JP7081404B2 - License plate identification device, license plate identification method and program - Google Patents

Description

The present invention relates to a license plate identification device, a license plate identification method and a program.

In recent years, an ETC (Electronic Toll Collection) system has been developed, and a free-flow ETC system is being developed. In the free flow ETC system, for example, a wireless antenna, a vehicle detection sensor, a camera, or the like is installed on a gate-shaped pillar installed across lanes on an exit ramp of a highway or on a main lane. The free-flow ETC system allows the driver to pass under the radio antenna without slowing down or temporarily stopping the vehicle, so that communication is performed between the radio antenna and the ETC on-board unit.

In the ETC system, the vehicle is identified by identifying the license plate. For example, in Patent Document 1, a license plate recognition device that uses a neural network to identify three types of license plates classified according to size and four types of backgrounds, and color-identifies the identified license plates. Is disclosed.

Japanese Unexamined Patent Publication No. 4-205284

However, in the license plate recognition device described in Patent Document 1, when the distance between the camera and the license plate changes, the size of the display area of the license plate displayed on the camera changes depending on this distance. Misrecognition may occur. Therefore, there is room for improvement in applying the technique described in Patent Document 1 to the free-flow ETC system.

Therefore, the present invention has been made in view of the above problems, and an object of the present invention is to be able to more accurately identify the type of license plate even when the shooting distance changes. To provide a license plate identification device, a license plate identification method and a program.

In order to solve the above problem, according to a certain viewpoint of the present invention, an image cutout portion for acquiring a cutout image which is an image cut out from one or a plurality of images, and a number based on the position of the plurality of cutout images. The image includes an image extraction unit that extracts an identification image that is an image that may include a region in which the plate is displayed, and an identification unit that identifies the type of the number plate based on the identification image. The cutout unit is provided with a number plate identification device that cuts out a plurality of cutout images having different sizes from the one or a plurality of images by means of a plurality of cutout frames having different sizes.

The image extraction unit may calculate a score for each of the cutout images based on the certainty of the license plate, and select a predetermined number of cutout images from the plurality of cutout images in descending order of the score. ..

The image extraction unit may extract the identification image of a predetermined size from the one or a plurality of images based on the positions of the predetermined number of cutout images.

The cutout frame may have a size corresponding to the area of the license plate.

The identification unit may use information about the cutout frame corresponding to the identification image as one of the input information to the classifier.

The identification unit identifies the license plate by classifying the identification image into at least one of three types, that is, two types of the license plate and an image that does not include the area where the license plate is displayed. You may.

Further, in order to solve the above-mentioned problems, according to another viewpoint of the present invention, it is possible to acquire a cut-out image which is an image cut out from one or a plurality of images, and based on the position of the plurality of the cut-out images. , Extracting an identification image, which is an image that may include a number plate image, and identifying the number plate from the identification image, the acquisition of the cutout image of size. A method for identifying a number plate is provided in which a plurality of cutout images having different sizes are cut out from the one or a plurality of images by a plurality of different cutout frames.

Further, in order to solve the above-mentioned problems, according to another viewpoint of the present invention, a computer is provided with an image cutting unit for acquiring a cut-out image which is an image cut out from one or a plurality of images, and a plurality of the cut-out images. It is made to function as an image extraction unit for extracting an identification image which is an image having a possibility of including an image of a number plate based on the position of the above, and an identification unit for identifying the number plate from the identification image. The image cutting unit is provided with a program for cutting out a plurality of cut-out images having different sizes from the one or a plurality of images by a plurality of cutting frames having different sizes.

As described above, according to the present invention, there is provided a license plate identification device, a license plate identification method and a program capable of more accurately identifying the type of license plate even when the shooting distance changes. It becomes possible.

