Passage determination device, passage determination system, and passage determination method

Abstract

To distinguish trailers traveling in close proximity with high accuracy.SOLUTION: A passage determination device acquires a first captured image in which an area including a passage determination position of a trailer is repeatedly captured from the front in the traveling direction of the trailer. The passage determination device also performs a process to detect the trailer in the first captured image. The passage determination device is provided with a processor that determines that the trailer detected in the first captured image taken at the first time has passed the passage determination position when the trailer is detected in the first captured image taken at the first time and when the trailer is not detected in a first captured image taken within a first hour from a first time.SELECTED DRAWING: Figure 10

Description

[Technical field]

[0001] The present disclosure relates to a passage determination device, a passage determination system, and a passage determination method. [Background technology]

[0002] Patent Document 1 discloses an identical vehicle determination device that does not require calibration to adjust the installation distance of the front and rear cameras by performing accurate relative position measurement. The identical vehicle determination device includes a detection distance estimation unit that estimates the detection distance between the front camera and the rear camera based on the speed and passing time of a vehicle with a matching number calculated by a vehicle speed calculation unit. The identical vehicle determination device further includes an identical vehicle determination unit that calculates an estimated passing time of the vehicle based on the speed of the vehicle calculated by the vehicle speed calculation unit and the detection distance between the front camera and the rear camera estimated by the detection distance estimation unit when the front license plate number recognized by the first recognition unit and the rear license plate number recognized by the second recognition unit do not match. The identical vehicle determination unit determines that the rear license plate number recognized by the second recognition unit within the estimated passing time is the rear license plate number of the vehicle. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2020-154959 A Summary of the Invention [Problem to be solved by the invention]

[0004] In recent years, logistics using trailers has become popular, and there is an increasing need to efficiently manage the location of trailers. Here, a trailer means a towed vehicle towed by a towing vehicle (also called a tractor). A plurality of trailers may be gathered at a warehouse or at a terminal or the like located in various places (for example, a starting point, a destination point, or a relay point). Examples of the places where trailers gather include a distribution center where cargo to each delivery destination is loaded onto trailers, and a trailer station used as a temporary storage place for trailers. In the following, a distribution center is used as an example, but the places where trailers gather may be other facilities such as a trailer station. Since a large number of trailers gather at a distribution center, there is a need to automatically manage trailers entering and leaving the distribution center.

[0005] In Patent Document 1, the same vehicle determination device determines that the vehicles photographed by the front and rear cameras are the same vehicle when the number on the front license plate of the vehicle read by the front camera matches the number on the rear license plate of the vehicle read by the rear camera. However, when multiple vehicles are passing close to each other, there is a possibility that the separation between the vehicles will not be distinguished and the multiple vehicles will be erroneously determined to be one vehicle.

[0006] The present disclosure has been devised in view of the above-mentioned conventional situation, and has an object to distinguish trailers traveling in close proximity with high accuracy. [Means for solving the problem]

[0007] The present disclosure provides a passing determination device including a processor that obtains a first captured image by repeatedly capturing an area including a passing determination position of the trailer from ahead in the direction of travel of the trailer, performs processing to detect the trailer in the first captured image, and determines that the trailer detected from the first captured image taken at the first time has passed the passing determination position if the trailer is detected from the first captured image taken at the first time but not detected from the first captured image taken within one hour from the first time.

[0008] The present disclosure also provides a passing judgment system comprising a front camera that repeatedly captures an area including a passing judgment position of a trailer from ahead in the direction of travel of the trailer, and a processor that acquires a first captured image from the front camera, wherein the processor performs processing to detect the trailer in the first captured image, and if the trailer is detected from the first captured image taken at a first time, but not detected from the first captured image taken within one hour from the first time, it determines that the trailer detected from the first captured image taken at the first time has passed the passing judgment position.

[0009] The present disclosure also provides a passing determination method, which obtains a first captured image by repeatedly capturing an area including a passing determination position of the trailer from ahead in the direction of travel of the trailer, performs processing to detect the trailer in the first captured image, and if the trailer is detected in the first captured image captured at a first time but not detected in the first captured image captured within one hour from the first time, determines that the trailer detected in the first captured image captured at the first time has passed the passing determination position.

[0010] The present disclosure also provides a passing judgment device including a processor that acquires captured images from a camera that photographs a trailer at predetermined time intervals, sets coordinates in the captured images in a direction corresponding to the direction of travel of the trailer, calculates a first coordinate of the trailer, and judges that the trailer has passed a passing judgment position if the amount of change between the calculated first coordinate and a second coordinate of the trailer calculated from the captured image acquired immediately before is equal to or greater than a predetermined threshold.

[0011] The present disclosure also provides a passing judgment system comprising a camera that photographs a trailer and a processor that acquires the photographed image from the camera, wherein the processor acquires the photographed image from the camera at predetermined time intervals, sets coordinates in the photographed image parallel to the traveling direction of the trailer, calculates a first coordinate of the trailer, and determines that the trailer has passed a passing judgment position if the amount of change between the calculated first coordinate and a second coordinate of the trailer calculated from the photographed image acquired immediately before is greater than or equal to a predetermined threshold.

[0012] The present disclosure also provides a passing determination method, which acquires captured images from a camera that photographs a trailer at predetermined time intervals, sets coordinates in the captured images parallel to the direction of travel of the trailer, calculates a first coordinate of the trailer, and determines that the trailer has passed a passing determination position if the amount of change between the calculated first coordinate and a second coordinate of the trailer calculated from the captured image acquired immediately before is equal to or greater than a predetermined threshold.

[0013] These comprehensive or specific aspects may be realized as a system, an apparatus, a method, an integrated circuit, a computer program, or a recording medium, or may be realized as any combination of a system, an apparatus, a method, an integrated circuit, a computer program, and a recording medium. Effect of the Invention

[0014] According to the present disclosure, trailers traveling in close proximity can be distinguished with high accuracy. [Brief description of the drawings]

[0015] [Figure 1] A diagram showing the text included in the trailer [Diagram 2] Block diagram of a character recognition system according to a first embodiment. [Diagram 3] Flowchart of processing executed by the character recognition device [Figure 4] A diagram showing attributes that are presumed to be trailer IDs and attributes that are presumed not to be trailer IDs. [Diagram 5] Flowchart for classifying whether a string is likely to be a trailer ID [Figure 6] An example of how vehicles entering and leaving a distribution center can be photographed. [Figure 7] Block diagram of a passage determination device according to a second embodiment and a third embodiment. [Figure 8] FIG. 1 is a diagram showing an example of time-series data of images captured by a front camera and a rear camera; [Figure 9] A diagram showing the size of the trailer detection frame when multiple vehicles are photographed. [Figure 10] A diagram showing several cases of trailer passing judgment. [Figure 11] A diagram showing the first case where the trailer is not detected for more than a second period of time. [Figure 12] FIG. 2 shows a second case where the trailer is not detected for a second period of time. [Figure 13] A flowchart showing an example of a process for determining whether a trailer is passing. [Figure 14] A diagram showing an example of time-series data of images of two trailers passing close to each other [Figure 15] FIG. 13 shows a passage determination method according to the third embodiment. [Figure 16] A flowchart showing an example of a trailer passage determination process according to the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] Hereinafter, with reference to the drawings as appropriate, an embodiment specifically disclosing a passage determination device, a passage determination system, and a passage determination method according to the present disclosure will be described in detail. However, more detailed explanation than necessary may be omitted. For example, detailed explanation of already well-known matters and duplicate explanation of substantially the same configuration may be omitted. This is to avoid the following explanation becoming unnecessarily redundant and to facilitate understanding by those skilled in the art. Note that the attached drawings and the following explanation are provided to enable those skilled in the art to fully understand the present disclosure, and are not intended to limit the subject matter described in the claims.

[0017] <Background to the first embodiment> Patent Document 2, relating to Japanese Patent Laid-Open Publication No. 2008-217347, discloses a license plate recognition device that detects multiple quadrangles of license plate area candidates from an input image and performs character recognition on character areas included in the license plate area candidates. The license plate recognition device selects a license plate area candidate to be output from the multiple detected license plate area candidates based on the character recognition results and quadrangle information for each license plate area candidate. The license plate recognition device outputs information about the selected license plate area candidate.

[0018] In recent years, logistics using trailers has become popular, and there is an increasing need to efficiently manage the location of each trailer. Here, a trailer means a towed vehicle towed by a towing vehicle (also called a tractor). A trailer may be loaded with a container for storing various items. Generally, an identification (hereinafter referred to as "ID") related to the identification of the trailer or container is written on a trailer or container, and it is considered that the trailer or container can be efficiently managed by recognizing this ID from a photographed image of a vehicle or the like. However, a photographed image may contain various characters, and a character string that is successfully recognized is not necessarily an ID related to the identification of the trailer or container. Therefore, when a trailer or container, which is an object to be photographed, contains multiple character strings of various sizes, colors, and designs, there is a need to detect an ID related to the identification of the trailer or container from the multiple character strings and identify the object to be photographed. Hereinafter, an ID related to the identification of a trailer is referred to as a "trailer ID" and an ID related to the identification of a container is referred to as a "container ID."

[0019] In Patent Document 2, a quadrilateral that satisfies a predetermined condition is determined to be a quadrilateral that is a candidate for a license plate area. However, since there is currently no unified standard for trailer IDs, the position, size, number of characters, etc. are not fixed. Therefore, even if a technology similar to Patent Document 1 is applied, it is difficult to detect the trailer ID from among various character strings written at any position on the photographed object. In addition, there are multiple standards for container IDs, and it is difficult to detect them from among various character strings, just like trailer IDs.

[0020] In the following first embodiment, an example of a character recognition device and a character recognition method for detecting a character string that is considered to be a trailer ID or a container ID from among a plurality of character strings will be described.

[0021] <Embodiment 1> The character strings included in the trailer TR will be described with reference to Fig. 1. Fig. 1 is a diagram showing the character strings included in the trailer TR.

[0022] The trailer TR is the whole object towed by the tractor LE. Generally, a vehicle towed by a tractor is called a trailer. The trailer may be configured as a set of a container that stores cargo but does not have a mechanism for moving such as wheels, and a dolly on which the container is loaded. The trailer may also be configured as a vehicle in which a part for storing cargo and a mechanism for moving are integrated. When the tractor LE tows a set of a container and a dolly, only the dolly may be called a trailer. However, in this specification, for the sake of convenience, the whole object towed by the tractor LE, that is, both the set of the dolly and the container, or the vehicle having a part for storing cargo and a mechanism for moving, are called trailers. In addition to the trailer ID, the trailer TR contains a company name, a telephone number, a number indicating the height of the trailer, a uniform resource locator (hereinafter referred to as "URL"), and a character string related to words or text. The character strings written in the trailer TR are not limited to the above-mentioned examples.

[0023] The character strings ST1, ST2, ST3, and ST4 are character strings indicating company names. The character strings ST1, ST2, ST3, and ST4 are character strings that are decorated with italics.

[0024] The character string ST5 is a character string expressed in English words and is written below the character string ST2 indicating the company name.

[0025] The character strings ST6 and ST7 are character strings related to a number for identifying the tractor LE.

[0026] The character string ST8 is a character string indicating the trailer ID. The character string ST8 is written vertically on a corner of the trailer TR (that is, a boundary part of the surface forming the casing of the trailer TR, which is located on the side of the trailer TR and is long in the direction of gravity).

[0027] In this way, the trailer TR contains multiple character strings.

