KR20220030110A - Method for updating reference vehicle identification number using for idntifying vehicle identification number - Google Patents

Abstract

The present invention relates to a method of updating a reference vehicle identification number image used to recognize a vehicle identification number carved in a vehicle. More specifically, since a reference vehicle identification number image used to recognize a vehicle identification number of a vehicle is updated to an image of characters carved in the vehicle on an actual site, the image can be updated to an optimized reference vehicle identification number image considering an environment of the site in which the vehicle is manufactured and the vehicle identification number is carved therein, and also, the reference vehicle identification number image is divided into a registered vehicle identification number image and a candidate vehicle identification number image and a determination as to whether to update the registered vehicle identification number image to the candidate vehicle identification number image is made, and, as a result, the reference vehicle identification number image can be updated to an optimized image among character images carved in the vehicle on the actual site.

Description

{Method for updating reference vehicle identification number using for idntifying vehicle identification number}

The present invention relates to a method of updating a reference vehicle identification number image used to recognize a vehicle identification number engraved on a vehicle, and more specifically, to a reference vehicle identification number image used for vehicle identification number recognition as a character engraved on a vehicle at the actual site By updating the image of By comparing the vehicle identification number image and the reference vehicle identification number image, it is decided whether to update the image with high correlation among the candidate vehicle identification number image and the reference vehicle identification number image. It relates to a method of updating a reference vehicle identification number image capable of updating

Vehicle Identification Number (VIN) is attached or engraved on the dash panel, cross member, or upper part of the firewall. Used for tracking, etc.

The vehicle identification number was devised by ISO (International Organization for Standardization) in 1980 and has been applied to vehicles since 1981. , vehicle model, year of manufacture, etc.

When the vehicle manufacturer manufactures a vehicle, the vehicle identification number is engraved on the vehicle using an engraving device, and it is necessary to check whether the vehicle identification number engraved on the vehicle is correctly formed using the recognition device.

In general, a recognition device recognizes a vehicle identification number attached or engraved on a vehicle using lighting and a camera.

The recognition device acquires an image of the vehicle identification number actually engraved on the vehicle by using a light and a camera, and recognizes the vehicle identification number engraved on the vehicle by comparing the obtained vehicle identification number image with the stored reference vehicle identification number image, where the reference vehicle identification number image uses the image provided by the manufacturer of the recognition device. The manufacturer of the recognition device creates and creates a reference vehicle identification number image for the engraved text after engraving on the same material as the vehicle for each letter used as the vehicle identification number.

However, the environment for recognizing the vehicle identification number engraved at the vehicle manufacturing site, that is, the location of the lighting or camera of the recognition device, the brightness of the lighting, etc., is different from the environment of the recognition device manufacturer, and the environment is different Moreover, the materials used to manufacture the vehicle for each manufacturer are different from each other, or the material used to manufacture the vehicle is changed or improved.

If the image set by the recognition device manufacturer is fixed and used as the reference vehicle identification number image, the vehicle identification number cannot be accurately recognized or the vehicle is manufactured because the vehicle identification number and the reference vehicle identification number image engraved on the vehicle are different in the actual vehicle manufacturing environment. Since the material used for this purpose is changed or improved, the vehicle identification number engraved on the vehicle and the reference vehicle identification number image are different from each other, so that the vehicle identification number cannot be accurately recognized.

As a prior document of the present invention, Korean Patent Application Laid-Open No. 10-2020-0045023 exists.

The present invention is to solve the problems of the above-mentioned conventional vehicle identification number recognition method, and an object of the present invention is to convert the reference vehicle identification number image used for vehicle identification number recognition of the characters engraved on the vehicle in the actual field. It is to provide an update method of the reference vehicle identification number image optimized for each site by updating the image.

Another object of the present invention is to classify a separate candidate vehicle identification number image and compare the image of the characters engraved on the vehicle with the candidate vehicle identification number image and the reference vehicle identification number image, respectively, so that the correlation between the candidate vehicle identification number image and the reference vehicle identification number image It is to provide a method of updating the reference vehicle identification number image with a high-optimized image.

Another object of the present invention is to generate a split image by dividing by characters constituting the vehicle identification number engraved on the vehicle, and considering the engraving position of the divided image in the vehicle identification number, even if the lighting for each location is different, the optimal reference vehicle identification number image to provide a way to update it.

Another object of the present invention is to stop the update of the reference vehicle identification number image when the update to the reference vehicle identification number image of the required quality is completed, thereby preventing wastage of resources due to unnecessary reference vehicle identification number image update. It is to provide a method for updating the vehicle identification number image.