It is a block diagram which shows an example of the structure of the license plate identification system which concerns on one Embodiment of this invention. It is a schematic diagram which shows the application example of the license plate identification system which concerns on the same embodiment. It is a block diagram which shows an example of the structure of the license plate identification apparatus which concerns on the same embodiment. It is explanatory drawing for demonstrating the cut-out image which concerns on the same embodiment. It is a figure which shows the list of the cut-out image in FIG. It is explanatory drawing for demonstrating the identification method of the license plate by the identification part which concerns on the same embodiment. It is a flow chart which shows the flow of license plate identification by the license plate identification apparatus which concerns on the same embodiment. It is explanatory drawing explaining the license plate identification method by the license plate identification apparatus which concerns on the same embodiment. It is a block diagram which shows an example of the hardware composition of the license plate identification apparatus which concerns on the dynamic embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the present specification and the drawings, components having substantially the same functional configuration are designated by the same reference numerals, so that duplicate description will be omitted.

Further, when it is not necessary to particularly distinguish each of the plurality of components having substantially the same functional configuration, only the same reference numerals are given to each of the plurality of components.

<1. Background ＞
There are three types of license plates: a vehicle registration number plate attached to a registered vehicle, a vehicle number plate attached to a light vehicle, and a taxation sign attached to a motorized bicycle or the like. There are two types of vehicle registration number plates, large number plates (large plates) and medium-sized number plates (middle plates), depending on the size. The vehicle number plate includes a medium-sized number plate (middle plate) and a small number plate (small plate) depending on the size. Further, there are two types of small plates in which the arrangement of characters differs depending on the type of vehicle. The size of the large plate is 220 mm in length and 440 mm in width, the size of the middle plate is 165 mm in length and 330 mm in width, and the size of the small plate is 125 mm in length and 230 mm in width. The large plate is attached to a large vehicle such as a truck, the middle plate is attached to an ordinary automobile, a light automobile, or the like, and the small plate is attached to a two-wheeled vehicle. In addition, each of the above-mentioned large plate, medium plate, and small plate has a different color scheme between a private car and a commercial car. The size, shape, or color of the taxable sign differs depending on each municipality, and the size of the taxable sign includes, for example, 100 mm in length and 170 mm in width.

In addition, the large plate, the middle plate, the small plate, and the taxation sign have different character arrangements, and the background color or the character color may be different.

Since the camera provided in the free-flow ETC system described above acquires an image of a moving vehicle, the shooting distance from the camera to the license plate of the vehicle is not constant. Therefore, the size of the license plate area on the image acquired by the camera is affected by the size of the license plate and the shooting distance. Therefore, in the image acquired by the camera, for example, the area of the image of the small plate having a short shooting distance and the area of the image of the large plate having a long shooting distance may be substantially equal. In this case, in the general method of identifying the license plate, for example, a large plate having a long shooting distance may be mistakenly recognized as a small plate. On the contrary, in the general method of identifying the license plate, a small plate having a short shooting distance may be mistakenly recognized as a large plate. As described above, in the general method of identifying the license plate, there is a high possibility that erroneous recognition will occur.

The present inventors may cut out a plurality of cutout images of different sizes using a plurality of cutout frames of different sizes from an image acquired by a camera, and include a number plate image from the cutout images. By extracting an identification image having a Hereinafter, the license plate identification device according to the position embodiment of the present invention will be described.

<2. Configuration of license plate identification system 1>
First, with reference to FIGS. 1 and 2, an example of the license plate identification system 1 to which the license plate identification device 10 according to the present embodiment can be applied will be described. FIG. 1 is a block diagram showing an example of the configuration of the license plate identification system 1. FIG. 2 is a schematic diagram showing an application example of the license plate identification system 1 according to the present embodiment.

The license plate identification system 1 identifies the type of license plate of a moving vehicle. The license plate identification system 1 includes a license plate identification device 10, a camera 20, and a network 30.

The license plate identification system to which the license plate identification device 10 according to the present embodiment is applied is not limited to the license plate identification system 1 described here.

Although the details will be described later, the license plate identification device 10 acquires an image captured by the camera 20 via the network 30, extracts an image of the license plate that can be included in the image, and corresponds to the image of the license plate. Identify the type of license plate.

The camera 20 photographs the vehicle. The camera 20 is provided at a position where the license plate of the road on which the vehicle travels can be imaged. The camera 20 may image the front surface of the vehicle or may image the rear surface of the vehicle. The camera 20 may be provided, for example, on a gate-shaped pillar installed across lanes as shown in FIG.