[0028] In recent years, as the logistics industry has become more active, there is an increasing need to efficiently manage a large number of trailers TR. For example, at a location where multiple trailers TR are gathered, it is necessary to identify each trailer TR. However, since trailer TRs have multiple character strings (e.g., character strings ST1, ST2, ST5, ST8), there is a problem that it is difficult to detect the trailer ID from among the multiple character strings.

[0029] Next, a block diagram of the character recognition system 1 according to the first embodiment will be described with reference to Fig. 2. Fig. 2 is a block diagram of the character recognition system 1 according to the first embodiment.

[0030] The character recognition system 1 includes at least a photographing device 10, a database 20, a display device 30, and a character recognition device 40.

[0031] The photographing device 10 is a device that photographs a trailer. The photographing device 10 is configured to have at least a lens (not shown) and an image sensor (not shown) as optical elements. The lens receives light reflected by an object within the angle of view of the area photographed by the photographing device 10, and forms an optical image of the object on the light receiving surface of the image sensor (in other words, the photographing surface). The image sensor is a solid-state photographing element such as a Charged Coupled Device (hereinafter referred to as "CCD") or a Complementary Metal Oxide Semiconductor (hereinafter referred to as "CMOS"). The image sensor converts the optical image formed on the photographing surface through the lens into an electrical signal every predetermined time (for example, 1 / 30 (second)). For example, when the predetermined time is 1 / 30 (second), the frame rate of the photographing device 10 is 30 fps. The photographing device 10 may also generate image data (video data) by performing predetermined signal processing on the electrical signal every predetermined time. Note that the image data is a still image, and the video data is a video. The photographing device 10 outputs image data (video data) to the character recognition device 40. Hereinafter, the image data and video data will be referred to as a photographed image.

[0032] The database 20 is configured using a storage medium (e.g., a flash memory, a hard disk drive (hereinafter referred to as an "HDD"), or a solid state drive (hereinafter referred to as an "SSD")). The database 20 stores a table (see FIG. 4 and hereinafter referred to as the classification table) that classifies whether or not the attributes of a character string are estimated to be a trailer ID. The database 20 also stores the results of a determination made by the character recognition device 40 as to whether or not a character string is a trailer ID. The database 20 may be incorporated into the character recognition device 40.

[0033] The display device 30 displays the result of the determination as to whether or not the character string executed by the character recognition device 40 is a trailer ID. The display device 30 is, for example, a touch panel display, a display, a mobile phone, etc. The display device 30 may be configured integrally with the character recognition device 40.

[0034] The character recognition device 40 includes at least a processor 41, a communication I / F 42, and a memory 43. The character recognition device 40 is, for example, a personal computer (hereinafter, referred to as a "PC").

[0035] The processor 41 is, for example, a central processing unit (hereinafter referred to as "CPU"), a digital signal processor (hereinafter referred to as "DSP"), a graphical processing unit (hereinafter referred to as "GPU"), or a field programmable gate array (hereinafter referred to as "FPGA"). The processor 41 functions as a controller that manages the overall operation of the character recognition device 40. The processor 41 performs control processing for managing the operation of each part of the character recognition device 40, input / output processing of data between each part of the character recognition device 40, calculation processing of data, and storage processing of data. The processor 41 operates according to a program stored in the memory 43. The processor 41 uses the memory 43 during operation, and temporarily stores data generated or acquired by the processor 41 in the memory 43. The processor 41 realizes the functions of the character string detection unit 41A, the character string recognition unit 41B, the attribute determination unit 41C, and the trailer ID estimation unit 41D by using the programs and data stored in the memory 43.

[0036] The character string detection unit 41A, which is an example of a detection unit, detects a character string from a captured image acquired from the imaging device 10. The character string detection unit 41A detects a character string using a known technique such as machine learning.

[0037] The character string recognizing unit 41B, which is an example of a recognition unit, performs character recognition on the character string detected by the character string detecting unit 41 A. The character string recognizing unit 41B performs character recognition on the character string using a known technique such as Optical Character Recognition (hereinafter referred to as "OCR").

[0038] Attribute determination unit 41C, which is an example of a determination unit, determines the attribute of the character string recognized by character string detection unit 41A. The attribute is, for example, the writing style of the characters, the position of the character string, or the content indicated by the character string. The attribute will be described later.

[0039] The trailer ID estimation unit 41D, which is an example of a determination unit, estimates whether each detected and recognized character string is a trailer ID based on the attributes. The trailer ID estimation unit 41D calculates the likelihood that the character string is a trailer ID as a score based on the attributes. The trailer ID estimation unit 41D classifies the character strings into those estimated to be trailer IDs, those estimated not to be trailer IDs, those that are likely to be trailer IDs based on the calculated scores, and those that are unlikely to be trailer IDs based on the calculated scores. The trailer ID estimation unit 41D sorts (i.e., rearranges) the classified character strings in order of likelihood that they are trailer IDs (for example, in descending order of score). The trailer ID estimation unit 41D outputs the estimation result, the result of classifying or sorting the character strings, to the communication I / F 42.

[0040] The communication I / F 42 is an interface circuit that performs wireless or wired communication between the character recognition device 40 and the photographing device 10, between the character recognition device 40 and the database 20, and between the character recognition device 40 and the display device 30. Here, I / F refers to an interface. The communication between the character recognition device 40 and the photographing device 10, between the character recognition device 40 and the database 20, and between the character recognition device 40 and the display device 30 may be via a network. The communication method by the communication I / F 42 is, for example, a Wide Area Network (hereinafter referred to as "WAN"), a Local Area Network (hereinafter referred to as "LAN"), Long Term Evolution (hereinafter referred to as "LTE"), mobile communication such as 4G and 5G, power line communication, short-range wireless communication (e.g., Bluetooth (registered trademark) communication), or communication for mobile phones.

[0041] Memory 43 is configured using, for example, Random Access Memory (hereinafter referred to as "RAM") and Read Only Memory (hereinafter referred to as "ROM"), and temporarily stores programs necessary for the operation of character recognition device 40, as well as data generated during operation. The RAM is, for example, a work memory used during operation of character recognition device 40. The ROM stores and holds, for example, programs for controlling character recognition device 40 in advance.

[0042] Next, a flowchart of the process executed by the character recognition device 40 will be described with reference to Fig. 3. Fig. 3 is a flowchart of the process executed by the character recognition device 40. Each process of the flowchart in Fig. 3 is executed by the processor 41.

[0043] The processor 41 acquires a captured image from the imaging device 10 (step St100).

[0044] The processor 41 detects a character string from the captured image acquired in the process of step St100 (step St101).

[0045] The processor 41 performs character recognition on the character string detected in the process of step St101 (step St102).

[0046] The processor 41 determines the attribute of the character string recognized in the process of step St102 (step St103).

[0047] The processor 41 reads the classification table stored in the database 20 (step St104). Note that the process of step St104 may be omitted.

[0048] The processor 41 estimates whether the character string is a trailer ID based on the attribute determined in the process of step St103. The processor 41 estimates whether each of the multiple character strings detected in the process of step St101 is a trailer ID, and classifies the character strings into those estimated to be trailer IDs and those estimated not to be trailer IDs (step St105). The process of step St105 will be described in detail with reference to FIG. 5.

[0049] The processor 41 outputs and writes the character string classified in the process of step St105 to the database 20 (step St106).

[0050] The processor 41 outputs the result of step St105 to the display device 30 (step St107). Note that the process of step St107 may be omitted.

[0051] Next, an attribute presumed to be a trailer ID and an attribute presumed not to be a trailer ID will be described with reference to Fig. 4. Fig. 4 is a diagram showing an attribute presumed to be a trailer ID and an attribute presumed not to be a trailer ID.

[0052] The classification table shown in FIG. 4 is a table in which each attribute is classified according to whether it is an attribute presumed to be a trailer ID or an attribute presumed not to be a trailer ID.

[0053] It will be explained whether each attribute related to "character writing style" is presumed to be a trailer ID or not.

[0054] The fact that a character string is written vertically is one of the attributes that suggests it is a trailer ID. In general, it is unlikely that information such as a company name, telephone number, or URL would be written vertically, and license plate characters are also required to be written horizontally. Therefore, it is unlikely that characters other than the trailer ID would be written vertically, and vertically written character strings are highly likely to be trailer IDs.

[0055] Characters in a string that are the same size, evenly spaced, and in a single color or Gothic font are among the attributes that indicate that the string is a trailer ID. Trailer IDs, by their very nature, are unlikely to be used for promotional purposes, so they are unlikely to have special designs and are often written in a simple format. Therefore, the fact that a string is written in such a simple format as described above can be a basis for indicating that the string is a trailer ID.

[0056] The fact that a character string is written inside a rectangle is one of the attributes that indicates that it is a trailer ID. If the color of the trailer and the trailer ID are similar, measures may be taken to improve visibility. As a measure, a rectangle of a different color than the color of the character string may be drawn as the background of the trailer ID. Therefore, the fact that a character string is written inside a rectangle can be a basis for assuming that the character string is a trailer ID. Other measures, such as bordering the characters, are also considered, but these are unlikely to be adopted for trailer IDs, which place more importance on visibility than design.

[0057] A character string larger than a predetermined size determined in advance by a user (e.g., a person who manages the character recognition device 40) is one of the attributes that indicates that the character string is a trailer ID. This is because trailer IDs are often written large with an emphasis on visibility.

[0058] The fact that the character string is a certain magnification to the trailer is one of the attributes that suggests it is a trailer ID. This is because trailer IDs are often written large for ease of visibility.

[0059] A string in a decorative or italic font is one attribute that indicates it is not a trailer ID. A trailer ID is a string written to make the ID itself visible, so there is no need to use an unnecessarily complicated font.

[0060] Next, it will be explained whether each attribute related to "character position" is presumed to be a trailer ID or not.

[0061] The fact that the character string is located above the license plate is one of the attributes that suggests that it is a trailer ID. This is because the trailer ID is unlikely to be written at a lower position than the license plate, which is generally placed at the bottom of the vehicle body. Note that there are various known methods for detecting the position of the license plate, so details are omitted.

[0062] A character string written vertically in a corner of a trailer is one of the attributes that indicates that the character string is a trailer ID. The corner of a trailer is a place that is difficult to see unless you pay attention to it, so it is unlikely to be selected as a position for writing information that you want to publicly transmit, such as a company name or a phone number. Therefore, the character string written in this place is likely to be a character string used by a company for a special purpose, such as a trailer ID. As described above, a vertically written character string is likely to be a trailer ID. In this embodiment, a character string written vertically in a corner is used as an example of an attribute, but a character string written in a corner may be used as an attribute that indicates that the character string is a trailer ID, regardless of the direction of the character string. As described above, the fact that the character string is written in a corner itself suggests that the character string is for a special purpose.

[0063] The fact that a character string is written on a tractor LE is one of the attributes that suggests that it is not a trailer ID. Because trailers can be moved by changing the tractor, it is unlikely that a trailer ID, which is a character string that identifies a trailer, would be written on a tractor.

[0064] Characters written above or below a company logo, mark, or other characters are one of the attributes that indicate that it is not a trailer ID. Characters written around a company logo, etc. are likely to be information that should be sent to the public, such as a phone number or advertising slogan, and are unlikely to contain information for businesses, such as a trailer ID.

[0065] Next, it will be explained whether each attribute related to the "content indicated by the character string" is presumed to be a trailer ID or not.