In order to achieve the object of the present invention, according to the present invention, the method of updating the reference vehicle identification number image used for recognizing the vehicle identification number engraved on the vehicle is engraved on the vehicle, the vehicle identification number is taken by photographing the vehicle identification number consisting of a plurality of characters. Generating an image and generating a segmented image for each letter constituting the vehicle identification number from the vehicle identification number image, Comparing the first reference vehicle identification number image and the divided image, calculating the edge (contour) correlation of the letter; , Comparing the second reference vehicle identification number image and the segmented image to calculate the brightness correlation of the character, and calculating the overall correlation from the edge correlation and the lightness correlation to the character having the highest overall correlation of the segmented image Comprising the steps of recognizing a character, and updating and registering a second reference vehicle identification number image as a candidate vehicle identification number image by comparing the second reference vehicle identification number image and the candidate vehicle identification number image corresponding to the recognized letter with the divided image, The first reference vehicle identification number image is a reference image generated graphically by software using text information of a preset vehicle identification number, and the second reference vehicle identification number image is a reference image generated by photographing the vehicle identification number engraved on the actual vehicle. characterized in that

Preferably, the step of changing and updating the second reference vehicle identification number image to the candidate vehicle identification number image includes: searching for a second reference vehicle identification number image corresponding to the recognized text and a candidate vehicle identification number image corresponding to the recognized text; Calculating the registration brightness correlation between the second reference vehicle identification number image and the divided image, calculating the candidate brightness correlation between the searched candidate vehicle identification number image and the divided image, and the registration brightness correlation and candidate brightness correlation and comparing the sizes of the drawings and updating and registering a second reference vehicle identification number image as a candidate vehicle identification number image.

Here, when the registered brightness correlation is greater than the candidate brightness correlation, the second reference vehicle identification number image is maintained.

Here, when the registered brightness correlation is smaller than the candidate brightness correlation, the second reference vehicle identification number image is updated and registered as the candidate vehicle identification number image.

When the registered brightness correlation is greater than or less than the candidate brightness correlation, the candidate vehicle identification number image is updated and registered as a split image.

Here, the divided image is generated as a divided depth image having depth information and a divided brightness image having brightness information, and the second reference vehicle identification number image is generated as a registered depth image having depth information and a registered brightness image having brightness information, and a candidate The vehicle identification number image is characterized in that it is generated as a candidate depth image having depth information and a candidate brightness image having brightness information.

Here, by comparing the first registered brightness correlation between the divided depth image and the registered depth image with the first candidate brightness correlation between the divided depth image and the candidate depth image to update and register the registered depth image as a candidate depth image, the divided brightness image and comparing the second registered brightness correlation between the and the registered brightness image with the second candidate brightness correlation between the divided brightness image and the candidate brightness image to update and register the registered brightness image as the candidate brightness image.

Preferably, the method for updating the reference vehicle identification number image according to an embodiment of the present invention further comprises the steps of comparing the registration brightness correlation and the update threshold value, and when the registration brightness correlation exceeds the update threshold value, the candidate vehicle identification number It is characterized in that the renewal registration of the second reference vehicle identification number image as the number image is stopped.

Preferably, in another embodiment of the present invention, the method of updating the reference vehicle identification number image further comprises the step of determining whether an update period has arrived, and when the update period has arrived, a second reference vehicle identification number image as a candidate vehicle identification number image for a set time It is characterized in that the image is updated and registered.

Preferably, in another embodiment of the present invention, the method of updating the reference vehicle identification number image further comprises counting the number of recognitions for each character, and when the number of times of recognition for each character exceeds the threshold number, it is replaced with a candidate vehicle identification number image. 2 It is characterized in that the renewal registration of the reference vehicle identification number image is stopped.

Preferably, in another embodiment of the present invention, the method of updating the reference vehicle identification number image further includes determining whether an update stop command is input, and when the update stop command is input, the second reference vehicle identification number image as a candidate vehicle identification number image It is characterized in that the renewal registration of the number image is stopped.

The method for updating the reference vehicle identification number image according to the present invention has various effects as follows.

First, the method for updating the reference vehicle identification number image according to the present invention updates the reference vehicle identification number image used for vehicle identification number recognition with an image of characters engraved on the vehicle at the actual site, thereby manufacturing a vehicle and imprinting the vehicle identification number. It can be updated with the optimized reference vehicle identification number image considering the environment of the site.

Second, the method for updating the reference vehicle identification number image according to the present invention distinguishes a separate candidate vehicle identification number image and compares the image of the characters engraved on the vehicle with the candidate vehicle identification number image and the reference vehicle identification number image, respectively, to compare the candidate vehicle identification number image and the reference By updating the reference vehicle identification number image with an image with high correlation among vehicle identification number images, the reference vehicle identification number image can be updated with an optimized image.

Third, the updating method of the reference VIN image according to the present invention generates a divided image by dividing by characters constituting the VIN number engraved on the vehicle, and by updating the reference VIN image in consideration of the engraving position of the divided image in the VIN. , even if the lighting for each location is different, it can be updated with the optimal reference vehicle identification number image.

Fourth, the update method of the reference vehicle identification number image according to the present invention stops the update of the reference vehicle identification number image when the update to the reference vehicle identification number image of the required quality is completed, thereby wasting resources due to unnecessary update of the reference vehicle identification number image can prevent

1 is a functional block diagram for explaining a vehicle number engraving system according to the present invention.
2 is a functional block diagram illustrating a recognition device according to the present invention.
3 is a functional block diagram illustrating an example of an update unit according to the present invention.
4 is a flowchart for explaining a method of updating a reference vehicle identification number according to the present invention.
5 is a flowchart for explaining an example of a step of calculating a correlation according to the present invention.
6 is a view for explaining a first reference vehicle identification number image stored in the first database unit.
7 is a view for explaining a second reference vehicle identification number image that is initially registered and stored in the second database unit.
8 is a flowchart for explaining an example of the step of updating the second reference vehicle identification number image according to the present invention.
9 is a view for explaining an example of a method of recognizing the characters of the divided image by comparing the divided image for each vehicle identification number character generated from the vehicle identification number image with a first reference vehicle identification number image and a second reference vehicle identification number image.
10 is a view for explaining an example of updating a second reference vehicle identification number image of a second reference database to a candidate vehicle identification number image in the present invention.
11 is a flowchart for explaining an example of stopping the update in the method for updating the reference vehicle identification number image according to the present invention.

