Vehicle maintenance system and vehicle maintenance method

Abstract

To provide a vehicle maintenance system capable of suppressing a reduction in inspection accuracy for a vehicle compared to a conventional system.SOLUTION: A vehicle maintenance system includes a vehicle state information acquisition unit 301 that acquires vehicle state information indicating a state of a vehicle, a passenger existence estimation unit 302 that compares the acquired vehicle state information with predetermined information indicating timing when no passenger is in the vehicle to estimate the presence / absence of a passenger in the vehicle, an image data acquisition unit 303 that, when it is estimated that no passenger is in the vehicle, acquires image data indicating an image of the vehicle captured by an imaging device, and an inspection unit 304 that inspects the vehicle on the basis of the acquired image data and reference image data indicating an image of the vehicle preliminary captured by the imaging device in a state where no abnormality is in the vehicle.SELECTED DRAWING: Figure 5

Description

[Technical Field]

[0001] The present disclosure relates to a vehicle maintenance system and a vehicle maintenance method. [Background technology]

[0002] Patent Document 1 discloses a measuring device that calculates the current and future wear amounts of a measurement object. The measurement object is, for example, a railway vehicle (hereinafter simply referred to as a "vehicle") and the pantograph slider of the vehicle. This measuring device calculates current shape information of the measurement object from an image of the measurement object taken by, for example, a stereo camera, and calculates an estimated future wear amount of the measurement object by comparing the calculated shape information with past shape information of the measurement object. The measuring device then compares the calculated estimated wear amount with a wear amount threshold, and performs an error determination if the calculated estimated wear amount is greater than the wear amount threshold. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Laid-Open No. 2016-191648 Summary of the Invention [Problem to be solved by the invention]

[0004] One possible approach to automating vehicle maintenance work is to photograph the vehicle using an imaging device, such as a security camera, which is increasingly being installed inside the passenger compartment of a vehicle, and then analyze the captured images using a computer to inspect the vehicle. In this case, if the imaging device photographs the interior of the vehicle when passengers are present, the passengers may partially obstruct the image capture. Therefore, in this case, it is desirable for the imaging device to photograph the vehicle when there are no passengers in the passenger compartment. However, the measurement device described in Patent Document 1 does not calculate the estimated future wear amount of the measurement target using images captured by a stereo camera, taking into account the presence or absence of passengers in the vehicle. Therefore, if the measurement device described in Patent Document 1 is used in vehicle inspections, the accuracy of the vehicle inspection may be reduced.

[0005] The present disclosure has been made to solve the above-mentioned problems, and aims to provide a vehicle maintenance system that can suppress a decrease in vehicle inspection accuracy compared to conventional systems. [Means for solving the problem]

[0006] The vehicle maintenance system according to the present disclosure includes vehicle status information indicating the status of a vehicle. Vehicle status information is distributed while the vehicle is traveling and includes information indicating the status of the vehicle. a vehicle state information acquisition unit that acquires the vehicle state information acquired by the vehicle state information acquisition unit and a preset While driving Compared with information indicating when the vehicle is empty, While driving The vehicle is characterized by comprising a passenger presence / absence estimation unit that estimates whether there are passengers in the vehicle, an image data acquisition unit that acquires image data showing an image of the vehicle taken by an imaging device when the passenger presence / absence estimation unit estimates that there are no passengers in the vehicle, and an inspection unit that inspects the vehicle based on the image data acquired by the image data acquisition unit and reference image data showing an image of the vehicle taken by the imaging device when there are no abnormalities in the vehicle. [Effects of the Invention]

[0007] According to the present disclosure, the above-described configuration makes it possible to suppress a decrease in vehicle inspection accuracy compared to conventional methods. [Brief explanation of the drawings]