The camera 20 may be a known image pickup device having an image pickup function, such as a digital video camera, a digital still camera, or the like. Further, the camera 20 may be an infrared camera. A license plate with a design is scheduled to be distributed in the future, but since this design transmits infrared rays, the design will not be displayed in the captured image when the license plate with a design is imaged with an infrared camera. Therefore, by using an infrared camera for the camera 20, more accurate license plate identification becomes possible.

The network 30 is realized by using a known information and communication technique such as the Internet, and connects the license plate identification device 10 and the camera 20 to each other.

<3. Configuration of license plate identification device 10>
Next, the license plate identification device 10 according to the present embodiment will be described with reference to FIGS. 3 to 6. FIG. 3 is a block diagram showing an example of the configuration of the license plate identification device 10 according to the present embodiment. FIG. 4 is an explanatory diagram for explaining a cut-out image according to the present embodiment. FIG. 5 is a diagram showing a list of cutout images in FIG. FIG. 6 is an explanatory diagram for explaining a method of identifying a license plate by the identification unit 125 according to the present embodiment.

As shown in FIG. 3, the license plate identification device 10 according to the present embodiment includes an image acquisition unit 110, a processing unit 120, a storage unit 130, and a communication unit 140.

The image acquisition unit 110 receives the image captured by the camera 20. The acquired image, which is an image acquired by the image acquisition unit 110, may be one image or a plurality of images. The plurality of acquired images received by the image acquisition unit 110 may be, for example, a plurality of images continuously captured by the camera 20.

The processing unit 120 receives the acquired image from the image acquisition unit 110 and identifies the license plate that can be included in the acquired image. The processing unit 120 includes an image cutting unit 121, an image extraction unit 123, and an identification unit 125.

The image cutting unit 121 cuts out a plurality of cutout images having different sizes from one or a plurality of images by using a plurality of cutting frames having different sizes. For example, the image cutting unit 121 may cut out a predetermined number of cutout images having different sizes from one acquired image, or the size is different from a plurality of acquired images so that the total number is a predetermined number. Cutout An image may be cut out.

Further, the image cutting unit 121 can cut out, for example, a region of the acquired image having a high edge density.

As the cutting frame, for example, a cutting frame having a plurality of sizes according to the area of the license plate may be used. For the cutting frame, for example, a large frame that is a frame of a size corresponding to a large plate, a middle frame that is a frame of a size corresponding to a middle plate, and a small frame that is a frame of a size corresponding to a small plate are used. May be done.

As shown in FIG. 4, the image cutting unit 121 can cut out a part of the image captured by the camera 20 by using the large frame L, the middle frame M, and the small frame S, for example. In FIG. 4, the cutout image L1 to the cutout image L3 are cut out by the large frame L, the cutout image M1 to the cutout image M3 are cut out by the middle frame M, and the cutout image S1 to the cutout image S3 are cut out by the small frame S. ..

For example, when the type of the vehicle displayed in the acquired image can be specified in advance, the image cutting unit 121 may set a cutting frame according to the size of the license plate of the vehicle. Further, when the camera 20 is an acquired image obtained by capturing the front surface of the vehicle, the image cutting portion 121 does not need to cut out the image by the small frame S because the small plate S is attached to the rear portion of the motorcycle. Therefore, the image cutting portion 121 may omit the small frame S and cut out the image by the large frame L and the middle frame M. Further, the size of the cutout frame may be set by the user, for example.

The image extraction unit 123 extracts an identification image, which is an image that may include a region in which a license plate is displayed, from one or a plurality of images based on the positions of the plurality of cutout images. The image extraction unit 123 classifies the cutout image into two types, for example, an image including a license plate area and an image not including a license plate area (background image) by a classifier, and identifies which image the cutout image is. And the identification image may be extracted.