It should be noted that the technical terms used in the present invention are only used to describe specific embodiments, and are not intended to limit the present invention. In addition, the technical terms used in the present invention should be interpreted as meanings generally understood by those of ordinary skill in the art to which the present invention belongs, unless otherwise specifically defined in the present invention, and excessively comprehensive It should not be construed in the meaning of a human being or in an excessively reduced meaning. In addition, when the technical term used in the present invention is an incorrect technical term that does not accurately express the spirit of the present invention, it should be understood by being replaced with a technical term that can be correctly understood by those skilled in the art.

Also, the singular expression used in the present invention includes the plural expression unless the context clearly dictates otherwise. In the present invention, terms such as “consisting of” or “comprising” should not be construed as necessarily including all of the various components or various steps described in the invention, some of which components or some steps are included. It should be construed that it may not, or may further include additional components or steps.

In addition, it should be noted that the accompanying drawings are only for easy understanding of the spirit of the present invention, and should not be construed as limiting the spirit of the present invention by the accompanying drawings.

1 is a functional block diagram for explaining a vehicle number engraving system according to the present invention.

Referring to FIG. 1 in more detail, the engraving device 10 receives information on the vehicle identification number to be engraved, and when the vehicle is disposed at the engraving position, the vehicle number is engraved on the set part of the vehicle or the set position. The engraving device 10 engraves the vehicle identification number by engraving through a laser or a pin, etc. according to the material of the vehicle, the engraving position, or the engraving method. In this way, the vehicle identification number may be engraved, which is within the scope of the present invention.

When the imprinting of the vehicle identification number on the vehicle is completed through the engraving device 10 , the recognition device 30 captures the vehicle identification number engraved on the vehicle to generate a vehicle identification number image, and the generated vehicle identification number image and the first reference vehicle identification number The edge correlation is calculated by comparing the images, and the brightness correlation is calculated by comparing the generated vehicle identification number image with the second reference vehicle identification number image.

By calculating the overall correlation from the calculated edge correlation and brightness correlation, the vehicle identification number engraved on the vehicle is recognized for each character. By comparing the recognized vehicle identification number and the received vehicle identification number information, it is determined whether the same vehicle identification number as the received vehicle identification number is engraved on the vehicle.

Here, the first reference vehicle identification number image is a reference image generated graphically by software using text information of a preset vehicle identification number, and the second reference vehicle identification number image is a reference image generated by photographing the vehicle identification number engraved on the actual vehicle. am. The first reference vehicle identification number image is a reference image that does not change because it is graphically generated by software using text information of a preset vehicle identification number, whereas the second reference vehicle identification number image is a photographed image of the vehicle identification number engraved on the actual vehicle. It is a reference image that is updated and registered as an image with higher correlation by comparing the previously stored reference vehicle identification number image with the image.

The manager terminal 50 is a terminal used by the manager, and a personal computer, a laptop computer, a smart phone, etc. may be used, and the manager terminal 50 sends an alarm message when receiving a signal that an incorrect vehicle identification number is engraved from the recognition device 30 . You can create it and output it to the manager.

2 is a functional block diagram illustrating a recognition device according to the present invention.

Referring to FIG. 2 in more detail, the image generating unit 110 generates a vehicle identification number image by photographing the vehicle identification number engraved on the vehicle.

Here, the image generating unit 110 is provided with a camera for photographing the surface on which the vehicle identification number is engraved and a light irradiating light to the surface on which the vehicle identification number is engraved. It is equipped with at least three or more lights to irradiate.

The image generating unit 110 generates a brightness image having brightness information of the surface on which the vehicle identification number is engraved and a depth image having depth information of the surface on which the vehicle identification number is engraved using lighting and a camera, and a plurality of lights are turned on. In this state, the vehicle identification number is photographed to generate a brightness image having brightness information, and a plurality of lights are sequentially turned on, each vehicle identification number is photographed, and a depth image having depth information is generated from them.

Here, the brightness image is an image having information on the reflection intensity of a specific pixel from a grayscale point of view.

Here, the depth image is an image having depth information of the surface on which the vehicle identification number is engraved using the photometric stereo technique. This technique extracts the three-dimensional shape of an object using images obtained by sequentially irradiating the above lights. As the number of lights increases, the three-dimensional shape of the object can be extracted more reliably. The photometric stereo technique is a well-known technique and a detailed description thereof will be omitted.

The division unit 120 generates the vehicle identification number image generated by the image generation unit 110 as a split image for each character by using edge information of an ideal vehicle identification number image generated from vehicle identification number information to be engraved on the vehicle. A divided brightness image is generated by dividing the medium brightness image by characters constituting the vehicle identification number, and a divided depth image is generated by dividing the vehicle identification number by characters constituting the vehicle identification number from the depth image of the vehicle identification number image.