[0008] [Figure 1] 1 is a diagram illustrating an example of the configuration of a vehicle maintenance system according to a first embodiment. [Figure 2] FIG. 10 is a diagram showing an example of the arrangement of an IP camera, a recording device, and an analysis device when viewed from above inside a guest room. [Figure 3] 4 is a diagram showing an example of vehicle state information stored in a vehicle state DB according to the first embodiment. FIG. [Figure 4] 4 is a diagram showing an example of reference image data and related information stored in a reference condition DB according to the first embodiment. FIG. [Figure 5] 1 is a functional block diagram of a vehicle maintenance system according to a first embodiment. [Figure 6] FIG. 10 is a diagram showing an example in which background subtraction is used for comparison and analysis in the first embodiment. [Figure 7] FIG. 4 is a diagram showing an example of notification of an abnormality location in the first embodiment. [Figure 8] FIG. 4 is a diagram showing an example of notification of an abnormality location in the first embodiment. [Figure 9] 4 is a flowchart showing an example of the operation of the vehicle maintenance system according to the first embodiment. [Figure 10] 10A and 10B are diagrams illustrating an example of a hardware configuration of the analysis device according to the first embodiment. [Figure 11] FIG. 10 is a diagram illustrating an example of the configuration of a vehicle maintenance system according to a second embodiment. [Figure 12] FIG. 11 is a diagram showing an example of reference sensor data and related information stored in a reference state DB according to the second embodiment. [Figure 13] FIG. 10 is a functional block diagram of a vehicle maintenance system according to a second embodiment. [Figure 14] FIG. 10 is a diagram showing an example in which background subtraction is used for comparison and analysis in the second embodiment. [Figure 15] 10 is a flowchart showing an example of the operation of the vehicle maintenance system according to the second embodiment. [Figure 16]FIG. 10 is a diagram illustrating an example of the configuration of a vehicle maintenance system according to a third embodiment. [Figure 17] 13 is a diagram showing an example of reference image data and related information stored in a reference condition group DB according to the third embodiment. FIG. [Figure 18] FIG. 11 is a diagram showing an example of organization information stored in an organization information DB in the third embodiment. [Figure 19] FIG. 10 is a functional block diagram of a vehicle maintenance system according to a third embodiment. [Figure 20] FIG. 11 is a diagram showing an example in which background subtraction is used for comparison and analysis in the third embodiment. [Figure 21] 11 is a flowchart showing an example of the operation of the vehicle maintenance system according to the third embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments will be described in detail with reference to the drawings.

[0010] Embodiment 1 Fig. 1 is a diagram showing an example of the configuration of a vehicle maintenance system according to embodiment 1. As shown in Fig. 1, the vehicle maintenance system includes IP (Internet Protocol) cameras 1 to 4, a recording device 5, a vehicle state information management device 10, a monitor device 20, an analyzer 30, a vehicle state database (DB) 31, a reference state database (DB) 32, wireless devices 40 to 41, and a terminal 50. Of these, the wireless device 41 and the terminal 50 are installed at ground bases (for example, a vehicle depot or a control center), and the devices other than the wireless device 41 and the terminal 50 are installed in the vehicle.

[0011] In the first embodiment, the vehicle is a railway vehicle, and the vehicle is a single formation made up of one or more cars (for example, cars 1 to 10).

[0012] IP cameras 1-4 are installed in each car of the train, and periodically capture images of the interior of each car (hereinafter simply referred to as "interior of the passenger compartment"). For example, IP cameras 1-4 installed in car 1 periodically capture images of the interior of car 1. Furthermore, IP cameras 1-4 installed in car 2 periodically capture images of the interior of car 2. IP cameras 1-4 are communicably connected to recording device 5, and output data showing images captured of the interior of each car (hereinafter simply referred to as "image data") to recording device 5 as needed.

[0013] IP cameras 1 to 4 are installed in positions such as shown in FIG. 2 so that they can capture the entire interior of each car. FIG. 2 shows a top view of a certain car. In FIG. 2, the car moves, for example, from the right end to the left end. IP cameras 1 to 4 are installed alternately on the left and right sides along the direction of travel of the car, on the lintel K above the door of each car, as shown in FIG. 2. This allows IP cameras 1 to 4 to capture the entire interior of each car. In FIG. 2, the straight lines extending from each IP camera indicate the capture range of each IP camera.

[0014] For example, one or more recording devices 5 are installed in a vehicle (see also FIG. 2). The recording device 5 records image data output from IP cameras 1 to 4 installed in each vehicle for a certain period of time. In addition, the recording device 5 distributes the recorded image data to another device (for example, analysis device 30) in response to a request from the other device.

[0015] For example, one vehicle state information management device 10 is installed in a vehicle. The vehicle state information management device 10 manages information indicating the vehicle state. Here, the information indicating the vehicle state includes, for example, information indicating the current distance traveled by the vehicle, information indicating the vehicle's next stop, and information indicating the vehicle's running state, and in particular refers to information that makes it possible to estimate whether or not there are passengers in the vehicle, including the passenger compartment. Note that the running state of the vehicle refers to, for example, whether the vehicle is in out-of-service operation or in commercial operation. Hereinafter, for ease of explanation, the information indicating the vehicle state managed by the vehicle state information management device 10 will be referred to as "vehicle state information."