The image extraction unit 123 may calculate a score based on the certainty of the license plate for each of the cutout images, and select a predetermined number of cutout images from the plurality of cutout images in descending order of the score. Further, the image extraction unit 123 may extract an identification image of a predetermined size from the acquired image based on the positions of the selected predetermined number of cutout images. The identification image may be extracted according to each of the selected cutout images, or may be extracted from the acquired image so as to have a predetermined size with respect to the center of the cutout image, for example. Therefore, the quantity of the identification images is the same as the quantity of the selected cutout images, and the sizes of the identification images may be the same.

A classifier may be used to calculate the score. As the classifier, various machine learning algorithms may be applied, and for example, machine learning algorithms such as neural networks, random forests, SVMs (Support Vector Machines), and deep learning may be applied.

In the score calculation by the image extraction unit 123, for example, in FIG. 5, the cutout image M2 includes the entire license plate area in a size close to the area of the cutout frame. Therefore, the image extraction unit 123 may calculate the score of the cutout image M2 to the highest value. The cutout image L1, the cutout image L3, the cutout image M1, the cutout image M3, and the cutout image S1 do not include the license plate region. Therefore, the image extraction unit 123 may calculate the scores of these cutout images to a low value. Since the cutout image S2 includes only a part of the license plate area, the image extraction unit 123 may calculate the score of the cutout image S2 to a low value. Although the cutout image S3 includes the license plate region, the image is not clear, so the image extraction unit 123 may calculate the score of the cutout image S2 to a low value. The cutout image L2 includes the entire license plate area, but the ratio of the area of the license plate area to the size of the cutout frame is small. Therefore, the image extraction unit 123 may calculate the score of the cutout image L2 to be smaller than the score of the cutout image M2 and larger than the score of the cutout image M2.

When selecting two cutout images, the image extraction unit 123 can select the cutout image M2 having the highest score and the cutout image L2 having the next highest score after the cutout image M2 from the result of the score calculation. ..

The image extraction unit 123 may extract an identification image of a predetermined size from the acquired image, for example, based on the position of the cutout image M2 and the position of the cutout image L2. For example, as shown in FIG. 6, the image extraction unit 123 acquires an identification image based on the center of the cutout image L2 and an identification image based on the center of the cutout image M2 by the common frame C. It may be extracted from the image. With such a configuration, it can be expected that the information that can be used for identification will be increased when the identification unit 125 identifies the type of license plate. Specifically, assuming that the closer the license plate size is to the cutout frame, the higher the score, if an image with a high score on a small plate is cut out in the common frame C, the license plate in the image cut out in the common frame C will be smaller. However, if an image with a high score on a large plate is cut out in the common frame C, the license plate in the image cut out in the common frame C will appear large. As described above, the identification unit 125 can identify the type of license plate by utilizing the feature that the physical size of the license plate differs depending on the type. Further, for example, in the case of a small plate, since the peripheral area of the license plate is widely projected in the image cut out by the common frame C, for example, the driver's foot may be reflected in the cutout image by the common frame C. be. Since the size of the license plate is different between the truck and the passenger car, the size of the peripheral area of the license plate is significantly different, so that the identification unit 125 can utilize these characteristics.

A different classifier may be applied to the classifier applied to the image extraction unit 123 depending on the size of the cut-out image, or a common classifier may be applied. When a common classifier is applied to the image extraction unit 123, the cutout images of different sizes may be changed to the same size and the cutout images may be classified by the classifier.

The image extraction unit 123 makes it possible to identify the license plate in the identification unit 125, which will be described later, with higher accuracy.

The identification unit 125 identifies the type of license plate based on the identification image. The identification unit 125 performs image processing on the identification image and identifies the type of license plate from the arrangement of characters displayed on the identification image.

The identification unit 125 classifies the identification image by a classifier, for example, and identifies the type of license plate. Various machine learning algorithms can be applied to the classifier of the identification unit 125. Machine learning algorithms such as neural networks, random forests, k-nearest neighbors, and deep learning may be applied to the classifier of the discriminator 125.

Further, the identification unit 125 may use information about a cutout frame corresponding to the identification image as one of the input information to the classifier. The information regarding the cutout frame corresponding to the identification image can improve the identification accuracy, and examples thereof include the area of the cutout frame corresponding to the identification image, the position of the identification image in the input image, and the like. For example, when a neural network is applied as a machine learning algorithm, the identification unit 125 may change the weight in the neural network according to the information about the cutout frame corresponding to the identification image. By using the information related to the cutout frame as one of the input information to the classifier, it is possible to improve the accuracy of various identification in the identification unit 125.