The edge comparison unit 130 compares the divided depth image generated from the depth image with the first reference vehicle identification number image stored in the first reference database unit 140 to calculate a first edge correlation (CP), and brightness A second edge correlation (CI) is calculated by comparing the divided brightness image generated from the image with the first reference vehicle identification number image stored in the first reference database unit 140 . The edge comparison unit 130 calculates the edge correlation for each character of the character corresponding to the split image and the first reference vehicle identification number image from the first edge correlation diagram and the second edge correlation diagram.

The brightness comparison unit 150 compares the divided depth image generated from the depth image with the registered depth image among the second reference vehicle identification number images stored in the second reference database unit 160 to obtain a first brightness correlation (GP _R ) and calculates the second brightness correlation (GI _R ) by comparing the divided brightness image generated from the brightness image with the registered brightness image among the second reference vehicle identification number images stored in the second reference database unit 160 . . The brightness comparison unit 160 calculates the brightness correlation for each character of the character corresponding to the divided image and the second reference vehicle identification number image from the first brightness correlation diagram and the second brightness correlation diagram.

The recognition unit 170 calculates an overall correlation for each character from the edge correlation and the brightness correlation, and recognizes a character having the highest overall correlation as a character corresponding to the divided image. When the recognition unit 170 completes recognizing all the characters corresponding to each segmented image, it compares the received vehicle identification number information with the vehicle identification number consisting of the recognized characters to determine whether the same vehicle identification number as the received vehicle identification number is engraved on the vehicle. judge When the recognition unit 170 does not recognize some characters of the engraved vehicle identification number or some characters of the engraved vehicle identification number are erroneously engraved, the recognition unit 170 may generate alarm information and transmit the generated alarm information to the manager terminal.

Meanwhile, the update unit 190 searches the candidate database unit 180 for a candidate vehicle identification number image corresponding to the recognized character. The update unit 180 calculates the candidate brightness correlation between the segmented image and the candidate vehicle identification number image corresponding to the recognized character, and the registration brightness correlation between the second reference vehicle identification number image corresponding to the recognized character and the divided image is compared with the candidate brightness correlation, and the second reference vehicle identification number image stored in the second reference database unit 160 is updated and registered as a candidate vehicle identification number image based on the comparison result.

3 is a functional block diagram illustrating an example of an update unit according to the present invention.

Referring to FIG. 3 in more detail, the search unit 191 searches for the second reference vehicle identification number image corresponding to the recognized character in the second reference database unit 160 and the candidate chassis corresponding to the recognized character. The number image is searched for in the candidate database unit 180 . Here, the search unit 191 searches for a second reference vehicle identification number image and a candidate vehicle identification number image having the same character as the recognized character and the same position identifier as the position identifier of the divided image.

The registered brightness comparison unit 193 calculates the registered brightness correlation ( _GR ) between the divided image used to recognize the character and the searched second reference vehicle identification number image, and the candidate brightness comparison unit 195 recognizes the character Calculate the candidate brightness correlation (G _C ) between the segmented image used to perform the search and the searched candidate vehicle identification number image. According to the field to which the present invention is applied, the registered brightness comparison unit 193 may extract the brightness correlation between the split image previously used to recognize the character and the second reference vehicle identification number image and use it as the registered brightness correlation diagram.

More specifically, the registered brightness comparison unit 193 compares the first registered brightness between the divided depth image among the divided images used to recognize the character and the registered depth image among the searched second reference vehicle identification number image (GP _R ) and calculates the second registered brightness correlation (GI _R ) between the divided brightness image among the divided images used to recognize the character and the registered brightness image among the searched second reference vehicle identification number image. On the other hand, the candidate brightness comparison unit 195 calculates the first candidate brightness correlation (GP _C ) between the divided depth image among the divided images used to recognize the character and the candidate depth image among the searched candidate vehicle identification number images, and A second candidate brightness correlation (GI _C ) between the divided brightness image among the divided images used for recognition and the candidate brightness image among the searched candidate vehicle identification number image is calculated.

The update registration unit 197 updates and registers the second reference vehicle identification number image retrieved from the second reference database as a candidate vehicle identification number image based on the registration brightness correlation and candidate brightness correlation, and the candidate vehicle identification number image retrieved from the candidate database unit. is updated and registered as a split image.

Depending on the field to which the present invention is applied, the update registration unit 197 may batch update the registered depth image and the registered brightness image to the candidate depth image and the candidate brightness image based on the registered brightness correlation and the candidate brightness correlation, or the first registered brightness correlation Based on the degree and the first candidate brightness correlation, only the registered depth image can be individually updated to the candidate depth image, or only the registered brightness image can be individually updated to the candidate brightness image based on the second registered brightness correlation and the second candidate brightness correlation. there is.

4 is a flowchart for explaining a method of updating a reference vehicle identification number according to the present invention.

Referring to FIG. 4 in more detail, a vehicle identification number image is generated by photographing the vehicle identification number, and the vehicle identification number image is generated as a depth image having depth information and a brightness image having brightness information (S110).