[0066] The fact that a character string consists only of capital letters and numbers is one of the attributes that indicates that the character string is a trailer ID. If a character string that contains a mixture of easily confused characters such as capital letters and lowercase letters is used as an ID, multiple IDs with the same pronunciation will exist, which will cause confusion, so the type of characters used as an ID is generally limited. Here, capital letters are often used from the viewpoint of visibility, and numbers are generally also used in combination to generate more IDs, so that the fact that a character string consists only of capital letters and numbers can be a basis for indicating that the character string is a trailer ID. In the first embodiment, a character string consisting only of capital letters and numbers is one of the attributes that indicates that the character string is a trailer ID. However, if it is known that a character string consisting only of capital letters, lowercase letters, or numbers tends to be used as a trailer ID, the fact that the character string consists of that character type may be one of the attributes that indicates that the character string is a trailer ID. In addition, the fact that a character string does not contain a character type that is used only in some languages, such as Chinese characters or Hangul, may be one of the attributes that indicates that the character string is a trailer ID. This is because, particularly when transporting globally, it is unlikely that characters that can only be read in a limited number of languages ​​will be used as a trailer ID.

[0067] One of the attributes that indicates that a trailer ID is likely to be a trailer ID is that the number of characters is within a certain range that is determined in advance by the user. This is because, unlike text, trailer IDs often have a limited number of characters.

[0068] One of the attributes that suggests a character string is a trailer ID is that it conforms to specific company rules. If the company rules are known, then a character string that conforms to those rules is likely to be a trailer ID.

[0069] The fact that a character string is a container ID is one of the attributes that indicates that the character string is not a trailer ID. A container ID is a code that is written according to a predetermined rule and is used to manage and identify a container. In the first embodiment, the container ID is assumed to be a code of the international standard defined by ISO6346. A container ID is an ID used to manage a container, and is a character string for which a unified standard has been established among businesses. If each business adopts a trailer ID that conforms to this standard as its own trailer ID, it will cause problems in managing containers among businesses, so it is unlikely that a trailer ID that conforms to the container ID standard will be adopted. Therefore, the fact that a character string is a container ID can be a basis for indicating that the character string is not a trailer ID.

[0070] The fact that a character string indicates the height of a trailer is one of the attributes that indicates that the character string is not a trailer ID. Whether or not a character string indicates the height of a trailer can be determined from the presence or absence of a character indicating a unit and the number of digits of the value. In addition, if there is a possibility that a character string indicating the height of something other than a trailer (for example, the height of a tractor) is written, the fact that the character string indicates the height of a tractor may be used as an attribute that indicates that the character string is not a trailer ID.

[0071] The fact that a string is a common name is one attribute that makes it unlikely to be a trailer ID, since trailer IDs are often constructed as meaningless strings to prevent confusion with other strings.

[0072] The fact that a character string is a company name is one of the attributes that are presumed not to be a trailer ID. This is because trailer IDs are often composed of meaningless character strings to prevent confusion with other character strings. Here, whether or not a character string is a company name can be presumed, for example, by comparing the recognized character string with a database of company names. Note that in order to ensure the uniqueness of trailer IDs, a company name may be included as part of a trailer ID. Therefore, if it is known in advance that a trailer ID that includes a company name exists, the fact that the character string is a company name may be excluded from the attributes that are presumed not to be a trailer ID.

[0073] The fact that a string is a phone number is one of the attributes that indicates that it is not a trailer ID. Whether or not a string is a phone number can be inferred by the number of digits in the recognized string, the format (e.g., the presence or absence of hyphens), or by comparing the contents of the string with a database of phone numbers.

[0074] The fact that a character string is a license plate is one of the attributes that indicates that it is not a trailer ID. This is because the characters that may be written on license plates are often regulated by law, and it is unlikely that a trailer ID is written on a license plate.

[0075] The fact that the character string is a timestamp relating to the date and time, etc., written on the captured image is one of the attributes that indicates that the character string is not a trailer ID.

[0076] The attributes marked with a star in FIG. 4 (character string being written vertically, character string being a container ID) are attributes that are determined with priority over other attributes when estimating whether or not a character string is a trailer ID. Here, the attribute that is determined with priority may be an attribute that is more likely to identify a trailer ID with a higher probability than other attributes or combinations of attributes. In the first embodiment, the reasons why the attributes are used as the attributes are that the character string is written vertically and that the character string is a container ID are as follows. First, character strings written on containers are generally written horizontally, and vertical character strings are less likely to be character strings indicating a company name or the like. Therefore, by selecting vertical character strings, character strings that are less likely to be trailer IDs can be excluded. However, since it is known that container IDs may be written vertically, container IDs may not be excluded solely based on the fact that the character string is written vertically. However, since container IDs must be written in a format defined by a standard, it is possible to reliably determine whether or not a character string is a container ID based on whether or not the format matches. In other words, a character string that is written vertically and is not a container ID can be estimated with a high probability to be a trailer ID.

[0077] The attributes marked with double circles in Figure 4 (the string represents the height of the trailer, and the string is a common name, company name, phone number, license plate, or timestamp) are attributes that are determined not to be a trailer ID if the string matches any of the attributes.

[0078] Next, a process for classifying whether or not a character string is presumed to be a trailer ID will be described with reference to Fig. 5. Fig. 5 is a flowchart for classifying whether or not a character string is presumed to be a trailer ID. Each process in the flowchart of Fig. 5 is executed by processor 41.

[0079] The processor 41 determines whether the character string is a container ID (step St200). If the processor 41 determines that the character string is a container ID (step St200, YES), the processor 41 determines that the character string is not a trailer ID (step St207).

[0080] When the processor 41 determines that the character string is not a container ID (step St200, NO), the processor 41 determines whether the character string is written vertically (step St201).

[0081] When processor 41 determines that the character string is written vertically (YES in step St201), processor 41 determines that the character string is a trailer ID (step St202).

[0082] When the processor 41 determines that the character string is not written vertically (step St201, NO), the processor 41 determines whether the content of the character string is other than a trailer ID (step St203). That is, in the process of step St203, the processor 41 determines whether the character string corresponds to each attribute marked with a double circle in the table of FIG.

[0083] When processor 41 determines that the character string corresponds to any of the attributes marked with a double circle in the table of FIG. 4 (i.e., the character string is other than a trailer ID) (step St203, YES), processor 41 determines that the character string is not a trailer ID (step St207).

[0084] When the processor 41 determines that the character string does not correspond to any of the attributes marked with a double circle in the table of FIG. 4 (i.e., the character string is not other than a trailer ID) (step St203, NO), it determines whether or not the character string is likely to be a trailer ID by using attributes not marked with a symbol in the table of FIG. 4 (hereinafter referred to as normal attributes) (step St204). The processor 41, for example, calculates the degree of possibility that the character string is a trailer ID as a score. For example, the processor 41 adds a score when the character string corresponds to a normal attribute that is estimated to be a trailer ID, and subtracts a score when the character string corresponds to a normal attribute that is estimated not to be a trailer ID. Note that the method of calculating the score is not limited to the above example, and the score may be calculated by combining addition, subtraction, and multiplication. The higher the score, the higher the possibility that the character string is a trailer ID. The processor 41 may sort the character strings based on the calculated score (for example, arrange them in descending order of score).

[0085] When the processor 41 determines that the character string is highly likely to be a trailer ID using the normal attribute (step St204, YES), the processor 41 classifies the character string as being highly likely to be a trailer ID (step St205). For example, the processor 41 determines that the character string is highly likely to be a trailer ID when the calculated score is equal to or greater than a predetermined threshold. The processor 41 may also sort the character strings determined to be highly likely to be trailer IDs in order of score (i.e., arranging in order of high or low score).

[0086] When the processor 41 determines that the character string is unlikely to be a trailer ID using the normal attribute (step St204, NO), the processor 41 classifies the character string as unlikely to be a trailer ID (step St206). For example, when the calculated score is less than a predetermined threshold, the processor 41 determines that the character string is unlikely to be a trailer ID. The processor 41 may also sort the character strings determined to be unlikely to be trailer IDs in order of score (i.e., arranging in order of high score or low score).

[0087] <Other variations> The attribute presumed to be a trailer ID and the attribute presumed not to be a trailer ID in the above-mentioned embodiment are merely examples, and the attribute presumed to be a trailer ID and the attribute presumed not to be a trailer ID based on other conditions may be adopted. Also, the attribute presumed to be a trailer ID and the attribute presumed not to be a trailer ID may be reversed and used as an attribute having an equivalent meaning. For example, instead of using the attribute that a character string is written vertically as the attribute that is presumed to be a trailer ID, the attribute that a character string is written horizontally may be used as the attribute that is presumed not to be a trailer ID.

[0088] In the above-described embodiment, the fact that the character string is written vertically and that the character string is a container ID are exemplified as attributes that are determined to be given priority over other attributes. However, either one of the attributes or another attribute may be used as the attribute that is determined to be given priority. As described above, if the character string is written vertically and is not a container ID, it can be said that the character string is a trailer ID with a high probability. However, for example, if it is known that a vertically written character string does not include a container ID, only the fact that the character string is written vertically may be used as an attribute that is determined to be given priority. Also, for example, if it is known that the trailer ID is written in a special manner, the trailer ID can be recognized more efficiently by using these attributes as attributes that are determined to be given priority.

[0089] In the above-described embodiment, the attributes for recognizing a trailer ID have been mainly described. However, the attributes used for recognizing a trailer ID are not limited to those described above. Similarity to the format of a character string known as a trailer ID may be used as an attribute. For example, since trailer IDs are often configured as a character string consisting of an alphabet followed by a number, the character recognition device 40 may determine that a character string that matches this configuration is highly likely to be a trailer ID. Furthermore, since trailer IDs tend to be written within a certain range of character numbers, the character recognition device 40 may use the number of characters in a character string as an attribute for determining whether or not the trailer ID has been recognized.

[0090] In addition, if the trailer ID is composed of only numbers or written in a notation or character string similar to other known character strings, it may be difficult to recognize the trailer ID based on the above-mentioned attributes. In such a case, the character recognition device 40 may determine whether the recognized character string is a trailer ID based on the similarity between the recognized character string and a character string recorded as a correct trailer ID.

[0091] In the above embodiment, the configuration for recognizing the trailer ID has been described, but the same configuration may be used as the configuration for recognizing the container ID. Note that when identifying the trailer TR, the container ID may be used instead of the trailer ID, so that when managing the trailer TR, the trailer ID and the container ID may be handled without distinction.

[0092] (Summary of the first embodiment) The above description of the first embodiment discloses the following technology.

[0093] <Technology 1> A character recognition device (e.g., character recognition device 40) according to embodiment 1 includes a recognition unit (e.g., character string recognition unit 41B) that recognizes at least one character string from a captured image including a trailer captured by an imaging device (e.g., imaging device 10), an attribute determination unit (e.g., attribute determination unit 41C) that determines an attribute of the character string recognized by the recognition unit, and a trailer ID estimation unit (e.g., trailer ID estimation unit 41D) that estimates whether the character string is a trailer ID based on the attribute of the character string determined by the attribute determination unit.

[0094] As a result, when multiple character strings are written in a trailer at arbitrary positions in various sizes, fonts, or writing styles, the character recognition device according to the first embodiment can detect a character string that is considered to be a trailer ID from among the multiple character strings. In other words, the character recognition device can detect a trailer ID with high accuracy from among multiple character strings written in the trailer.