The identification unit 125 may classify the identification image into at least three types of groups of two types of license plates and images that do not include the area where the license plates are displayed, and identify the license plates. For example, the identification unit 125 may classify the identification image into three types of groups of a large plate, a middle plate, and a background image that does not include a license plate area, and identify the license plate. Further, the identification unit 125 may classify the identification image into five groups of a large plate, a middle plate, two types of small plates, and a background image to identify the license plate, or may identify the license plate. And business vehicles may be distinguished and classified into 9 types of groups to identify license plates. The identification unit 125 may further classify the identification image into 10 types of groups in which the taxation sign is added to the above 9 types for identification. Even when the image extraction unit 123 extracts the background image as the identification image, the identification unit 125 can classify the identification image into the background image, so that the type of license plate can be selected with high accuracy. It is possible to identify. When the identification image is classified into nine types of groups for identification, the identification unit 125 may identify the identification image based on, for example, the arrangement of characters and the brightness of the image.

The identification unit 125 calculates a score based on the certainty of the license plates classified for the identification image. The identification unit 125 can use the group to which the identification image having the highest calculated score belongs as the identification result. Thereby, the type of the license plate corresponding to the license plate region included in the image captured by the camera 20 is identified.

The storage unit 130 is an example of a storage device included in the license plate identification device 10 according to the present embodiment. Various programs, databases, and the like used by the license plate identification device 10 when performing various processes as described above are appropriately recorded in the storage unit 130. The storage unit 130 may record an acquired image received by the image acquisition unit 110, a cutout image cut out by the image cutting unit 121, or the like. Further, the storage unit 130 may record a classifier applied to the image extraction unit 123 or a classifier applied to the identification unit 125. Further, the storage unit 130 contains various information such as data for associating the acquired image with the cut-out image, coordinate values of the cut-out image, and a score output by the classifier of the image extraction unit 123 or the classifier of the identification unit 125. May be recorded as historical information. In the storage unit 130, for example, various parameters that each of the image cutting unit 121, the image extraction unit 123, and the identification unit 125 need to save when performing each process, the progress of the process, and the like are appropriately recorded. You may. Not limited to the processing executed by the image cutting unit 121, the image extraction unit 123, and the identification unit 125, various parameters and processes that need to be saved when the license plate identification device 10 according to the present embodiment performs some processing. The progress etc. may be recorded as appropriate. In the storage unit 130, the image acquisition unit 110, the image cutting unit 121, the image extraction unit 123, and the identification unit 125 can freely perform read / write processing.

The communication unit 140 performs information communication with the camera 20 via the network 30. Specifically, the communication unit 140 receives an image from the camera 20. Further, the communication unit 140 may transmit a control signal to the camera 20 to control its drive.

According to the license plate identification device 10 according to the present embodiment, it is possible to more accurately identify the type of license plate even when the shooting distance changes.

Further, since the license plate identification device 10 identifies the type of license plate, if the type of license plate is identified, it is possible to acquire information regarding the arrangement of characters on the license plate. Therefore, for example, when the process of acquiring the detailed information of the license plate is performed after the type of the license plate is identified, the image analysis may be performed on the portion corresponding to the position of the character on the license plate. As a result, it becomes possible to analyze the number information with higher accuracy and higher speed. Up to this point, the configuration of the license plate identification device 10 has been described in detail.

<4. Operation of license plate identification device 10>
Subsequently, the operation of the license plate identification device 10 according to the present embodiment will be described with reference to FIGS. 7 and 8. FIG. 7 is a flow chart showing the flow of license plate identification by the license plate identification device according to the present embodiment. FIG. 8 is an explanatory diagram illustrating a license plate identification method by the license plate identification device according to the present embodiment.