[0016] The vehicle state information management device 10 acquires the above-mentioned vehicle state information by a known method and manages the acquired vehicle state information. The vehicle state information management device 10 also distributes the acquired vehicle state information to other devices (e.g., the analysis device 30) at regular intervals.

[0017] For example, one vehicle status DB 31 is installed in each vehicle. The vehicle status DB 31 stores information indicating when no passengers are present in the vehicle. For ease of explanation, the information indicating when no passengers are present in the vehicle will be referred to as "absence information" below. The absence information is created in advance by an administrator of the vehicle maintenance system and is stored in the vehicle status DB 31 using, for example, the analysis device 30.

[0018] The absence information stored in the vehicle status DB 31 is configured to include, for example, items similar to the above-mentioned vehicle status information. An example of the absence information is shown in FIG. 3. As shown in FIG. 3, the absence information is configured to include, for example, three items similar to the above-mentioned vehicle status information, namely, the vehicle's current kilometer range, the vehicle's next stop, and the vehicle's running status. In the example of FIG. 3, the absence information is identified by one or more of these three items.

[0019] For example, the first item of absence information shown in Figure 3 indicates that there are no passengers in the vehicle when the vehicle is in out-of-service operation, regardless of the vehicle's current kilometer distance and next stop. The second item of absence information shown in Figure 3 indicates that there are no passengers in the vehicle when the vehicle's next stop is the starting station, regardless of the vehicle's current kilometer distance and running state. The third item of absence information shown in Figure 3 indicates that there are no passengers in the vehicle when the vehicle is running within a specific kilometer range (100.0 to 120.0 in the example of Figure 3), even when the vehicle is in commercial operation.

[0020] Note that the vehicle status information and unattended information exemplified above are merely examples, and can be created based on various other conditions. Furthermore, the three items included in the vehicle status information and unattended information exemplified above are merely examples, and items other than these may be included in the vehicle status information and unattended information. For example, the vehicle status information may include the vehicle's current kilometer range, the vehicle's next stop, and the vehicle's running status, as well as the vehicle's running time zone. Similarly, the unattended information may include the vehicle's current kilometer range, the vehicle's next stop, and the vehicle's running status, as well as the vehicle's running time zone.

[0021] For example, one reference state DB 32 is installed in each vehicle. The reference state DB 32 stores images of the interior of the passenger compartment of each car taken by the IP cameras 1 to 4 in a state where it has been confirmed in advance that there are no abnormalities in the vehicle, including the interior of the passenger compartment. Here, a state where it has been confirmed in advance that there are no abnormalities in the vehicle refers to, for example, a state immediately after a maintenance inspection of the vehicle or immediately after a new vehicle has been put into service.

[0022] For ease of explanation, the state in which it has been confirmed in advance that there are no abnormalities in the vehicle will be referred to as the "reference state." In addition, in the reference state, the images of the interior of each car taken by IP cameras 1 to 4 will be referred to as "reference images," and the data representing the reference images will be referred to as "reference image data."

[0023] Furthermore, the reference state DB 32 stores information related to the reference image in association with the reference image data. The information related to the reference image is, for example, a number indicating the train formation (train formation number) of the vehicle from which the reference image was taken, the car number from which the reference image was taken, the number of the IP camera from which the reference image was taken, and the date and time when the reference image data was saved (updated) in the reference state DB 32. Hereinafter, for ease of explanation, information related to the reference image will be referred to as "related information."

[0024] An example of reference image data and related information stored in the reference state DB 32 is shown in Figure 4. As shown in Figure 4, the reference image data is saved with a file name such as "A001-1-1.jpg." Furthermore, according to the related information associated with this reference image data, it can be seen that this reference image was taken by IP camera 1 installed in car No. 1 of a vehicle with the formation number "A001," and that the data was saved in the reference state DB 32 at "2021 / 12 / 10 23:10."

[0025] Similarly, the reference image data is saved with a file name such as "A001-1-2.jpg." Furthermore, according to the related information associated with this reference image data, it is clear that this reference image was taken by IP camera 2 installed in car No. 1 of a vehicle with the train formation number "A001," and that the data was saved in reference state DB 32 at "2021 / 12 / 3 23:00."

[0026] The reference image data and related information are created in advance by the administrator of the vehicle maintenance system based on images taken by IP cameras 1 to 4 when the vehicle is in a reference state, and are stored in the reference state DB 32 using, for example, the analysis device 30.

[0027] The file name of the reference image data exemplified above is merely an example, and other file names may be used. The related information exemplified above is also an example, and may include information other than the train formation number, car number, IP camera number, and storage date and time of the reference image data.