[0095] <Technology 2> In the character recognition device described in Technology 1, the trailer ID estimation unit estimates that a character string is a trailer ID if the character string is written vertically and the content indicated by the character string is not a container ID that conforms to a predetermined rule.

[0096] As a result, the character recognition device according to the first embodiment can reliably determine whether a character string is a container ID based on whether the character string matches a format defined by a predetermined standard. When a container ID is written vertically, the character recognition device can estimate with a high probability that the character string is not a trailer ID. In other words, the character recognition device can eliminate container IDs with a high probability from multiple character strings and detect character strings that are considered to be trailer IDs with a high degree of accuracy.

[0097] <Technology 3> In the character recognition device described in Technology 1 or 2, if a character string is written vertically and the content indicated by the character string is not a container ID defined in ISO 6346, the trailer ID estimation unit estimates that the character string is the trailer ID.

[0098] As a result, the character recognition device according to the first embodiment can exclude container IDs from a plurality of character strings and detect character strings that are likely to be trailer IDs with high accuracy.

[0099] <Technology 4> In a character recognition device described in any one of Techniques 1 to 3, a trailer ID estimation unit estimates that a character string is not the trailer ID if the character string is not written vertically, the content indicated by the character string is not the container ID, and the content indicated by the character string is something other than a trailer ID.

[0100] As a result, the character recognition device according to the first embodiment can eliminate any horizontally written character string that is not a container ID, if the character string indicates a content other than a trailer ID. In other words, the character recognition device can prevent erroneous detection of a character string other than a trailer ID as a trailer ID.

[0101] <Technology 5> In the character recognition device described in any one of Techniques 1 to 4, the thing indicating something other than the trailer ID is a character string indicating the height of the trailer, indicating a general name, indicating a company name, indicating a telephone number, indicating a license plate, or indicating a time embedded in a captured image by a photographing device.

[0102] This makes it possible for the character recognition device according to the first embodiment to prevent erroneous detection of a character string other than a trailer ID as a trailer ID.

[0103] <Technology 6> In a character recognition device described in any one of Techniques 1 to 5, when a character string is not written vertically, the content indicated by the character string is not a container ID, and the content indicated by the character string is not something other than a trailer ID, a trailer ID estimation unit estimates whether or not the content indicated by the character string is likely to be a trailer ID based on at least one attribute of an attribute related to the writing of the characters, an attribute related to the position of the character string, and an attribute that indicates that the content indicated by the character string is estimated to be a trailer ID.

[0104] As a result, the character recognition device according to the first embodiment can classify multiple character strings written in a trailer with high accuracy based on whether or not they are likely to be trailer IDs.

[0105] <Technology 7> In the character recognition device according to any one of the first to sixth techniques, the trailer ID estimation unit calculates a score representing a likelihood that the character string is a trailer ID.

[0106] In this way, the character recognition device according to the first embodiment can classify character strings written in trailers by assigning scores according to the likelihood that the character strings are trailer IDs.

[0107] <Technology 8> In the character recognition device according to any one of the first to seventh techniques, the trailer ID estimation unit estimates that a character string is likely to be a trailer ID if all characters are the same size and equally spaced apart.

[0108] In this way, the character recognition device according to the first embodiment can extract character strings that are highly likely to be trailer IDs from among a plurality of character strings written in the trailer.

[0109] <Technology 9> In the character recognition device according to any one of the first to eighth aspects, the trailer ID estimation unit determines that a character string is unlikely to be a trailer ID when the character string is decorated.

[0110] In this way, the character recognition device according to the first embodiment can extract and eliminate character strings that are unlikely to be trailer IDs from among a plurality of character strings written in a trailer.

[0111] <Technology 10> In a character recognition device described in any one of Techniques 1 to 9, a trailer ID estimation unit estimates that a character string is highly likely to be a trailer ID when the character string is located above the license plate or within the trailer.

[0112] In this way, the character recognition device according to the first embodiment can extract character strings that are highly likely to be trailer IDs from among a plurality of character strings written in the trailer.

[0113] <Technology 11> In a character recognition device described in any one of Techniques 1 to 10, a trailer ID estimation unit estimates that a character string is unlikely to be a trailer ID if the character string is written on a tractor or above or below a company logo.

[0114] In this way, the character recognition device according to the first embodiment can extract and eliminate character strings that are unlikely to be trailer IDs from among a plurality of character strings written in a trailer.

[0115] <Technology 12> In a character recognition device described in any one of Techniques 1 to 11, a trailer ID estimation unit estimates that a character string is likely to be a trailer ID if the character string consists only of uppercase letters and numbers, if the number of characters in the character string is less than a predetermined number, or if the character string conforms to specific rules for each company.

[0116] In this way, the character recognition device according to the first embodiment can extract character strings that are highly likely to be trailer IDs from among a plurality of character strings written in the trailer.

[0117] <Technology 13> In the character recognition device according to any one of the first to twelfth techniques, the trailer ID estimation unit estimates that the character string is not a trailer ID if the character string is a container ID that complies with a predetermined rule.

[0118] In this way, the character recognition device according to the first embodiment can eliminate character strings that are presumed to be container IDs from among a plurality of character strings written in the trailer.

[0119] <Technology 14> In the character recognition device according to any one of the first to thirteenth aspects, the attribute is a writing style of a character, a position of a character string, or a content indicated by the character string.

[0120] In this way, the character recognition device according to the first embodiment can estimate whether a plurality of character strings written in a trailer is a trailer ID based on each attribute. In other words, the character recognition device can detect a trailer ID from among a plurality of character strings with high accuracy.

[0121] <Technology 15> In the character recognition device according to any one of the first to fourteenth aspects, the trailer ID estimation unit outputs the estimation result to a display device.

[0122] In this way, the character recognition device according to the first embodiment can assist the user who uses the character recognition device in identifying the trailer by using the trailer ID.

[0123] <Background to the second and third embodiments> Patent Document 1 discloses an identical vehicle determination device that does not require calibration to adjust the installation distance of the front and rear cameras by performing accurate relative position measurement. The identical vehicle determination device includes a detection distance estimation unit that estimates the detection distance between the front camera and the rear camera based on the speed and passing time of a vehicle with a matching number calculated by a vehicle speed calculation unit. The identical vehicle determination device further includes an identical vehicle determination unit that calculates an estimated passing time of the vehicle based on the speed of the vehicle calculated by the vehicle speed calculation unit and the detection distance between the front camera and the rear camera estimated by the detection distance estimation unit when the front license plate number recognized by the first recognition unit and the rear license plate number recognized by the second recognition unit do not match. The identical vehicle determination unit determines that the rear license plate number recognized by the second recognition unit within the estimated passing time is the rear license plate number of the vehicle.

[0124] In recent years, logistics using trailers TR has become popular, and there is an increasing need to efficiently manage the location of trailers TR. Here, a trailer TR means a towed vehicle towed by a towing vehicle (also called a tractor LE). A trailer TR may be a combination of a container for storing cargo and a dolly on which the container is mounted, or it may be a vehicle equipped with a portion for storing cargo. A plurality of trailers TR may be gathered at warehouses or terminals located in various locations (e.g., starting points, destination points, relay points, etc.). Since a large number of trailers TR gather at a distribution center 100, there is a need to automatically manage the trailers TR entering and leaving the distribution center 100.

[0125] In Patent Document 1, the same vehicle determination device determines that the vehicles photographed by the front and rear cameras are the same vehicle when the number of the license plate on the front of the vehicle read by the front camera matches the number of the license plate on the rear of the vehicle read by the rear camera. However, when multiple vehicles pass close to each other, the separation between the vehicles may not be distinguished and the multiple vehicles may be erroneously determined as one vehicle. In addition, the license plate number of the tractor LE towing the trailer TR may not match the license plate number attached to the rear of the trailer TR. When the trailer TR is distinguished using the license plate, the license plate numbers attached to the tractors LE at the front and rear of the vehicle do not match the license plate numbers attached to the trailer TR, so that the trailer TR and the tractor LE may be erroneously determined to be different vehicles even though they are the same vehicle.

[0126] In the following second and third embodiments, an example of a passage determination device that distinguishes between trailers traveling in close proximity with high accuracy will be described.

[0127] <Embodiment 2> First, an example of a method for photographing a vehicle V entering and leaving the distribution center 100 will be described with reference to Fig. 6. Fig. 6 is a diagram showing an example of a method for photographing a vehicle V entering and leaving the distribution center 100.

[0128] A vehicle V is traveling on a road RO in the direction of an arrow. In FIG. 6, the vehicle V is traveling in a direction approaching the distribution center 100 and heading toward the distribution center 100.

[0129] A front camera CA1 that photographs the vehicle V from diagonally forward and a rear camera CA2 that photographs the vehicle V from diagonally rear are installed on one side of the road RO. The front camera CA1 and the rear camera CA2 are installed at positions that determine the passage of the trailer TR (hereinafter, referred to as "passage determination positions"). In this way, the vehicle V entering the distribution center 100 is photographed from the front and rear by the two cameras.

[0130] The front camera CA1 is defined as a camera that photographs the trailer TR from the front with respect to the traveling direction. The rear camera CA2 is defined as a camera that photographs the trailer TR from the rear with respect to the traveling direction. The example shown in FIG. 6 shows a case where the trailer TR enters the distribution center 100, but when the trailer TR leaves the distribution center 100, the rear camera CA2 shown in FIG. 6 functions as a camera that photographs the trailer TR from the front, and the front camera CA1 functions as a camera that photographs the trailer TR from the rear. Separate cameras may be used for the entrance and exit of the trailer TR. For the sake of convenience, the following description will be given using a case where the trailer TR enters the distribution center 100. The passage determination device 50 (see FIG. 7) performs a passage determination of the trailer TR using images captured by the front camera CA1 and the rear camera CA2.

[0131] When the vehicle V enters the distribution center 100, if there is no following vehicle or if there is a sufficient distance between the following vehicle and the trailer TR, the passage determination device 50 can photograph the trailer TR with one camera and determine the passage of the trailer TR using the photographed image. However, when multiple trailers TR pass the position where the passage is determined in a close state, it is difficult for the passage determination device 50 to distinguish each trailer TR. The reason for this will be explained below. First, since the shapes of trailers TR are generally similar, it is difficult to distinguish the trailers using the shapes as a clue, and as described above, it is also difficult to accurately obtain information that serves as a clue to distinguish the trailers, such as the trailer ID or container ID. That is, due to the nature of the trailer TR, it is difficult to distinguish each of the multiple trailers TR. In addition, if an attempt is made to distinguish the trailers TR using the color of the trailer TR, etc., there is a risk that the same trailer TR may be determined to be a different trailer TR due to changes in the environment such as lighting or the influence of shadows, or that the trailer TR may be erroneously determined to be an object other than a trailer TR. Therefore, the function of detecting the trailer TR is configured to absorb minute differences between the trailer TRs so as to reliably distinguish between the trailer TRs and objects other than the trailer TRs. Therefore, even if different trailer TRs are included in the multiple captured images, the passage determination device 50 may only be able to recognize that some trailer TR has been detected from each of the multiple captured images. This result cannot be distinguished from a case where the same trailer TR is detected multiple times, so the passage determination device 50 may detect multiple trailer TRs as one trailer TR. In this case, even if multiple trailer TRs have passed, there is a possibility that an erroneous determination may occur in which it is determined that only a smaller number of trailer TRs have passed than the correct number. If an erroneous determination occurs, the passage determination device 50 may have a problem of being unable to correctly manage the trailer TRs in the distribution center 100. Examples of cases where the passage determination device 50 is unable to correctly manage the trailer TRs in the distribution center 100 include a case where a trailer TR with no entry record exists in the distribution center 100, or a trailer TR that should have entered has no record.In view of this situation, there is a need to be able to distinguish between the individual trailers TR and determine their passage with high accuracy, even when multiple trailers TR pass close to each other at a position where their passage is to be determined.