A split image for each character constituting the vehicle identification number is generated from the depth image and the brightness image (S120). A divided depth image for each letter constituting the vehicle identification number is generated from the depth image, and a divided brightness image for each letter constituting the vehicle identification number is generated from the brightness image. When generating a divided depth image or divided brightness image for each character from a depth image or a brightness image, the location identifier is mapped and recorded in the divided depth image and the divided brightness image in consideration of the order or position of the characters. For example, if the vehicle identification number consists of 19 characters, the brightness image is divided into each letter to generate 19 divided brightness images from the brightness image. 1" is mapped and recorded, and the location identifier "2" can be mapped and recorded in the divided brightness image corresponding to the character in the second position. For example, if the vehicle identification number consists of 19 characters, the depth image is divided into characters to generate 19 divided depth images from the depth image. 1" is mapped and recorded, and the location identifier "2" can be mapped and recorded in the divided depth image corresponding to the second character.

By comparing the split depth image and the split brightness image with the first reference vehicle identification number image, the edge correlation for each character is calculated, and the split depth image and the divided brightness image are compared with the second reference vehicle identification number image to calculate the brightness correlation for each character, , calculates the total correlation for each character from the edge correlation and the brightness correlation (S150).

Based on the overall correlation for each character, the character having the highest overall correlation is recognized as a character corresponding to the divided image (S170).

By comparing the second reference vehicle identification number image corresponding to the recognized character and the candidate vehicle identification number image with the divided image, an image with high correlation with the divided image among the second reference vehicle identification number image and the candidate vehicle identification number image is selected as the second reference vehicle identification number image to register for renewal (S190).

5 is a flowchart for explaining an example of a step of calculating a correlation according to the present invention.

Referring to FIG. 5 in more detail, by comparing the divided depth image generated from the depth image among the divided images with the first reference vehicle identification number image stored in the first reference database unit, the first edge correlation (CP) for each character is calculated (S131), and a second edge correlation (CI) for each character is calculated by comparing the divided brightness image generated from the brightness image among the divided images with the first reference vehicle identification number image stored in the first reference database unit. (S133).

Meanwhile, the location identifier of the divided image is determined (S135), and a registered depth image and a registered brightness image corresponding to the same location as the divided image are searched for in the second reference database unit 160 based on the determined location identifier (S136). The first brightness correlation (GP _R ) for each character is calculated by comparing the divided depth image with the searched registered depth image (S137), and the second brightness correlation for each character is compared with the registered brightness image for which the divided brightness image is searched (S137). GI _R ) is calculated (S137).

Here, the first edge correlation CP is calculated as in Equation (1) below.

[Equation 1]

here

t(r,c) is a template, a reference vehicle identification number image, i(r,c) is a split image, r is a row, and c is a column.

The second edge correlation (CI), the first brightness correlation (GP _R ), and the second brightness correlation (GI _R ) are also calculated as in Equation (1) above, but the edge correlation is only correlated with edge information. The degree of correlation is calculated, and the degree of correlation is calculated for the entire brightness image.

Depending on the field to which the present invention is applied, the first edge correlation, the second edge correlation, the first lightness correlation, and the second lightness correlation may be calculated in various ways, which fall within the scope of the present invention.

An overall correlation for each character is calculated from the first edge correlation (CP), the second edge correlation (CI), the first brightness correlation (GP _R ), and the second brightness correlation ( _GIR ) (S139). Here, the total correlation TC may be calculated as in Equation (2) below.

[Equation 2]

TC=(CP+CI+GP _R +GI _R )/4

Depending on the field to which the present invention is applied, the overall correlation may be calculated in various ways, and this falls within the scope of the present invention.

6 is a view for explaining a first reference vehicle identification number image stored in the first database unit.

Referring to FIG. 6 , the characters engraved on the vehicle are uppercase alphabetic characters, special characters, numbers, etc., and the characters engraved on the vehicle include preset character information for each character, for example, vector font, font width, font height, font pitch. , is imprinted on the vehicle through an imprinting device based on character information such as font thickness. The first reference vehicle identification number image stored in the first database unit is a graphic character image generated from the same character information for each character engraved on the vehicle. Here, the graphic character image may be generated through various applications, and the generated graphic character image is an ideal character image having accurate edges.

7 is a view for explaining a second reference vehicle identification number image that is initially registered and stored in the second database unit.

Referring to FIG. 7 , a manufacturer of a recognition device creates a second reference vehicle identification number image from an image taken after engraving an actual character on a steel plate of the same material as the vehicle with preset character information, and then generating a second reference vehicle identification number image in the second reference database Register to save.

The same characters are repeatedly engraved on the steel plate by the same number as the number of characters constituting the vehicle identification number, and the engraved characters are photographed to generate a depth image and a brightness image, and a split image for each character from the depth image and the brightness image is initially second It is created as a reference vehicle identification number image. The second reference vehicle identification number image is generated as a registered depth image (D) generated from the divided depth image and a registered brightness image (I) generated from the divided brightness image. In consideration of the position of the segmented image, the location identifier of the segmented image is mapped and recorded in the registered depth image and registered brightness image.

For example, if the vehicle identification number consists of 19 characters, after engraving 19 "A", one of the letters used as the vehicle identification number, on the same iron plate as the vehicle, photograph the iron plate engraved with 19 A's to provide depth information A depth image with , and a brightness image with brightness information are generated. The depth image is divided by character to generate 19 divided depth images, and the brightness image is divided for each character to generate 19 divided brightness images. Create 19 registered depth images by mapping the location identifier of the split image to the split depth image for "A" considering the location of the split image, and consider the location of the split image to split the image into split brightness image for "A" 19 registered brightness images mapped with location identifiers of

By repeating the same operation for all characters used in the vehicle identification number, the registration depth image and registration brightness image by location identifier for all letters are generated as the second reference vehicle identification number image, and the generated second reference vehicle identification number image is stored in the second database. save in wealth

In this way, by generating a split image of a character for each position where the character is engraved in the vehicle identification number, even if the lighting is non-uniform for each position where the character is engraved in the vehicle identification number, it is possible to more accurately recognize each character.