[0028] For example, one analysis device 30 is installed in a vehicle (see also FIG. 2). The analysis device 30 inspects the vehicle based on image data captured by the IP cameras 1 to 4 and the reference image data and related information stored in the reference condition DB 32.

[0029] Specifically, the analysis device 30 acquires vehicle state information periodically distributed from the vehicle state information management device 10. Next, the analysis device 30 compares the acquired vehicle state information with the unoccupied state information stored in the vehicle state DB 31. Then, based on the result of the comparison, the analysis device 30 estimates whether the timing at which the vehicle state information was acquired was a timing at which there were no passengers in the vehicle, including the passenger compartments of each car.

[0030] If the analysis device 30 estimates that the timing at which the vehicle status information was acquired was a timing at which no passengers were in the vehicle, it sends a request to the recording device 5 to acquire image data captured by the IP cameras 1 to 4 of each car, which is recorded in the recording device 5. Then, the analysis device 30 acquires the image data captured by the IP cameras 1 to 4 of each car, which is distributed from the recording device 5 in response to this acquisition request.

[0031] Then, the analysis device 30 compares and analyzes the image data acquired from the recording device 5 with the reference image data stored in the reference condition DB 32 using a method such as background subtraction, and obtains an analysis result indicating whether or not there is an abnormality in the vehicle, including the passenger compartment of each car.

[0032] Furthermore, analysis device 30 transmits data indicating the obtained analysis results (hereinafter also referred to as "analysis result data") to monitor device 20 as the vehicle inspection results. Analysis device 30 also transmits the obtained analysis result data as the vehicle inspection results to terminal 50 via wireless devices 40 and 41. The analysis method used by analysis device 30 will be described in detail later.

[0033] One monitor device 20 is installed in a vehicle, for example. The monitor device 20 receives the analysis result data transmitted from the analysis device 30, and displays the analysis results indicated by the received analysis result data on a screen.

[0034] For example, one wireless device 40 is installed in a vehicle, and wireless device 40 performs wireless communication with wireless device 41 installed at a ground base.

[0035] For example, one wireless device 41 is installed at a ground base, and wireless device 41 performs wireless communication with wireless device 40 installed in a vehicle.

[0036] For example, one terminal 50 is installed at a ground base. The terminal 50 communicates with a device (e.g., analysis device 30) installed in a vehicle via wireless devices 40 and 41. For example, the terminal 50 receives analysis result data transmitted from the analysis device 30 via the wireless devices 40 and 41, and displays the analysis result indicated by the received analysis result data on a monitor device (not shown) provided in the terminal 50.

[0037] Next, a functional block diagram of the vehicle maintenance system according to the first embodiment will be described with reference to FIG.

[0038] As shown in Figure 5, the vehicle maintenance system includes a vehicle condition information acquisition unit 301, a passenger presence / absence estimation unit 302, an image data acquisition unit 303, an inspection unit 304, an inspection result notification unit 305, a vehicle condition memory unit 310, and a reference condition memory unit 320.

[0039] Of these, the vehicle state information acquisition unit 301, passenger presence / absence estimation unit 302, image data acquisition unit 303, inspection unit 304, and inspection result notification unit 305 are included in, for example, the analysis device 30. The vehicle state storage unit 310 is configured, for example, by a vehicle state DB 31. The reference state storage unit 320 is configured, for example, by a reference state DB 32.

[0040] The vehicle state information acquisition unit 301 acquires the vehicle state information distributed periodically from the vehicle state information management device 10 .

[0041] The passenger presence / absence estimation unit 302 compares the vehicle state information acquired by the vehicle state information acquisition unit 301 with the unoccupied information stored in the vehicle state storage unit 310 (vehicle state DB 31). Then, based on the result of the comparison, the passenger presence / absence estimation unit 302 estimates whether or not the timing at which the vehicle state information was acquired was a timing at which no passengers were present in the vehicle, including the passenger compartments of each car. In other words, the passenger presence / absence estimation unit 302 estimates the presence or absence of passengers at the timing at which the vehicle state information was acquired.

[0042] 3, assume that the unoccupied state information stored in vehicle state DB 31 includes information indicating that the vehicle's running state is "in deadhead operation," and the vehicle state information acquired by vehicle state information acquisition unit 301 indicates that the vehicle's running state is in deadhead operation. In this case, since the unoccupied state information includes the vehicle state information, passenger presence / absence estimation unit 302 estimates that there were no passengers in the vehicle at the time the vehicle state information was acquired.