[0132] Next, a block diagram of the passage determination device 50 according to the second and third embodiments will be described with reference to Fig. 7. Fig. 7 is a block diagram of the passage determination device 50 according to the second and third embodiments. Note that the same components as those in the character recognition system 1 shown in Fig. 2 are denoted by the same reference numerals and will not be described.

[0133] The passage determination system 2 includes a front camera CA1, a rear camera CA2, a database 20, a display device 30, and a passage determination device 50.

[0134] The front camera CA1 is a device that photographs the trailer TR from the front in the traveling direction. The rear camera CA2 is a device that photographs the trailer TR from the rear in the traveling direction. The front camera CA1 and the rear camera CA2 are configured to have at least a lens (not shown) and an image sensor (not shown) as optical elements. The lens forms an optical image of the object on the light receiving surface (in other words, the photographing surface) of the image sensor by receiving light reflected by an object within the angle of view of the area photographed by the front camera CA1 and the rear camera CA2. The image sensor is, for example, a solid-state photographing element such as a CCD or a CMOS. The image sensor converts the optical image formed on the photographing surface through the lens into an electrical signal every predetermined time (for example, 1 / 30 (second)). For example, when the predetermined time is 1 / 30 (second), the frame rate of the front camera CA1 and the rear camera CA2 is 30 fps. In addition, the front camera CA1 and the rear camera CA2 may generate a photographed image by performing a predetermined signal processing on the electrical signal every predetermined time described above. The front camera CA1 and the rear camera CA2 output captured images to the passage determination device 50 via the communication I / F 42. In the following, the captured image captured by the front camera CA1 is referred to as a first captured image, and the captured image captured by the rear camera CA2 is referred to as a second captured image.

[0135] The database 20A is configured using a storage medium (e.g., a flash memory, a HDD), and an SSD. The database 20A stores identification information (e.g., a trailer ID or a container ID, etc.) of a trailer TR that is determined to have passed through a passage determination position by the passage determination device 50. That is, the database 20A stores identification information of a trailer TR that exists in the distribution center 100.

[0136] The display device 30A displays the determination result executed by the passage determination device 50. The display device 30A is, for example, a touch panel display, a display, a mobile phone, or the like. The display device 30A may be configured integrally with the passage determination device 50, or may be omitted from the passage determination system 2.

[0137] The passage determination device 50 includes a communication I / F 42, a memory 43, and a processor 51. The passage determination device 50 is, for example, a PC or a server device.

[0138] The processor 51 is, for example, a CPU, a DSP, a GPU, or an FPGA. The processor 51 functions as a controller that manages the overall operation of the passage determination device 50. The processor 51 performs control processing for managing the operation of each part of the passage determination device 50, data input / output processing between each part of the passage determination device 50, data calculation processing, and data storage processing. The processor 51 operates according to a program stored in the memory 43. The processor 51 uses the memory 43 during operation, and temporarily stores data generated or acquired by the processor 51 in the memory 43. The processor 51 realizes the functions of the character string detection unit 41A, the character string recognition unit 41B, the attribute determination unit 41C, the trailer ID estimation unit 41D, the tractor detection unit 41E, the passage determination unit 41F, and the database update unit 41G by using the program and data stored in the memory 43.

[0139] The tractor detection unit 41E detects the tractor LE from the captured images acquired from the front camera CA1 and the rear camera CA2. The tractor detection unit 41E detects the tractor from the captured images using a known artificial intelligence (hereinafter referred to as "AI") technique such as image recognition.

[0140] The passage determination unit 41F detects the trailer TR from the captured images acquired from the front camera CA1 and the rear camera CA2 using a known AI technology such as image recognition. The passage determination unit 41F performs a determination that the trailer TR has passed the passage determination position (hereinafter referred to as "passage determination") using information on the detected trailer TR. Here, the passage determination position is, for example, a gate provided at the entrance / exit of the distribution center 100 or a position on the way to the entrance / exit of the distribution center 100. The passage determination position can be set at any position by a user (for example, a business operator who introduces the passage determination system 2), and the front camera CA1 and the rear camera CA2 are installed at the passage determination position. The passage determination position is not limited to one position and may be provided in multiple positions. A detailed description of the passage determination performed by the passage determination unit 41F will be described later.

[0141] The database update unit 41G registers the identification information of the trailer TR that has been determined to have passed by the passage determination unit 41F in the database 20A. The database update unit 41G may register the identification information of the trailer TR in the database 20A by linking it to the position and time at which the passage was determined.

[0142] Next, an example of time series data of images captured by the front camera CA1 and the rear camera CA2 will be described with reference to Fig. 8. Fig. 8 is a diagram showing an example of time series data of images captured by the front camera CA1 and the rear camera CA2.

[0143] The captured images PF1, PF2, PF3, PF4, PF5, PF6, PF7, PF8, and PF9 are captured by the front camera CA1. The captured images PF1, PF2, PF3, PF4, PF5, PF6, PF7, PF8, and PF9 are arranged in chronological order in the direction of the arrows.

[0144] In the photographed image PF2, the tractor LE is detected within the detection frame F1, and the vehicle V begins to be detected. After that, in the photographed image PF3, the trailer TR is detected for the first time within the detection frame F2, and then the trailer TR is detected in the photographed images PF4 and PF5. In the photographed image PF6, the vehicle V passes the passage determination position and goes out of the frame of the photographed image.

[0145] The captured images PR1, PR2, PR3, PR4, PR5, PR6, PR7, PR8, and PR9 are captured by the rear camera CA2. The captured images PR1, PR2, PR3, PR4, PR5, PR6, PR7, PR8, and PR9 are arranged in chronological order in the direction of the arrows.

[0146] In the captured image PR7, the trailer TR is detected for the first time within the detection frame F2. After that, in the captured image PR8, the vehicle V moves away from the passage determination position, and the area of ​​the detection frame F4 for the trailer TR becomes smaller.

[0147] When the area within the frame is less than a predetermined threshold, such as detection frame F1 of photographed image PF2 and detection frame F4 of photographed image PR8, the passage determination device 50 may set it not to be included in the passing time of the trailer TR.

[0148] Next, detection conditions for the trailer TR will be described with reference to Fig. 9. Fig. 9 is a diagram showing the size of the detection frame for the trailer when multiple vehicles are photographed.

[0149] The captured image PF10 is an image of two vehicles (vehicle V and vehicle V2) captured by the front camera CA1.

[0150] The detection frame F5 is a rectangular frame set in the detected trailer TR of the vehicle V. The detection frame is, for example, a frame that includes the trailer TR inside. The detection frame may be a frame with the minimum area that includes the entire detected trailer TR inside, or may be a frame that is enlarged by a ratio arbitrarily set by the user, such as several percent, from the frame with the minimum area that includes the entire detected trailer TR inside. The detection frame may include the wheels, or may include only the container part of the trailer TR. However, when the detection frame F5 is used for some kind of judgment processing, if the range or size of the detection frame is allowed to be arbitrarily changed by the user, there is a risk that the accuracy of the judgment will deteriorate. Therefore, when the detection frame F5 is used for some kind of judgment processing, the setting of the detection frame F5 may not be changed by the user. As an example of using the detection frame F5 for some kind of judgment processing, it is possible to use the detection frame F5 as a search range for a trailer ID or the like, or to estimate whether the detection frame F can properly detect the trailer based on the validity of the area of ​​the detection frame F5.

[0151] Vehicles V and V2 are traveling side by side in the traveling direction of vehicle V. The vehicle subject to the passage judgment is vehicle V, but there is a possibility that adjacent vehicle V2 may be detected, which may cause an erroneous judgment. In order to prevent the detection results of vehicles other than the subject of the passage judgment from being used in the passage judgment, a threshold may be set for the size of the detection frame, and trailers TR for which a detection frame less than the threshold is set may be excluded from the passage judgment. This makes it possible for the passage judgment device 50 to detect trailers TR other than the subject of the passage judgment and to prevent erroneous judgments such as judging a trailer TR that has not passed as having passed, or linking the identification information of another trailer TR to a trailer TR that has been judged to have passed.

[0152] Next, several cases of the passage determination of the trailer TR will be described with reference to Fig. 10. Fig. 10 is a diagram showing several cases of the passage determination of the trailer TR.

[0153] The diagram shown in Fig. 10 is a diagram showing trailers A, B, C, and D detected by the front camera CA1 and the rear camera CA2 at times T1 to T25. Trailers A, B, C, and D are each subjected to a passage determination. In Fig. 10, a method for the passage determination device 50 to determine the passage of each of trailers A, B, C, and D will be described.

[0154] The passage determination device 50 performs a passage determination of the trailer TR using the first time P1 and the second time P2. After detecting the trailer TR from the first photographed image, the passage determination device 50 determines that the trailer TR detected within the first time P1 is the same trailer TR as the trailer TR detected immediately before. The passage determination device 50 also determines that the trailer TR detected from the second photographed image within the second time P2 after the last trailer TR was detected from the first photographed image is the same trailer TR as the trailer TR detected in the first photographed image. The second time P2 is a threshold value that is set from the time when the trailer TR detected in the first photographed image is no longer detected. The first time P1 is set to several seconds, for example, 3 seconds. The second time P2 is set to several tens of seconds, for example, 120 seconds. Here, if the first time P1 is too long, there is a high possibility that a subsequent trailer TR will be erroneously determined to be the same trailer TR, so the first time P1 is set to a short time. On the other hand, in the second photographed image in which the trailer TR is photographed from the rear, it is highly likely that the trailer TR can continue to be detected even if it is no longer detected in the first photographed image in which the trailer TR is photographed from the front. Therefore, if the second time P2 is too short, there is a high possibility that the trailer TR that is no longer detected in the first photographed image will be erroneously determined to be a different trailer TR, so the second time P2 is set to a long time. Note that the first time P1 and the second time P2 can be arbitrarily set by the user and are not limited to the above-mentioned example.

[0155] First, using an example of trailer A, a case will be described in which a trailer TR passes through the passage determination position alone without any trailer TR nearby.

[0156] The passage determination device 50 detects the trailer A from time T1 to T3 in the first captured image. Thereafter, the passage determination device 50 detects the trailer A from time T5 to T6 in the second captured image. The trailer TR detected from time T1 to time T3 is detected again within a first time P1 after it was detected once, and is therefore determined by the passage determination device 50 to be the same trailer A. The trailer TR detected from the second captured image at time T5 is detected within a second time P2 after the trailer A was detected by the front camera CA1 at time T3, and is therefore determined by the passage determination device 50 to be the same trailer A as the trailer TR detected in the first captured image at time T3. The passage determination device 50 performs a passage determination for the trailer A at time T5, when the trailer A is detected for the first time from the second captured image after the trailer A is detected from the first captured image.

[0157] Next, using an example of trailers B, C, and D, a case in which trailers TR pass through the passage determination position in close proximity will be described.