First, as shown in FIG. 7, the image acquisition unit 110 receives the acquired image captured by the camera 20 via the communication unit 140 (step S101). Subsequently, the image cutting unit 121 receives the acquired image from the image acquisition unit 110, and cuts out a plurality of cutout images having different sizes from one or two or more acquired images by a plurality of cutting frames having different sizes (step). S103).

Subsequently, the image extraction unit 123 acquires the cut-out image from the image cut-out unit 121, classifies the cut-out image into two types, an image including the license plate area and an image not including the license plate area, and cuts out the cut-out image. Identifying which image the image is (step S105). The image extraction unit 123 selects a predetermined number of cutout images from the cutout images classified into images including the license plate area (step S107), and extracts a predetermined number of identification images based on the positions of the selected cutout images. (Step S109).

Next, the identification unit 125 performs image processing on the identification image to identify the type of license plate (step S111).

The license plate identification device 10 according to the present embodiment can identify the type of license plate displayed in the image captured by the camera 20.

Further, as described above, the image cutting unit 121 cuts out a plurality of images having different sizes by a plurality of cutting frames having different sizes. The image extraction unit 123 identifies whether the cutout image cut out by each of the plurality of cutout frames is an image or a background image that may include a license plate area. As shown in FIG. 8, the image extraction unit 123 identifies and identifies each of the image cut out by the large frame L, the image cut out by the middle frame M, and the image cut out by the small frame S, and a predetermined number of them. Select the cutout image of. The image extraction unit 123 extracts an identification image from the selected cutout image by a common frame C having a predetermined size. Then, the identification unit 125 identifies the identification image. As described above, the license plate identification device 10 according to the present embodiment substantially identifies the license plate in two stages. Since the identification unit 125 does not perform various types of discrimination on the cutout image classified as the background image by the two types of the image extraction unit 123, the identification accuracy in the identification unit 125 is improved. Further, the calculation cost of the two-kind identification performed by the image extraction unit 123 is smaller than the calculation cost of the various types identification performed by the identification unit 125. Therefore, by identifying the type of license plate in two steps, it is possible to reduce the number of operations for various identification, which has a large calculation cost, and it is possible to shorten the time required for identifying the license plate.

<5. Hardware configuration>
The embodiment of the present invention has been described above. The various processes described above are realized by the cooperation between the software and the hardware of the license plate identification device 10 described below.

FIG. 9 is a block diagram showing a hardware configuration of the license plate identification device 10 according to the present embodiment. The number plate identification device 10 includes a CPU (Central Processing Unit) 901, a ROM (Read Only Memory) 902, a RAM (Random Access Memory) 903, a host bus 904, a bridge 905, an external bus 906, a storage device (HDD) 911, and a storage device (HDD) 911. A network interface 913 is provided. Further, the license plate identification device 10 may include an interface 907, an input device 908, a display device 909, an audio output device 910, and a drive 912.

The CPU 901 functions as an arithmetic processing device and a control device, and controls the overall operation of the license plate identification device 10 according to various programs. Further, the CPU 901 may be a microprocessor. The ROM 902 stores programs, calculation parameters, and the like used by the CPU 901. The RAM 903 temporarily stores a program used in the execution of the CPU 901, parameters that appropriately change in the execution, and the like. These are connected to each other by a host bus 904 composed of a CPU bus or the like. By the cooperation between the CPU 901, the ROM 902 and the RAM 903 and the software, for example, the function of the image acquisition unit 110 or the processing unit 120 can be realized.

The host bus 904 is connected to an external bus 906 such as a PCI (Peripheral Component Interconnect / Interface) bus via a bridge 905. It is not always necessary to separately configure the host bus 904, the bridge 905, and the external bus 906, and these functions may be implemented in one bus.

The input device 908 includes input means for the user to input information such as a mouse, keyboard, touch panel, buttons, microphone, sensor, switch and lever, and an input signal generated based on the input by the user and output to the CPU 901. It may be composed of a control circuit or the like.

The display device 909 may include, for example, a display device such as a CRT (Cathode Ray Tube) display device, a liquid crystal display (LCD) device, a projector device, and an OLED (Organic Light Emitting Diode) device. Further, the audio output device 910 may include an audio output device such as a speaker and headphones, and may output audio acquired by the camera 20.