[0043] 3, for example, assume that the absence information stored in vehicle state DB 31 includes information indicating that the vehicle's next stop is the "starting station," and the vehicle state information acquired by vehicle state information acquisition unit 301 indicates that the vehicle's next stop is the starting station. In this case, too, passenger presence / absence estimation unit 302 estimates that there were no passengers in the vehicle at the time the vehicle state information was acquired, because the absence information includes the vehicle state information.

[0044] On the other hand, in the above example, suppose that the vehicle status information acquired by the vehicle status information acquisition unit 301 indicates that the next stop of the vehicle is other than the starting station. In this case, the passenger presence / absence estimation unit 302 estimates that there is a passenger in the vehicle at the time the vehicle status information is acquired, because the vehicle status information is not included in the absence information.

[0045] When the passenger presence estimation unit 302 estimates that there are no passengers in the vehicle at the time the vehicle status information is acquired, the image data acquisition unit 303 transmits to the recording device 5 a request to acquire image data captured by the IP cameras 1 to 4 in each car that is recorded in the recording device 5. Then, the image data acquisition unit 303 acquires the image data captured by the IP cameras 1 to 4 in each car that is distributed from the recording device 5 in response to this acquisition request. At this time, the image data acquisition unit 303 may acquire multiple pieces of image data from the recording device 5.

[0046] The inspection unit 304 inspects the vehicle based on the image data acquired by the image data acquisition unit 303 and the reference image data stored in the reference state storage unit 320 (reference state DB 32). Specifically, the inspection unit 304 compares and analyzes the image data and the reference image data using a background subtraction method to obtain an analysis result indicating whether or not there is an abnormality in the vehicle. Furthermore, if the inspection unit 304 obtains an analysis result indicating there is an abnormality in the vehicle, it identifies the difference between the image data and the reference image data as the abnormality location.

[0047] Here, an example of an inspection method by the inspection unit 304 will be described. Here, as an example, an example will be described in which the inspection unit 304 compares and analyzes image data captured by the IP cameras 1 to 4 of car No. 1 with reference image data stored in the reference condition DB 32 using a background subtraction method. Also, it is assumed that the reference image data and related information stored in the reference condition DB 32 are in the state shown in FIG. 4.

[0048] In this case, for example, the inspection unit 304 compares and analyzes, by background subtraction, the image data captured by the IP camera 1 in car No. 1 with the reference image data (file name "A001-1-1.jpg") captured by the IP camera 1 in car No. 1, which is stored in the reference condition DB 32. Then, the inspection unit 304 obtains an analysis result indicating whether or not there is an abnormality within the range captured by the IP camera 1 in the passenger compartment of car No. 1.

[0049] For example, in Fig. 6, reference numeral 601 denotes a reference image taken by IP camera 1 on car No. 1. Reference numeral 602 denotes an image taken by IP camera 1 on car No. 1. Image 602 captures abnormality E in two locations. Reference numeral 603 denotes the results of comparing and analyzing reference image 601 and image 602 by background subtraction.

[0050] 6, when anomaly E is captured in image 602, the anomaly E is extracted as the difference between image 602 and reference image 601, as shown in result 603. When the difference between image 602 and reference image 601 is extracted in this manner, inspection unit 304 obtains an analysis result indicating that an anomaly exists within the passenger compartment of car No. 1 within the range of photography by IP camera 1. Furthermore, inspection unit 304 identifies the location where anomaly E has occurred based on comparison and analysis result 603.

[0051] On the other hand, if no difference is extracted between the image 602 and the reference image 601, the inspection unit 304 obtains an analysis result indicating that no abnormality is found within the range of the IP camera 1 in the passenger compartment of the first car.

[0052] Similarly, the inspection unit 304 compares and analyzes the image data captured by IP camera 2 in car No. 1 with the reference image data (file name "A001-1-2.jpg") captured by IP camera 2 in car No. 1, which is stored in the reference state DB 32, using background difference, and obtains an analysis result indicating whether or not there is an abnormality within the range captured by IP camera 2 within the passenger compartment of car No. 1.

[0053] In addition, the inspection unit 304 compares and analyzes the image data captured by the IP camera 3 in car No. 1 with the reference image data (file name "A001-1-3.jpg") captured by the IP camera 3 in car No. 1, which is stored in the reference state DB 32, using background difference, and obtains an analysis result indicating whether or not there is an abnormality within the passenger compartment of car No. 1 within the range captured by the IP camera 3.