[0158] The passage determination device 50 detects the trailer B from time T9 to time T11 in the first captured image. After that, the passage determination device 50 detects the trailer B from time T13 to time T14 in the second captured image. The trailer TR detected in the first captured image from time T9 to time T11 is detected within a first time P1 after it was detected once, and is therefore determined by the passage determination device 50 to be the same trailer B. The trailer TR detected in the second captured image at time T13 is detected within a second time P2 after the trailer B was detected by the front camera CA1 at time T11, and is therefore determined by the passage determination device 50 to be the same trailer B as the trailer TR detected by the front camera CA1 at time T11. The passage determination device 50 performs a passage determination for the trailer B at time T13, when the trailer B is detected for the first time from the second captured image after the trailer B is detected in the first captured image.

[0159] The passage determination device 50 detects trailer C in the first captured image from time T14 to time T18. Thereafter, the passage determination device 50 detects trailer C in the second captured image from time T19 to time T20. The trailer TR detected in the first captured image from time T14 to time T18 is determined to be the same trailer C by the passage determination device 50 because it was detected within a first time P1 after it was detected once. However, although trailer C is actually captured by the front camera CA1 from time T14 to time T16, trailer D exists closely behind trailer C, so the division of trailer TR is not identified, and trailer TR detected from time T17 to time T18 is also recognized as trailer C.

[0160] After detecting trailer C in the first captured image, the passage determination device 50 detects trailer TR in the second captured image for the first time at time T19. Since the passage determination device 50 detects trailer TR in the second captured image at time T19, it determines that trailer TR detected in the first captured image at time T19 is trailer D separate from trailer C. Since the passage determination device 50 detects trailer C in the second captured image at time T19, it performs a passage determination of trailer C at time T19.

[0161] The passage determination device 50 detects the trailer D from time T19 to time T20 in the first captured image. Thereafter, the passage determination device 50 detects the trailer D from time T23 to time T24 in the second captured image. Because the trailer D is traveling close to the trailer C, the passage determination device 50 determines that the trailer TR shown in the first captured image from time T17 to time T18 is actually the trailer D, but the passage determination device 50 determines that the trailer TR is the trailer C. Since the trailer C is first detected in the second captured image at time T19, the passage determination device 50 determines that the trailer TR detected in the first captured image at time T19 is the trailer D. Furthermore, the trailer TR detected from the second captured image at time T23 was detected within the second time P2 after the trailer D was detected from the first captured image at time T20, and the passage determination device 50 determines that the trailer TR is the same trailer D as the trailer TR detected in the first captured image at time T20.

[0162] In this manner, in this embodiment, even in a situation where it is difficult to distinguish between trailer C and trailer D using only the first captured image, it is possible to determine that trailer D was captured by using information from the second captured image. Note that in this embodiment, although it is difficult to accurately determine the timing at which trailer D began to be captured in the first captured image, it is possible to at least recognize that trailer D exists, and therefore the accuracy of trailer recognition is improved compared to a case where only trailer C is erroneously recognized using only the first captured image.

[0163] Next, a case where the trailer TR is not detected for the second time P2 or more will be described with reference to Figures 11 and 12. Figure 11 is a diagram showing a first case where the trailer TR is not detected for the second time P2 or more. Figure 12 is a diagram showing a second case where the trailer TR is not detected for the second time P2 or more.

[0164] 11, the passage determination device 50 detects trailer E in the first captured image from time T26 to time T28. Thereafter, from time T29 to time T31, the passage determination device 50 does not detect trailer TR in either the first captured image or the second captured image. Thereafter, at time T32, the passage determination device 50 detects trailer TR from the first captured image. At time T32, since a time equal to or longer than the second time P2 has elapsed since trailer E was detected at time T28, the passage determination device 50 determines that trailer TR detected at time T32 is trailer F, which is different from trailer E, and performs a passage determination for trailer E.

[0165] 12, the passage determination device 50 detects trailer E in the first captured image from time T26 to time T28. Thereafter, from time T29 to time T33, the passage determination device 50 does not detect trailer TR in either the first captured image or the second captured image. Thereafter, at time T34, the passage determination device 50 detects trailer TR in the second captured image. At time T34, since a time equal to or longer than the second time P2 has elapsed since trailer E was detected at time T28, the passage determination device 50 determines that trailer TR detected at time T34 is trailer F, which is different from trailer E, and performs a passage determination for trailer E.

[0166] As described above, the passage determination device 50 detects the trailer TR from the images captured by the front camera CA1 and the rear camera CA2, and performs passage determination using the first time P1 and the second time P2.

[0167] Next, an example of the process of determining whether a trailer TR passes will be described with reference to Fig. 13. Fig. 13 is a flowchart showing an example of the process of determining whether a trailer TR passes.

[0168] The processor 51 acquires captured images from the front camera CA1 and the rear camera CA2 (step St101). It is assumed that the processor 51 acquires captured images from the front camera CA1 and the rear camera CA2 at predetermined time intervals while executing the process of the flowchart of FIG.

[0169] The processor 51 associates the acquired captured image with the current time as the capture time (step St102).

[0170] The passage determination unit 41F analyzes the photographed image acquired in the process of step St101 and determines whether or not a trailer TR has been detected (step St103).

[0171] When the passage determination unit 41F determines that the trailer TR has not been detected (step St101, NO), it executes the process of step St101 again. The passage determination unit 41F determines that the trailer TR has not been detected when, for example, no vehicle is captured in the captured image or a tractor LE not towing a trailer TR (i.e., only a tractor LE) or a vehicle not including a trailer TR, such as a passenger car, is detected.

[0172] When the passage determination unit 41F determines that the trailer TR has been detected (step St101, YES), it outputs an instruction to the character string detection unit 41A to detect a character string written on the trailer TR. The processor 51 uses the functions of the character string detection unit 41A, the character string recognition unit 41B, the attribute determination unit 41C, and the trailer ID estimation unit 41D to detect a trailer ID as identification information of the trailer TR from among the character strings written on the trailer TR by the method described in the first embodiment (step St104). Note that the processor 51 may detect a container ID as identification information of the trailer TR, not limited to the trailer ID.

[0173] The passage determination unit 41F determines whether or not the trailer TR is detected again within the first hour after the trailer TR is detected in the first captured image (step St105).

[0174] If the passing determination unit 41F does not detect the trailer TR again within the first hour after detecting the trailer TR in the first captured image (step St105, NO), it determines whether the trailer TR has been detected in the first or second captured image within the second hour after the trailer TR was last detected in the first captured image (step St106).

[0175] When the trailer TR is detected in the first captured image or the second captured image within the second hour after the trailer TR was last detected in the first captured image (step St106, YES), the passage determination unit 41F determines that the trailer TR has passed (step St110). The cases in which the result of the process in step St106 is YES are, for example, the cases of trailer A, trailer B, and trailer D shown in FIG.

[0176] When the passage determination unit 41F does not detect a trailer TR in the first or second photographed image within the second hour (step St106, NO), when the trailer TR is detected in the first or second photographed image at a time exceeding the second hour, the passage determination unit 41F detects the trailer TR as the next trailer TR (step St107). The trailer TR detected as the next trailer TR in the process of step St107 is, for example, the trailer F shown in FIG. 11 and FIG. 12. After the process of step St107, the passage determination unit 41F executes a passage determination of the trailer TR detected by the front camera CA1 (step St110). Note that the passage determination unit 41F may execute a passage determination of the trailer TR at the time when the trailer TR is not detected by the rear camera CA2 beyond the second threshold in the process of step St106.

[0177] If the passage determination unit 41F detects a trailer TR again within one hour after detecting the trailer TR in the first captured image (step St105, YES), it determines whether or not the trailer TR is detected in the second captured image (step St108).

[0178] When the passage determination unit 41F determines that the trailer TR has not been detected in the second captured image (step St108, NO), it executes the process of step St105 again. A case where the process of step St108 is NO is, for example, a case in which the trailer C has been detected within the first captured image within the first hour from time T14 to time T18 in FIG. 10, but the trailer C has not yet been detected in the second captured image. In other words, the passage determination unit 41F repeatedly executes the process of step St105 and the process of step St108 while the trailer TR continues to be detected within the first captured image within the first hour.

[0179] When the passage determination unit 41F determines that the trailer TR has been detected in the second captured image (step St108, YES), the passage determination unit 41F detects the trailer TR that appears in the first captured image at the time when the trailer TR was detected in the second captured image as the next trailer (step St109). The case in which the process of step St109 is performed corresponds to, for example, the case of trailers C and D at time T19 shown in FIG.

[0180] The passage determination unit 41F detects the trailer TR in the first captured image and performs a passage determination of the trailer TR (step St110). The passage determination unit 41F outputs information of the trailer TR that has been determined to have passed to the database update unit 41G. The information of the trailer TR that has been determined to have passed includes, for example, the identification information of the trailer TR that the passage determination unit 41F detects in the process of step St104, the time when it is determined that the trailer TR has passed, and the passage determination position.

[0181] The database update unit 41G associates the identification information of the trailer with the trailer TR determined to have passed, and registers them in the database 20A (step St111).

[0182] As described above, the passage determination device 50 can use the two cameras, the front camera CA1 and the rear camera CA2, to distinguish between multiple trailers TR traveling close to each other with high accuracy and perform passage determination for the trailers TR. Since the passage determination device 50 can determine with high accuracy which trailers TR have passed the passage determination position, it can accurately manage the trailers TR in the distribution center 100. As a result, the passage determination device 50 can support the user in managing the trailers TR in the distribution center 100.

[0183] Furthermore, the passage determination device 50 can determine whether the trailers detected in the first photographed image and the second photographed image are the same trailer or different trailers by the process of step St106. This allows the passage determination device 50 to manage which photographed image shows the trailer that has been determined to have passed. In other words, the passage determination device 50 can manage which photographed image shows the trailer that has been determined to have passed, and therefore can store and manage the image of the trailer that has been determined to have passed in the database 20A without mixing it with images of different trailers. Furthermore, the passage determination device 50 may adopt an ID that is likely to be correct among IDs detected from a group of images determined to be the same trailer as the trailer ID or container ID of each trailer (see embodiment 1 for determining whether the ID is correct or not). This allows the passage determination device 50 to detect the trailer ID or container ID more accurately, since the number of images that are candidates for detecting the trailer ID or container ID increases, even if the trailer ID or container ID is written in a position that is difficult to detect at a specific angle. In particular, the passage determination device 50 determines that the first captured image captured by the front camera CA1 and the second captured image captured by the rear camera CA2 include the same trailer TR under a predetermined condition. Therefore, even if the trailer ID or container ID is written in a position that can only be detected from the front or rear, the trailer ID or container ID can be detected from the captured image captured by either camera.

[0184] Furthermore, the passage determination device 50 may assign the same trailer ID or container ID to images determined to be of the same trailer and record them in the database 20A. This allows the passage determination device 50 to assign a trailer ID or container ID to captured images in which a trailer ID or container ID could not be detected, even if the trailer ID or container ID can only be detected from some of a plurality of images determined to be of the same trailer. This allows the passage determination device 50 to extract captured images taken from various viewpoints all at once, for example, when it is desired to check images of trailers to which the same ID has been assigned.