Thereafter, the second reference vehicle identification number image initially stored in the second database unit is updated with an image of characters actually engraved on the vehicle at the site where the actual recognition device is installed and operated. As such, by updating the initially stored second reference vehicle identification number image with the image of characters actually engraved on the vehicle at the site where the actual recognition device is installed and operated, the optimized second It can be updated with the reference vehicle identification number image.

8 is a flowchart for explaining an example of the step of updating the second reference vehicle identification number image according to the present invention.

Referring to FIG. 8 in more detail, a second reference vehicle identification number image corresponding to the recognized character is searched for in the second reference database, and a candidate vehicle identification number image corresponding to the recognized letter is searched for in the candidate database (S191). A second reference vehicle identification number image and a candidate vehicle identification number image of the same text as the recognized text and the same location identifier as the location identifier of the split image of the recognized text are searched for.

The registration brightness correlation between the searched second reference vehicle identification number image and the divided image of the recognized character is calculated (S193), and the candidate brightness correlation between the searched candidate vehicle identification number image and the divided image of the recognized character is calculated (S195).

By comparing the size of the registered brightness correlation and the candidate brightness correlation (S196), if the candidate brightness correlation is greater than the registered brightness correlation, the searched second reference vehicle identification number image is updated and registered as the candidate vehicle identification number image (S197). However, if the candidate brightness correlation is smaller than the registered brightness correlation, the existing second reference vehicle identification number image is maintained without updating (S198).

On the other hand, the searched candidate vehicle identification number image is updated as a divided image of the recognized character and stored and registered in the candidate database unit (S199).

The divided image used to recognize the character is composed of a divided depth image having depth information and a divided brightness image having brightness information, and the second reference vehicle identification number image is a registered depth image having depth information and a registered brightness image having brightness information. , and the candidate vehicle identification number image is composed of a candidate depth image having depth information and a candidate brightness image having brightness information.

In an embodiment, the registered depth image and the registered brightness image of the second reference vehicle identification number image may be collectively updated to the candidate depth image and the candidate brightness image. For example, the registered brightness calculated from the sum of the first registered brightness correlation between the registered depth image and the divided depth image (GP _R ) and the second registered brightness correlation between the registered brightness image and the divided brightness image ( _GIR ). The correlation (G _R =GP _R +GI _R ) was calculated as the first candidate lightness correlation between the candidate depth image and the split depth image (GI _C ) and the second candidate lightness correlation between the candidate lightness image and the split lightness image (GI). When the candidate brightness correlation is greater than the registration brightness correlation (G _C >GR ) compared to the candidate brightness correlation calculated by the sum of _C ) (G _{C =} GP _C +GI _c ), the registration depth image and the registration brightness The images are collectively updated with a candidate depth image and a candidate brightness image.

In another embodiment, the registered depth image and the registered brightness image of the second reference vehicle identification number image may be individually updated as a candidate depth image and a candidate brightness image. For example, by comparing the first registered brightness correlation (GP _R ) and the first candidate brightness correlation (GP _C ), when the first candidate brightness correlation is greater than the first registered brightness correlation, only the registered depth image is individually If the second candidate brightness correlation is larger than the second registered brightness correlation by updating the candidate depth image or comparing the second registered brightness correlation (GI _R ) and the second candidate brightness correlation (GI _C ), the registered brightness image can be individually updated with candidate brightness images.

9 is a view for explaining an example of a method of recognizing the characters of the divided image by comparing the divided image for each vehicle identification number character generated from the vehicle identification number image with a first reference vehicle identification number image and a second reference vehicle identification number image.

Referring to FIG. 9, the depth image having depth information in the vehicle identification number image is divided by letter to generate a divided depth image (D) for each letter, and the brightness image having brightness information in the vehicle identification number image is divided by letter. Create a split brightness image (I) for each character.

The first edge correlation for each character is calculated by comparing the split depth image with the first reference vehicle identification number image for each character stored in the first reference database unit DB1, and the divided brightness image is compared to the first reference database unit DB1 The second edge correlation for each character is calculated by comparing it with the first reference vehicle identification number image for each character stored in the .

In addition, by comparing the divided depth image with the registered depth image mapped to the same position identifier as the divided depth image among the second reference vehicle identification number images for each character stored in the second reference database unit DB2, the correlation of the first brightness for each character was obtained. Calculation and comparing the divided brightness image with the registered brightness image mapped with the same location identifier as the divided brightness image among the second reference vehicle identification number images for each character stored in the second reference database unit DB2 to correlate the second brightness for each character calculate the degree

A character having the highest overall correlation from the first edge correlation, the second edge correlation, the first brightness correlation, and the second brightness correlation is recognized as the character of the divided image.

Here, on the basis of the registered brightness correlation and candidate brightness correlation, the registered depth image and the registered brightness image of each position stored in the second reference database unit are the candidate depth image and the candidate depth image of the same character stored in the candidate database unit. Candidate brightness may be updated as an image.