[0054] In addition, the inspection unit 304 compares and analyzes the image data captured by the IP camera 4 in car No. 1 with the reference image data (file name "A001-1-4.jpg") captured by the IP camera 4 in car No. 1, which is stored in the reference state DB 32, using background difference, and obtains an analysis result indicating whether or not there is an abnormality within the range captured by the IP camera 4 within the passenger compartment of car No. 1.

[0055] Similarly, the inspection unit 304 compares and analyzes the image data captured by each IP camera in each car with the reference image data captured by each IP camera in each car, which is stored in the reference state DB32, using background difference, and obtains an analysis result indicating whether or not there is an abnormality within the passenger compartment of each car within the range captured by each IP camera.

[0056] The inspection unit 304 may obtain the analysis results for each car and for each IP camera, or for each individual car. When obtaining the analysis results for each car, the inspection unit 304 obtains an analysis result indicating that there is no abnormality for a certain car only when there is no abnormality within the range of photography by all IP cameras installed in that car, for example.

[0057] Furthermore, the inspection unit 304 may obtain the analysis result on a vehicle-by-vehicle basis. When obtaining the analysis result on a vehicle-by-vehicle basis, the inspection unit 304 obtains the analysis result indicating that there is no abnormality in the vehicle only when there is no abnormality in, for example, all of the cars of the vehicle.

[0058] The inspection unit 304 may use other methods, such as machine learning, in addition to background subtraction as a comparison and analysis method. For example, when the inspection unit 304 uses machine learning as a comparison and analysis method, a model created in advance by machine learning may be stored in the reference state DB 32 instead of or in addition to the reference image data described above. This model is a model that is trained to receive image data captured by each IP camera in each car and output whether or not there is an abnormality in the passenger compartment shown in the image represented by this image data. The inspection unit 304 then inputs the image data captured by each IP camera in each car into this model and obtains the presence or absence of an abnormality output from this model as an analysis result.

[0059] The inspection result notification unit 305 transmits the analysis result data obtained by the inspection unit 304 to the monitor device 20 as the inspection result of the vehicle. As a result, the inspection result notification unit 305 notifies the conductor or the like of the inspection result. Furthermore, the inspection result notification unit 305 transmits the analysis result data obtained by the inspection unit 304 as the inspection result of the vehicle to the terminal 50 via the wireless devices 40 and 41. As a result, the inspection result notification unit 305 notifies the maintenance personnel or the like at the ground base of the analysis result.

[0060] In addition, when the inspection unit 304 obtains an analysis result indicating that there is an abnormality in the vehicle and identifies the abnormal location, the inspection result notification unit 305 includes the location identified as the abnormal location by the inspection unit 304 in the inspection result and transmits it to the monitor device 20 and the terminal 50.

[0061] 7, the inspection result notification unit 305 may generate data showing an image in which a rectangular frame 701 indicating the abnormality location and a message 702 indicating the details of the abnormality are displayed, and transmit the generated data as the inspection result to the monitor device 20, etc. Note that the inspection result notification unit 305 may store, for example, several types of messages 702 indicating the details of the abnormality in advance, and may select an appropriate one according to the analysis results obtained by the inspection unit 304.

[0062] In addition, when transmitting the inspection results, the inspection result notification unit 305 may generate data showing a diagram in which the frame 701 and message 702 are displayed on a schematic diagram 801 showing the interior of the guest room, as shown in Figure 8, and transmit the generated data to the monitor device 20, etc. as the inspection results.

[0063] Furthermore, when transmitting the inspection results, the inspection result notification unit 305 may generate text data indicating the inspection results and transmit the generated text data to the monitor device 20, etc. For example, the inspection result notification unit 305 may generate text data indicating the location and type of abnormality, or text data indicating that there are no abnormalities in the guest room, and transmit the generated text data to the monitor device 20, etc.

[0064] In the above description, an example has been described in which the vehicle maintenance system includes the inspection result notification unit 305. However, the inspection result notification unit 305 is not an essential component and may be omitted.

[0065] Next, an example of the operation of the vehicle maintenance system according to the first embodiment will be described with reference to the flowchart shown in FIG.

[0066] First, the vehicle state information acquisition unit 301 acquires the vehicle state information distributed periodically from the vehicle state information management device 10 (step ST1).

[0067] Next, the passenger presence / absence estimation unit 302 compares the vehicle state information acquired by the vehicle state information acquisition unit 301 in step ST1 with the unoccupied information stored in the vehicle state storage unit 310 (vehicle state DB 31). Then, based on the result of the comparison, the passenger presence / absence estimation unit 302 estimates whether or not there is a passenger in the vehicle at the time the vehicle state information was acquired (step ST2). As a result, if it is estimated that there is no passenger in the vehicle (step ST2; Yes), the process proceeds to step ST3. On the other hand, if it is estimated that there is a passenger in the vehicle (step ST2; No), the process returns to step ST1.

