of the image of the license plate in the captured image is stored;
a correction value generating unit that loads the image correction data corresponding to the position data from the database unit and generates an image correction value of the license plate based on the loaded image correction data;
a license plate image correction unit that corrects the license plate image based on the license plate image correction value;
Including,
The captured image is divided into a plurality of areas,
The image correction data is pre-calculated for each region and stored in the database unit,
The number plate image processing device for recognizing a vehicle registration number, wherein the correction value generating unit loads the image correction data corresponding to an area to which the position data belongs. The license plate image processing device for recognizing vehicle registration numbers according to claim 1, wherein the image correction data includes a first correction value for correcting a scale of the license plate image and a second correction value for rotationally correcting the license plate image. The database unit further stores a reference value of an aspect ratio calculated in advance for each position of the image of the license plate in the captured image,
3. The license plate image processing device for recognizing vehicle registration numbers according to claim 2, wherein the correction value generating unit corrects at least one of the first correction value and the second correction value based on a difference between the reference value of the aspect ratio corresponding to the position data and the aspect ratio of the image of the license plate. The vehicle image acquisition unit captures an image of the vehicle at a predetermined imaging position,
the correction value generation unit corrects the position data based on a difference in height between a ground surface at a position where the vehicle image acquisition unit is installed and a ground surface at the predetermined imaging position;
The number plate image processing device for vehicle registration number recognition according to claim 2 , wherein the correction value generating unit loads the image correction data corresponding to the corrected position data. The license plate image processing device according to claim 4 , wherein the correction value generating unit further corrects the position data by using inclination angle data of the ground at the predetermined imaging position. the position data is two-dimensional coordinate data including a horizontal axis value and a vertical axis value;
The number plate image processing device for recognizing an automobile registration number according to claim 4 , wherein the correction value generating unit corrects the vertical axis value of the position data based on the difference in height. The number plate image processing device for recognizing vehicle registration numbers according to claim 2, wherein the second correction value is corrected based on an angle between an imaging direction of the vehicle image acquisition unit and a direction of an access road in which the vehicle was located at the time of imaging. 3. The license plate image processing device for recognizing vehicle registration numbers according to claim 2, wherein the license plate image correction unit corrects the scale of the license plate image based on the first correction value after rotating the license plate image based on the second correction value. a) capturing an image of a vehicle to obtain an image including the vehicle, and detecting an image of a license plate from the captured image;
b) generating position data for a position of the image of the license plate within the captured image;
c) loading image correction data corresponding to the position data from a database unit in which image correction data calculated in advance for each position of the image of the license plate in the captured image is stored, and generating an image correction value of the license plate based on the loaded image correction data;
d) correcting the license plate image based on the license plate image correction value;
Including,
The captured image is divided into a plurality of areas,
The image correction data is pre-calculated for each region and stored in the database unit,
In the step c),
and loading the image correction data corresponding to the area to which the location data belongs from the database unit. The database unit further stores a reference value of an aspect ratio calculated in advance for each position of the image of the license plate in the captured image,
The step c) comprises:
loading image correction data corresponding to the position data from a database unit;
correcting the image correction data based on a difference between a reference value of the aspect ratio corresponding to the position data and an aspect ratio of the image of the license plate;
The method for correcting a license plate image for recognition of a vehicle registration number according to claim 9, comprising: The captured image is acquired using a vehicle image acquisition unit that captures an image of the vehicle,
The step c) comprises:
loading image correction data corresponding to the position data from a database unit;
correcting the image correction data based on an angle between an imaging direction of the vehicle image acquisition unit and a direction of an approach road on which the vehicle is located;
The method for correcting a license plate image for recognition of a vehicle registration number according to claim 9, comprising: The captured image is acquired using a vehicle image acquisition unit that captures an image of the vehicle,
The license plate image correction method for recognizing vehicle registration numbers according to claim 9, further comprising, between the steps b) and c), a step of correcting position data based on a difference in height between the ground at a position where the vehicle image acquisition unit is installed and the ground at a predetermined imaging position. The method of claim 12, wherein the position data is further corrected based on ground inclination angle data at the predetermined imaging position. The image correction data includes a first correction value for correcting a scale of the image of the license plate and a second correction value for rotationally correcting the image of the license plate,
In the step d),
The method for correcting a license plate image for recognizing a vehicle registration number according to claim 9, further comprising correcting a scale of the license plate image based on the first correction value after the rotation of the license plate image is corrected based on the second correction value.

Images