10 is a view for explaining an example of updating a second reference vehicle identification number image of a second reference database to a candidate vehicle identification number image in the present invention.

Referring to FIG. 10 in more detail, when the character located at the position identifier 7 in the vehicle identification number is first recognized as “H”, the registration depth located at the position identifier 7 in the second reference database unit and registered with the same character H Search for images and registered brightness images. The first registered brightness correlation between the divided depth image and the searched registered depth image is calculated, and the second registered brightness correlation between the divided brightness image and the searched registered brightness image is calculated, and the first registered brightness correlation and the second registered brightness are calculated. Calculate the registered brightness correlation from the correlation.

In addition, a candidate depth image and a candidate brightness image located at the position identifier 7 in the candidate database unit and registered with the same letter H are searched for. The first candidate brightness correlation and the second candidate brightness correlation are calculated by calculating the first candidate brightness correlation between the divided depth image and the searched candidate brightness image, and calculating the second candidate brightness correlation between the divided brightness image and the searched candidate brightness image. Calculate the candidate brightness correlation from the correlation. If the registered brightness correlation is greater than the candidate brightness correlation, the registered depth image and the registered brightness image are maintained without updating, but the candidate depth image and the candidate brightness image are updated and registered in the candidate database as a divided depth image and a divided brightness image do.

Repeat the same operation, and when the character located at position identifier 7 in the vehicle identification number is recognized as “H” for the third time, the registered depth image and registered brightness image registered as the same letter H as the position identifier 7 in the second reference database unit search for The first registered brightness correlation between the divided depth image and the searched registered depth image is calculated, and the second registered brightness correlation between the divided brightness image and the searched registered brightness image is calculated, and the first registered brightness correlation and the second registered brightness are calculated. Calculate the registered brightness correlation from the correlation.

Also, a candidate depth image and a candidate brightness image registered with the same letter H as the position identifier 7 are searched for in the candidate database unit. The first candidate brightness correlation and the second candidate brightness correlation are calculated by calculating the first candidate brightness correlation between the divided depth image and the searched candidate brightness image, and calculating the second candidate brightness correlation between the divided brightness image and the searched candidate brightness image. Calculate the candidate brightness correlation from the correlation. If the candidate brightness correlation is greater than the registered brightness correlation, the candidate depth image and the candidate brightness image are replaced with the registered depth image and the registered brightness image previously registered in the second reference database unit as the registered depth image and the registered brightness image. 2 Update registration in the reference database unit, and update and register the divided depth image and the divided brightness image as the candidate depth image and the candidate brightness image in the candidate database unit instead of the candidate depth image and the candidate brightness image previously registered in the candidate database unit .

11 is a flowchart for explaining an example of stopping the update in the method for updating the reference vehicle identification number image according to the present invention.

Referring to FIG. 11 in more detail, the number of times of recognition for each character at each position is counted to determine whether the number of times of recognition for each character at each position exceeds a threshold number (S211). For example, by counting the number of times that the character "H" located in the position identifier 7 is recognized, it is determined whether the number of times the character "H" located in the position identifier 7 is recognized exceeds a threshold number of times. If the number of recognition times for each character at each position does not exceed the threshold number, the second reference vehicle identification number image is continuously updated and registered as a candidate vehicle identification number image for the character recognized at the corresponding position (S219).

When the number of recognition times for each character in each position exceeds the threshold number, it is determined whether the registered brightness correlation exceeds the update threshold value by comparing the registered brightness correlation with the update threshold (S213). For example, it is determined whether the correlation between the registration brightness of the second reference vehicle identification number image with respect to the letter "H" located in the position identifier 7 exceeds a set update threshold.

Even if the number of recognition exceeds the threshold number, if the registration brightness correlation does not exceed the update threshold value, the second reference vehicle identification number image is continuously updated and registered as a candidate vehicle identification number image for the character recognized at the corresponding position (S219) .

On the other hand, when the number of recognition times for each character in each position exceeds the threshold number and the registration brightness correlation exceeds the update threshold value, a user command requesting to stop updating the second reference vehicle identification number image to the candidate vehicle identification number image is input It is determined whether or not (S215).

When a user command is input, updating the second reference vehicle identification number image to the candidate vehicle identification number image is stopped (S217).

In the present invention, when the number of times of recognition for each character at each location of the vehicle identification number exceeds the threshold number, it is unnecessary to continuously update the second reference vehicle identification number image to the candidate vehicle identification number image. can do.

In addition, in the present invention, when the registration brightness correlation for each character at each position of the vehicle identification number exceeds the set update threshold, the process and memory required to continuously update the second reference vehicle identification number image to the candidate vehicle identification number image unnecessarily. waste of resources can be avoided.

In addition, in the present invention, the second reference vehicle identification number image is updated to the candidate vehicle identification number image according to the user's command, or the second reference vehicle identification number image is updated with the candidate vehicle identification number image only for a set period in a set period Accordingly, it is possible to prevent wastage of resources such as a process and memory required to update the second reference vehicle identification number image into a candidate vehicle identification number image.

Meanwhile, the above-described embodiments of the present invention can be written as a program that can be executed on a computer, and can be implemented in a general-purpose digital computer that operates the program using a computer-readable recording medium.