The storage device 911 is a data storage device configured as an example of the storage unit 130 of the license plate identification device 10. The storage device 911 may include a storage medium, a recording device for recording data on the storage medium, a reading device for reading data from the storage medium, a deleting device for deleting data recorded on the storage medium, and the like. The storage device 911 is composed of, for example, an HDD (Hard Disk Drive), an SSD (Solid Storage Drive), or a memory having an equivalent function. The storage device 911 drives the storage and stores programs and various data executed by the CPU 901.

The drive 912 is a reader / writer for a storage medium, and may be built in or externally attached to the license plate identification device 10. The drive 912 may read the information recorded in the removable storage medium 914 such as the mounted magnetic disk, optical disk, magneto-optical disk, or semiconductor memory and output it to the RAM 903 or the storage device 911. The drive 912 can also write information to the removable storage medium 914.

The network interface 913 is, for example, a communication interface configured by a communication device or the like for connecting to the network 30. Further, the network interface 913 may be a wireless LAN (Local Area Network) compatible terminal or a wire terminal that performs wired communication.

<6. Conclusion>
As described above, according to the embodiment of the present invention, it is possible to more accurately identify the type of license plate even when the shooting distance changes.

A preferred embodiment of the present invention has been described in detail with reference to the accompanying drawings, but the present invention is not limited to this example. It is clear that a person having ordinary knowledge in the field of the art to which the present invention belongs can come up with various modifications or modifications within the scope of the technical idea described in the claims. , These are also naturally understood to belong to the technical scope of the present invention.

Further, it is possible to create a computer program for causing the hardware such as the CPU, ROM, and RAM built in the license plate identification device to exhibit the same functions as each configuration of the license plate identification device 10 described above. A storage medium for storing the computer program is also provided.

10 License plate identification device 110 Image acquisition unit 120 Processing unit 121 Image cutting unit 123 Image extraction unit 125 Identification unit 130 Storage unit 140 Communication unit

Claims (8)

An image cutout unit that acquires a cutout image that is an image cut out from one or more images, and an image cutout portion.
An image extraction unit that extracts an identification image, which is an image that may include a region in which a license plate is displayed from the one or a plurality of images based on the positions of the plurality of cutout images, and an image extraction unit.
An identification unit that identifies the type of license plate based on the identification image,
Equipped with
The image cutting unit is a license plate identification device that cuts out a plurality of cutout images having different sizes from the one or a plurality of images by means of a plurality of cutting frames having different sizes. The image extraction unit calculates a score based on the certainty of the license plate for each of the cutout images, and selects a predetermined number of cutout images from a plurality of the cutout images in descending order of the score. The license plate identification device according to 1. The license plate identification device according to claim 2, wherein the image extraction unit extracts the identification image of a predetermined size from the one or a plurality of images based on the positions of the predetermined number of cutout images. The license plate identification device according to any one of claims 1 to 3, wherein the cutout frame has a size corresponding to the area of the license plate. The license plate identification device according to any one of claims 1 to 4, wherein the identification unit uses information about the cutout frame corresponding to the identification image as one of input information to the classifier. The identification unit identifies the license plate by classifying the identification image into at least one of three types, two types of the license plate and an image not including the area where the license plate is displayed. , The license plate identification device according to any one of claims 1 to 5. Acquiring a cut-out image that is a cut-out image from one or more images,
Extracting an identification image, which is an image that may include a license plate image, based on the positions of the plurality of cutout images.
Including identifying the license plate from the identification image,
In the acquisition of the cutout image, a license plate identification method in which a plurality of cutout images having different sizes are cut out from the one or a plurality of images by a plurality of cutout frames having different sizes. Computer,
An image cutout unit that acquires a cutout image that is an image cut out from one or more images, and an image cutout portion.
An image extraction unit that extracts an identification image, which is an image that may include a license plate image, based on the positions of the plurality of cutout images.
It functions as an identification unit that identifies the license plate from the identification image.
The image cutting portion is a program for cutting out a plurality of cut-out images having different sizes from the one or a plurality of images by a plurality of cutting frames having different sizes.

Images