[0068] In step ST3, the image data acquisition unit 303 transmits to the recording device 5 a request to acquire image data captured by the IP cameras 1 to 4 of each car, which is recorded in the recording device 5. Then, the image data acquisition unit 303 acquires the image data captured by the IP cameras 1 to 4 of each car, which is distributed from the recording device 5 in response to this acquisition request (step ST3).

[0069] Next, the inspection unit 304 compares and analyzes the image data acquired by the image data acquisition unit 303 in step ST3 with the reference image data stored in the reference state storage unit 320 (reference state DB 32) using a background subtraction method, and inspects the vehicle including the interior of each car (step ST4). Furthermore, if the inspection unit 304 obtains an analysis result indicating an abnormality, it identifies the difference between the image data and the reference image data as the abnormal location.

[0070] Next, the inspection result notification unit 305 checks whether the analysis result obtained by the inspection unit 304 in step ST4 indicates that there is an abnormality in the vehicle (step ST5).

[0071] As a result, if the analysis result indicates that there is an abnormality in the vehicle (step ST5; Yes), the inspection result notification unit 305 transmits analysis result data indicating that there is an abnormality in the vehicle as the inspection result to the monitor device 20 and the terminal 50. At this time, the inspection result notification unit 305 also transmits the inspection result together with the abnormality location identified by the inspection unit 304 (step ST6).

[0072] On the other hand, if the above analysis result indicates that there is no abnormality in the vehicle (step ST5; No), the inspection result notification unit 305 transmits the analysis result data indicating that there is no abnormality in the vehicle to the monitor device 20 and the terminal 50 as the inspection result (step ST7).

[0073] Next, an example of the hardware configuration of the analysis device 30 according to the first embodiment will be described with reference to FIG. The functions of the vehicle state information acquisition unit 301, passenger presence / absence estimation unit 302, image data acquisition unit 303, inspection unit 304, and inspection result notification unit 305 in the analysis device 30 are realized by a processing circuit. The processing circuit may be dedicated hardware as shown in Fig. 10A, or may be a CPU (also referred to as a central processing unit, processing unit, arithmetic unit, microprocessor, microcomputer, processor, or DSP (Digital Signal Processor)) 52 that executes a program stored in memory 53 as shown in Fig. 10B.

[0074] When the processing circuit is dedicated hardware, the processing circuit 51 corresponds to, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or a combination thereof. The functions of each of the vehicle state information acquisition unit 301, the passenger presence / absence estimation unit 302, the image data acquisition unit 303, the inspection unit 304, and the inspection result notification unit 305 may be realized by the processing circuit 51 individually, or the functions of each unit may be realized by the processing circuit 51 collectively.

[0075] When the processing circuit is a CPU 52, the functions of the vehicle state information acquisition unit 301, passenger presence / absence estimation unit 302, image data acquisition unit 303, inspection unit 304, and inspection result notification unit 305 are realized by software, firmware, or a combination of software and firmware. The software and firmware are written as programs and stored in the memory 53. The processing circuit realizes the functions of each unit by reading and executing the programs stored in the memory 53. That is, the analysis device 30 includes a memory for storing a program that, when executed by the processing circuit, results in the execution of, for example, each step shown in FIG. 9 . Furthermore, these programs can also be said to cause a computer to execute the procedures and methods of the vehicle state information acquisition unit 301, passenger presence / absence estimation unit 302, image data acquisition unit 303, inspection unit 304, and inspection result notification unit 305. Here, examples of memory 53 include non-volatile or volatile semiconductor memory such as RAM (Random Access Memory), ROM (Read Only Memory), flash memory, EPROM (Erasable Programmable ROM), EEPROM (Electrically EPROM), magnetic disk, flexible disk, optical disk, compact disk, mini disk, or DVD (Digital Versatile Disc).

[0076] It should be noted that the functions of the vehicle state information acquisition unit 301, passenger presence / absence estimation unit 302, image data acquisition unit 303, inspection unit 304, and inspection result notification unit 305 may be partially realized by dedicated hardware and partially realized by software or firmware. For example, the function of the vehicle state information acquisition unit 301 may be realized by a processing circuit as dedicated hardware, and the functions of the passenger presence / absence estimation unit 302, image data acquisition unit 303, inspection unit 304, and inspection result notification unit 305 may be realized by the processing circuit reading and executing programs stored in the memory 53.