The computer-readable recording medium includes a magnetic storage medium (eg, ROM, floppy disk, hard disk, etc.), an optically readable medium (eg, CD-ROM, DVD, etc.) and a carrier wave (eg, Internet storage media such as transmission via

Although the present invention has been described with reference to the embodiment shown in the drawings, which is only exemplary, those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. Accordingly, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

10: engraving device 30: recognition device
50: manager terminal
110: image generating unit 120: dividing unit
130: edge comparison unit 140: first reference database unit
150: brightness comparison unit 160: second reference database unit
170: recognition unit 180: candidate database unit
190: update
191: search unit 193: registered name comparison unit
195: candidate brightness comparison unit 197: renewal register

Claims (12)

In the updating method of the reference vehicle identification number image used to recognize the vehicle identification number engraved on the vehicle,
generating a vehicle identification number image by photographing a vehicle identification number engraved on the vehicle, comprising a plurality of characters, and generating a divided image for each letter constituting the vehicle identification number from the vehicle identification number image;
Comparing the first reference vehicle identification number image and the divided image, calculating the edge (contour) correlation of the character;
Comparing the second reference vehicle identification number image and the divided image to calculate the brightness correlation of the character;
recognizing the character of the divided image as a character having the highest overall correlation by calculating an overall correlation from the edge correlation and the brightness correlation; and
Comprising the step of comparing a second reference vehicle identification number image and a candidate vehicle identification number image corresponding to the recognized character with the divided image to update and register a second reference vehicle identification number image as a candidate vehicle identification number image,
The first reference vehicle identification number image is a reference image generated graphically by software using text information of a preset vehicle identification number,
The method for updating the reference vehicle identification number image, characterized in that the second reference vehicle identification number image is a reference image generated by photographing the vehicle identification number engraved on the actual vehicle.
The method of claim 1, wherein the updating of the second reference vehicle identification number image to the candidate vehicle identification number image comprises:
retrieving a second reference vehicle identification number image corresponding to the recognized text and a candidate vehicle identification number image corresponding to the recognized text;
calculating a registration brightness correlation between the searched second reference vehicle identification number image and the divided image;
calculating a candidate brightness correlation between the searched candidate vehicle identification number image and the divided image; and
and comparing the size of the registered brightness correlation and the candidate brightness correlation to update and register the second reference vehicle identification number image as the candidate vehicle identification number image.
3. The method of claim 2,
The method of updating a reference vehicle identification number image, characterized in that the second reference vehicle identification number image is maintained when the registered brightness correlation is greater than the candidate brightness correlation.
4. The method of claim 3,
When the registered brightness correlation is smaller than the candidate brightness correlation, the second reference vehicle identification number image is updated and registered as the candidate vehicle identification number image.
5. The method of claim 4,
When the registered brightness correlation is greater or less than the candidate brightness correlation,
The updating method of the reference vehicle identification number image, characterized in that the candidate vehicle identification number image is updated and registered as the divided image.
6. The method of claim 5,
The divided image is generated as a divided depth image having depth information and a divided brightness image having brightness information,
The second reference vehicle identification number image is generated as a registered depth image having depth information and a registered brightness image having brightness information, and the candidate vehicle identification number image is generated as a candidate depth image having depth information and a candidate brightness image having brightness information A method of updating the reference vehicle identification number image, characterized in that
7. The method of claim 6,
comparing the first registered brightness correlation between the divided depth image and the registered depth image with a first candidate brightness correlation between the divided depth image and the candidate depth image to update and register the registered depth image as the candidate depth image;
Comparing the second registered brightness correlation between the divided brightness image and the registered brightness image with the second candidate brightness correlation between the divided brightness image and the candidate brightness image, updating and registering the registered brightness image as the candidate brightness image How to update the reference vehicle identification number image.
5. The method of claim 4, wherein the updating method of the reference vehicle identification number image comprises:
Further comprising the steps of comparing the registration brightness correlation and the update threshold,
When the registration brightness correlation exceeds the update threshold, updating and registering the second reference vehicle identification number image as the candidate vehicle identification number image is stopped.
5. The method of claim 4, wherein the updating method of the reference vehicle identification number image comprises:
Further comprising the step of determining whether the update cycle has arrived,
When the update period arrives, the method of updating the reference vehicle identification number image, characterized in that for a set period of time, the renewal registration of the second reference vehicle identification number image as the candidate vehicle identification number image.
5. The method of claim 4, wherein the updating method of the reference vehicle identification number image comprises:
Further comprising the step of counting the number of recognition for each character,
When the number of times of recognition for each character exceeds a threshold number of times, updating and registering the second reference vehicle identification number image as the candidate vehicle identification number image is stopped.
5. The method of claim 4, wherein the updating method of the reference vehicle identification number image comprises:
Further comprising the step of determining whether an update stop command is input,
When the update stop command is input, the method of updating a reference vehicle identification number image, characterized in that the renewal and registration of a second reference vehicle identification number image as the candidate vehicle identification number image is stopped.
12. The method according to any one of claims 1 to 11,
Further comprising the step of determining the location identifier of the divided image in the vehicle identification number image,
The second reference vehicle identification number image and the candidate vehicle identification number image corresponding to the recognized characters are the second reference vehicle identification number image and the candidate vehicle identification number image existing at the same location as the determined location identifier. method.

Images