[0185] In this embodiment, the trailer TR can be detected by a method of detecting an object that is likely to be a trailer TR from a captured image by image recognition, and by a method of regarding a trailer TR as having been detected when a trailer ID is detected from a captured image. The situations described in this embodiment can occur regardless of which method is adopted. However, the likelihood of the situation occurring may differ depending on the adopted method. For example, the situation in FIG. 11 is more likely to occur when a method of regarding the detection of a trailer ID as the detection of a trailer TR is adopted than when a method of detecting a trailer TR by image recognition is adopted. The reason for this will be explained below. When a trailer TR is traveling along a regular route, the trailer TR photographed by the front camera CA1 will always pass through the photographing range of the rear camera CA2 thereafter. Therefore, if the accuracy of image recognition is sufficiently high, a situation will not occur in which the trailer TR is detected from an image photographed by the front camera CA1 and cannot be detected from an image photographed by the rear camera CA2, unless the trailer TR turns back immediately after being photographed by the front camera CA1 or moves off the regular route. On the other hand, in a method in which a trailer TR is considered to have been detected when a trailer ID is detected, if the trailer ID is written in a position that is only visible from the front camera CA1, a situation will arise in which the trailer TR is detected from the image captured by the front camera CA1 but cannot be detected from the image captured by the rear camera CA2.

[0186] <Embodiment 3> In the first embodiment, a method for distinguishing between a plurality of vehicles passing closely through a passing judgment position using two cameras, a front camera CA1 and a rear camera CA2, and a first time P1 and a second time P2, was described. Next, in the third embodiment, a second method for distinguishing between a plurality of vehicles passing closely through a passing judgment position will be described.

[0187] First, an example of time series data of captured images of two trailers TR passing close to each other will be described with reference to Fig. 14. Fig. 14 is a diagram showing an example of time series data of captured images of two trailers TR passing close to each other.

[0188] The captured images IM1, IM2, IM3, IM4, IM5, IM6, and IM7 are captured by the front camera CA1. The captured images IM1, IM2, IM3, IM4, IM5, IM6, and IM7 are arranged in chronological order in the direction of the arrows.

[0189] The tractor detection unit 41E detects the tractor LE of the vehicle V1 in a detection frame F7 from the captured image IM1. The passage determination unit 41F detects the trailer TR in a detection frame F8 from the captured image IM1.

[0190] After that, the tractor detection unit 41E detects the tractor LE of the vehicle V2 in the detection frame F9 from the captured image IM3. In the captured images IM3 to IM5, the passage determination device 50 detects the vehicle V2 in the same captured image as the vehicle V1. In other words, the vehicle V2 is traveling close to the vehicle V1.

[0191] Thereafter, in the captured images IM6 to IM7, the trailer TR of the vehicle V1 disappears from the angle of view of the front camera CA1, and the passage determination device 50 detects only the vehicle V2.

[0192] In the third embodiment, a passage determination method for the case where a plurality of vehicles pass through the passage determination position in close proximity will be described.

[0193] Next, a passage determination method according to the third embodiment will be described with reference to Fig. 15. Fig. 15 is a diagram showing the passage determination method according to the third embodiment.

[0194] Schematic diagram S1 is a schematic diagram of the first captured image at time K1. Schematic diagram S2 is a schematic diagram of the first captured image at time K2. Schematic diagram S3 is a schematic diagram of the first captured image at time K3. Schematic diagram S4 is a schematic diagram of the first captured image at time K4. Time passes from times K1 to K4 in the direction of the arrows, and time K4 is later than time K1.

[0195] The passage determination device 50 sets the X-axis to one side of the outer periphery of the captured image that faces the traveling direction of the vehicle V. In the example of FIG. 15, the passage determination device 50 sets the X-axis to the forward direction opposite to the traveling direction of the vehicle V. Note that the method of setting the X-axis is not limited to the above example. For example, the passage determination device 50 may set the X-axis to any direction that corresponds to the traveling direction of the vehicle V. In other words, the set direction of the X-axis may be any direction that is correlated with the traveling direction of the vehicle, and does not necessarily have to be parallel to the traveling direction of the vehicle.

[0196] The passage determination device 50 calculates the X coordinate of the trailer TR from the first captured image. When multiple vehicles are detected in the first captured image, the passage determination device 50 calculates the X coordinate of only the trailer TR of one vehicle traveling ahead in the traveling direction. The passage determination device 50 may calculate the X coordinate of all vehicles shown in the captured image, not limited to the X coordinate of only one trailer TR. In this case, the passage determination device 50 sets, for example, the X coordinate of the trailer TR having the largest area of ​​the detection frame set in the trailer TR as the passage determination target. The passage determination device 50 may also set the trailer TR having the smallest detected X coordinate as the passage determination target. In the following, the passage determination device 50 calculates the X coordinate of only the trailer TR of one vehicle traveling ahead in the traveling direction.

[0197] At time K1, the passage determination device 50 detects the vehicle V1 and the vehicle V2. The passage determination device 50 sets the vehicle V1 ahead in the traveling direction as a vehicle to be subjected to passage determination. The passage determination device 50 calculates a coordinate Xn which is the X coordinate of the trailer TR of the vehicle V1 in the first captured image at each time. The passage determination device 50 calculates the X coordinate of the center of a detection frame (not shown) set in the trailer TR. Note that the X coordinate calculated by the passage determination device 50 is not limited to the center of the detection frame, and may be any position of the detection frame previously set by the user. For example, it is conceivable to calculate the smallest X coordinate, the largest X coordinate, or the X coordinate of a portion obtained by dividing the X coordinate of the detection frame at a predetermined ratio in the detection frame.

[0198] At time K2, the value of the coordinate Xn decreases as a result of the vehicle V1 moving forward.

[0199] At time K3, the vehicle V1 advances further and goes out of the field of view of the front camera CA1, and the passage determination device 50 detects only the vehicle V2. The passage determination device 50 calculates the coordinate Xn+1, which is the X coordinate of the trailer TR of the vehicle V2. Comparing the coordinate Xn at time K2 with the coordinate Xn+1 at time K3, the value of the coordinate Xn+1 is greater than the value of the coordinate Xn. Normally, the value of the X coordinate of the trailer TR of the same vehicle decreases as the vehicle advances. In other words, the increase in the value of the X coordinate indicates that the vehicle ahead of the vehicles traveling closely has gone out of the field of view of the front camera CA1, and the following vehicle is detected by the front camera CA1.

[0200] Here, the passage determination device 50 wants to determine that a vehicle ahead (i.e., a trailer TR ahead) has passed, triggered by an increase in the X coordinate of the trailer TR. However, the passage determination device 50 may erroneously detect a trailer TR that is different from the subject of passage determination that appears in the first captured image (e.g., a parked trailer TR, etc.) or an object other than the trailer TR, and determine that the X coordinate has increased. In order to prevent such erroneous determination, when the X coordinate of the trailer TR increases, the passage determination device 50 determines that the trailer TR ahead, which is the subject of passage determination, has passed by detecting a tractor LE at an X coordinate between the X coordinate before the increase and the X coordinate after the increase.

[0201] Next, an example of the passage determination process for a trailer TR according to the third embodiment will be described with reference to Fig. 16. Fig. 16 is a flowchart showing an example of the passage determination process for a trailer TR according to the third embodiment.

[0202] The processor 51 acquires a captured image from the front camera CA1 (step St301).

[0203] The passage determination unit 41F determines whether or not a trailer TR is detected from the captured image (step St302).

[0204] When determining that the trailer TR has not been detected from the captured image (step St302, NO), the passage determination unit 41F executes the process of step St301 again.

[0205] When the passage determination unit 41F determines that a trailer TR has been detected from the captured image (step St302, YES), it outputs an instruction to the character string detection unit 41A to detect a character string written on the trailer TR. The processor 51 uses the functions of the character string detection unit 41A, the character string recognition unit 41B, the attribute determination unit 41C, and the trailer ID estimation unit 41D to detect a trailer ID as identification information of the trailer TR from among the character strings written on the trailer TR by the method described in the first embodiment (step St303). Note that the processor 51 may detect a container ID as identification information of the trailer TR, not limited to the trailer ID.

[0206] The passage determination unit 41F calculates the X coordinate of the trailer TR that is the object of passage determination (step St304).

[0207] The passage determination unit 41F determines whether the X coordinate calculated in the process of step St304 has suddenly increased from the X coordinate calculated at the previous time (step St305). The processor 51 acquires images from the front camera CA1 at predetermined time intervals, and the X coordinate calculated at the previous time is the X coordinate of the trailer TR calculated from the image acquired immediately before the image used to calculate the X coordinate in the process of step St304.

[0208] When it is determined that the X coordinate has not increased suddenly (step St305, NO), the passage determination unit 41F determines whether the X coordinate has disappeared (step St306).

[0209] When the passage determination unit 41F determines that the X coordinate has not disappeared (that is, the trailer TR that is the subject of the passage determination has not gone out of the frame of the front camera CA1) (step St306, NO), the passage determination unit 41F acquires a photographed image from the front camera CA1 again (step St308). After executing the process of step St308, the passage determination unit 41F executes the process of step St304 again.

[0210] When it is determined that the X coordinate has disappeared (step St306, YES), the passage determination unit 41F assumes that the trailer TR that is the subject of the passage determination has gone out of the frame of the front camera CA1, and determines that the trailer TR has passed (step St307).

[0211] When it is determined that the X coordinate has suddenly increased (step St305, YES), the passage determination unit 41F determines whether or not the tractor LE has been detected between the X coordinates before and after the sudden increase (step St309).

[0212] When the passage determination unit 41F determines that the tractor LE has not been detected between the X coordinates before and after the sudden increase (step St309, NO), it executes the process of step St301 again. When the process of step St309 is NO, there is a high possibility that the sudden increase in the X coordinate is an erroneous determination. For this reason, the processor 51 starts again from the process of step St301. As will be described later, the passage determination process according to the third embodiment may be performed simultaneously with the passage determination process according to the second embodiment, and when the process of step St309 in FIG. 12 is NO, the passage determination process of the trailer TR is performed using the passage determination process according to the second embodiment.

[0213] When it is determined that the tractor LE is detected between the X coordinates before and after the sudden increase (step St309, YES), the passage determination unit 41F determines that the trailer TR has passed (step St310).

[0214] After determining in step St310 that the trailer TR subject to the passage determination has passed, the passage determination unit 41F proceeds to the passage determination process for the trailer TR of the following vehicle. The passage determination unit 41F detects the identification information of the trailer TR of the following vehicle (step St311). The process of step St311 is the same as the process of step St303.

[0215] The passage determination unit 41F acquires a captured image from the front camera CA1 (step St312). After the process of step St312, the passage determination unit 41F executes the process of step St304.

[0216] As a result, the passage determination device 50 can distinguish between multiple trailers TR traveling close to each other with high accuracy. This allows the passage determination device 50 to perform the passage determination process for the trailers TR with high accuracy and accurately manage the trailers TR present in the distribution center 100. Furthermore, the passage determination device 50 can perform the passage determination process for the trailers TR with high accuracy using one camera.

[0217] Here, in this embodiment, the passage determination device 50 also uses the coordinates at which each trailer is detected in the passage determination process, so that it is possible to accurately distinguish between trailers C and D, which was difficult to distinguish using the method of embodiment 2 based on the time the trailer was detected.

[0218] In the above-mentioned third embodiment, the direction opposite to the traveling direction of the vehicle V is set as the positive direction of the X-axis, but this is merely an example. Any of the directions on the set X-axis may be set as the positive direction. For example, if the direction along the traveling direction of the vehicle V is set as the positive direction of the X-axis, a similar passing determination process can be performed by replacing the sudden increase in the X-coordinate in the above-mentioned process with a sudden decrease in the X-axis. In addition, whether or not a sudden change (sudden increase or decrease) in the coordinate has occurred can be determined, for example, by whether or not the change amount of the coordinate in a predetermined time is equal to or greater than a predetermined threshold. The predetermined time may be, for example, a time corresponding to a certain number of frames set by the user. In addition, the predetermined threshold may be an arbitrary value set by the user, or may be a value based on the average value of the change amount of the coordinate in the past.

[0219] Furthermore, the passage determination device 50 may simultaneously execute the passage determination process described in the second embodiment and the passage determination process described in the third embodiment. This allows the passage determination device 50 to perform the passage determination process with high accuracy even for a plurality of trailers TR traveling closely. Furthermore, the passage determination device 50 preferentially adopts the result of the passage determination process described in the third embodiment in the passage determination process for a trailer TR traveling closely. In other words, when the passage determination device 50 determines that a trailer TR has passed in the passage determination process described in the third embodiment, it also determines that the trailer TR has passed in the passage determination process described in the second embodiment, and moves to the passage determination process for the following station.

[0220] (Summary of the second and third embodiments) The above description of the second and third embodiments discloses the following techniques.

[0221] <Technology 1> The passing determination device of this embodiment includes a processor that acquires first captured images obtained by repeatedly capturing an area including the passing determination position of the trailer from ahead in the direction of travel of the trailer, performs processing to detect the trailer in the first captured images, and determines that the trailer detected in the first captured image taken at the first time has passed the passing determination position if the trailer is detected in the first captured image taken at the first time but not detected in the first captured image taken within one hour from the first time.

[0222] This allows the passage determination device according to the present embodiment to distinguish closely-traveling trailers with high accuracy, thereby enabling the passage determination device to accurately identify trailers that have passed the passage determination position and to accurately manage trailers within the distribution center.

[0223] <Technology 2> In the passage determination device described in Technology 1, the processor further acquires second captured images obtained by repeatedly capturing an area including the passage determination position from behind the trailer's direction of travel, and performs processing to detect the trailer in the second captured images.If a trailer is detected from the first captured image taken within the first hour from the first time at which the trailer was detected in the first captured image, if the trailer is not detected from the second captured image taken at a time after the first time, it is determined that the trailer detected from the first captured image taken at the first time has not passed the passage determination position, and if a trailer is detected from the second captured image taken at a time after the first time, it is determined that the trailer detected from the first captured image taken at the first time has passed the passage determination position.

[0224] This makes it possible for the passage determination device according to the present embodiment to prevent erroneous determination that multiple trailers traveling close to each other are a single trailer. The passage determination device can reduce erroneous determinations and improve determination accuracy by distinguishing between trailers using two cameras.

[0225] <Technology 3> In the passage determination device described in Technology 1 or Technology 2, the processor determines that a trailer detected from a first captured image taken at a time before a second time, which is the capture time of the second captured image in which the trailer is detected, and a trailer detected from a first captured image taken after the second time are different trailers.

[0226] As a result, the passage determination device according to the present embodiment can accurately distinguish trailers traveling close to a trailer that is the subject of passage determination from the trailer that is the subject of passage determination.

[0227] <Technology 4> In the passage determination device described in any one of Techniques 1 to 3, the processor acquires second captured images obtained by repeatedly capturing an area including the passage determination position from behind in the trailer's traveling direction, and performs processing to detect the trailer in the second captured images. If a trailer is not detected from the first captured image taken within the first hour from the first time, but a trailer is detected from the second captured image taken within the second hour from the first time, it is determined that the trailer detected from the first captured image at the first time and the trailer detected from the second captured image are the same trailer, and if a trailer is not detected from the second captured image taken within the second hour from the first time, it is determined that the trailer detected from the first captured image at the first time and the trailer detected from the second captured image after the second time has elapsed are different trailers from the trailer detected from the first captured image at the first time and the trailer detected from the second captured image.

[0228] As a result, the passage determination device according to the present embodiment can determine whether the trailers detected in the first and second captured images are the same or different when there is a trailer traveling close to the trailer to be determined to be passed. As a result, the passage determination device can manage which captured image shows the trailer that has been determined to be passed. In other words, the passage determination device can store and manage the image of the trailer that has been determined to be passed in a database without mixing it with images of different trailers. For example, by detecting the ID of the trailer from multiple images including trailers determined to be the same, the ID of the trailer can be detected more accurately even if the ID is written in a position that can only be seen from one of the images. In particular, since the first captured image captured by the front camera and the second captured image captured by the rear camera are captured at different angles, the ID may be detected from the other captured image even if the ID cannot be detected from one captured image. In addition, the passing determination device can manage images taken by the front camera and the rear camera containing the same trailer as images containing the same trailer, so that when a user wishes to check images of the same trailer, images taken from a variety of viewpoints can be extracted all at once, improving user convenience.

[0229] <Technology 5> The passage determination device according to any one of Techniques 1 to 4 further includes a database, and the processor detects identification information of the trailer and stores the identification information of the trailer determined to have passed the passage determination position in the database.

[0230] In this way, the passage determination device according to the present embodiment can construct a database that stores information about trailers that exist within distribution center 100.

[0231] <Technology 6> The passing determination device of this embodiment includes a processor that acquires captured images from a camera that photographs the trailer at predetermined time intervals, sets coordinates in the captured images in a direction corresponding to the trailer's direction of travel, calculates a first coordinate of the trailer, and determines that the trailer has passed the passing determination position if the amount of change between the calculated first coordinate and a second coordinate of the trailer calculated from the most recently acquired captured image is greater than or equal to a predetermined threshold.

[0232] As a result, the passage determination device according to the present embodiment can distinguish between a plurality of trailers traveling in close proximity with high accuracy.

[0233] <Technology 7> In the passage determination device described in Technique 6, when a tractor is detected between the first coordinate and the second coordinate, it is determined that the trailer has passed the passage determination position.

[0234] As a result, even if the passage determination device according to the present embodiment erroneously detects a trailer at a position where the X coordinate increases, it is possible to prevent erroneous judgment that a trailer has passed by performing the passage determination process based on the detection of the tractor.

[0235] Although the embodiments have been described above with reference to the accompanying drawings, the present disclosure is not limited to such examples. It is clear that a person skilled in the art can conceive of various modifications, corrections, substitutions, additions, deletions, and equivalents within the scope of the claims, and it is understood that these also belong to the technical scope of the present disclosure. In addition, the components in the above-mentioned embodiments may be arbitrarily combined within the scope of the invention. [Industrial Applicability]

[0236] The technology disclosed herein is useful as a passage determination device, a passage determination system, and a passage determination method that can distinguish between trailers traveling in close proximity with high accuracy. [Explanation of symbols]

[0237] TR Trailer LE Tractor ST1,ST2,ST3,ST4,ST5,ST6,ST7,ST8 String 1. Character Recognition System 2 Passage Judgment System 10. Shooting Devices 20,20A Database 30,30A Display Device 40 Character recognition device 41,51 Processor 41A String detection section 41B Character string recognition section 41C Attribute determination section 41D Trailer ID Estimation Unit 41E Tractor detector 41F Passage judgment section 41G Database Update Division 42 Communication I / F 43 Memory CA1 Front Camera CA2 Rear Camera P1 1st Hour P2 2nd Hour

Claims

1. A first photographed image is obtained by repeatedly photographing an area including a passing determination position of the trailer from the front in the traveling direction of the trailer; performing a process of detecting the trailer in the first captured image; a processor that, when the trailer is detected from the first photographed image photographed at a first time and the trailer is not detected from the first photographed image photographed within a first hour from the first time, determines that the trailer detected from the first photographed image photographed at the first time has passed a passing determination position; Passage determination device.

2. The processor further comprises: A second photographed image is obtained by repeatedly photographing an area including the passage determination position from the rear in the traveling direction of the trailer; performing a process of detecting the trailer in the second captured image; When the trailer is detected from the first photographed image taken within a first hour from the first time at which the trailer is detected from the first photographed image, If the trailer is not detected from the second photographed image photographed at a time later than the first time, the trailer detected from the first photographed image photographed at the first time has not passed through the passage determination position, If the trailer is detected from the second photographed image photographed at a time later than the first time, it is determined that the trailer detected from the first photographed image photographed at the first time has passed through the passage determination position. The passage determination device according to claim 1 .

3. The processor, determining that the trailer detected from the first photographed image taken at a time before a second time, which is the photographing time of the second photographed image at which the trailer was detected, and the trailer detected from the first photographed image taken after the second time are different trailers; The passage determination device according to claim 2 .

4. The processor, A second photographed image is obtained by repeatedly photographing an area including the passage determination position from the rear in the traveling direction of the trailer; performing a process of detecting the trailer in the second captured image; If the trailer is not detected from the first captured image captured within a first hour from the first time, If the trailer is detected from a second captured image captured within a second hour from the first time, it is determined that the trailer detected from the first captured image at the first time and the trailer detected from the second captured image are the same trailer; The passing determination device of claim 1, wherein if the trailer is not detected from the second captured image taken within the second hour from the first time, it is determined that the trailer detected from the first captured image or the second captured image after the second time has elapsed is a different trailer from the trailer detected from the first captured image at the first time and the trailer detected from the second captured image.

5. Further equipped with a database, The processor, Detecting the trailer's identification information; storing the identification information of the trailer determined to have passed the passage determination position in the database; The passage determination device according to claim 1 .

6. a front camera that repeatedly captures an area including a trailer passage determination position from the front in the trailer's traveling direction; A processor that acquires a first captured image from the front camera, The processor, performing a process of detecting the trailer in the first captured image; if the trailer is detected from the first photographed image photographed at a first time and the trailer is not detected from the first photographed image photographed within a first hour from the first time, it is determined that the trailer detected from the first photographed image photographed at the first time has passed through a passing determination position; Passing judgment system.

7. A first photographed image is obtained by repeatedly photographing an area including a passing determination position of the trailer from the front in the traveling direction of the trailer; performing a process of detecting the trailer in the first captured image; if the trailer is detected from the first photographed image photographed at a first time and the trailer is not detected from the first photographed image photographed within a first hour from the first time, it is determined that the trailer detected from the first photographed image photographed at the first time has passed through a passing determination position; Passage judgment method.

8. Acquiring images of the trailer from a camera at predetermined intervals; A coordinate is set in the captured image in a direction corresponding to the traveling direction of the trailer; Calculating a first coordinate of the trailer; a processor that determines that the trailer has passed a passing determination position when a change amount between the calculated first coordinate and a second coordinate of the trailer calculated from a captured image acquired immediately before is equal to or greater than a predetermined threshold value; Passage determination device.

9. When the processor detects a tractor between the first coordinate and the second coordinate, the processor determines that the trailer has passed through the passage determination position. The passage determination device according to claim 8.

10. A camera to shoot the trailer, A processor for acquiring a captured image from the camera, The processor, acquiring the photographed images from the camera at predetermined time intervals; A coordinate is set in the captured image parallel to the traveling direction of the trailer; Calculating a first coordinate of the trailer; When a change amount between the calculated first coordinate and a second coordinate of the trailer calculated from a captured image acquired immediately before is equal to or greater than a predetermined threshold, it is determined that the trailer has passed through a passing determination position. Passing judgment system.

11. Acquiring images of the trailer from a camera at predetermined intervals; A coordinate is set in the captured image parallel to the traveling direction of the trailer; Calculating a first coordinate of the trailer; When a change amount between the calculated first coordinate and a second coordinate of the trailer calculated from a captured image acquired immediately before is equal to or greater than a predetermined threshold, it is determined that the trailer has passed through a passing determination position. Passage judgment method.

Images