[0077] Thus, the processing circuitry can implement each of the above-described functions by hardware, software, firmware, or a combination thereof.

[0078] As described above, in the vehicle maintenance system according to the first embodiment, the vehicle state information acquisition unit 301 acquires vehicle state information periodically distributed from the vehicle state information management device 10, and the passenger presence / absence estimation unit 302 estimates whether or not there are passengers in the vehicle, including the passenger compartment, at the time the vehicle state information is acquired, based on the acquired vehicle state information. If the inspection unit 304 estimates that there are no passengers in the vehicle, it inspects the vehicle based on the image data acquired by the image data acquisition unit 303 and the reference image data stored in the reference state storage unit 320 (reference state DB 32). This makes it possible in the first embodiment to perform vehicle inspection work that takes the presence of passengers into consideration, thereby suppressing a decrease in vehicle inspection accuracy compared to conventional methods.

[0079] In the above description, an example has been described in which the inspection unit 304 inspects a vehicle using image data captured by IP cameras 1 to 4 installed inside the passenger compartment of the vehicle. However, the installation locations of the IP cameras 1 to 4 in the first embodiment are not limited to inside the passenger compartment. For example, the IP cameras 1 to 4 may be installed outside the passenger compartment to capture images of the pantograph and underfloor of the vehicle. In this case, the inspection unit 304 can also inspect the pantograph and devices installed under the floor of the vehicle by using the image data captured by the IP cameras 1 to 4.

[0080] Also, in the above explanation, an example has been described in which the IP cameras 1 to 4 are directly connected to the recording device 5. However, the IP cameras 1 to 4 are not limited to this, and may be configured to be connected to a separately prepared PoE HUB or the like, and connected to the recording device 5 via the PoE HUB or the like over a network.

[0081] In the above explanation, four IP cameras are installed in each car. However, the number of IP cameras installed in each car is not limited to this, and it is sufficient that one or more IP cameras are installed in each car.

[0082] In the above description, an example has been described in which the IP cameras 1 to 4 periodically capture images of the interior of the guest room. However, the IP cameras 1 to 4 are not limited to this, and may constantly capture images of the interior of the guest room in real time and acquire live image data (so-called video data). In this case, the IP cameras 1 to 4 may output the acquired live image data to the recording device 5. Furthermore, when acquiring image data from the recording device 5, the image data acquisition unit 303 may appropriately extract image data from the live image data and acquire the extracted image data.

[0083] In the above description, an example has been described in which the image data acquisition unit 303 transmits to the recording device 5 a request to acquire image data captured by the IP cameras 1 to 4 of each car, and acquires the image data distributed from the recording device 5 in response to this acquisition request. However, the image data acquisition unit 303 is not limited to this, and may be configured to acquire image data directly from the IP cameras 1 to 4.

[0084] In the above description, an example has been described in which the analysis device 30 including the vehicle state information acquisition unit 301, the passenger presence / absence estimation unit 302, the image data acquisition unit 303, the inspection unit 304, and the inspection result notification unit 305 is installed in a vehicle. However, the analysis device 30 is not limited to this, and may be installed in another location that can communicate with the devices and equipment installed in the vehicle via the wireless devices 40 and 41. For example, the analysis device 30 may be installed at a ground base where the terminal 50 is installed, or on the cloud.

[0085] In the above description, an example has been described in which the vehicle state information acquisition unit 301, passenger presence / absence estimation unit 302, image data acquisition unit 303, inspection unit 304, and inspection result notification unit 305 are included in the analysis device 30. However, the above units are not limited to this, and may be incorporated into devices other than the analysis device 30, such as the IP cameras 1 to 4 and the recording device 5.

[0086] Furthermore, in the above description, the vehicle is a railroad car, but the vehicle is not limited to this and may be, for example, a bus or taxi that carries passengers.

[0087] As described above, according to the first embodiment, the vehicle maintenance system includes a vehicle state information acquisition unit 301 that acquires vehicle state information indicating the state of the vehicle, a passenger presence / absence estimation unit 302 that estimates the presence or absence of a passenger in the vehicle by comparing the vehicle state information acquired by the vehicle state information acquisition unit 301 with preset information indicating a timing when no passengers are present in the vehicle, an image data acquisition unit 303 that acquires image data indicating an image of the vehicle captured by an imaging device when the passenger presence / absence estimation unit 302 estimates that no passengers are present in the vehicle, and an inspection unit 304 that inspects the vehicle based on the image data acquired by the image data acquisition unit 303 and reference image data indicating an image of the vehicle captured by the imaging device when the vehicle is in a normal state. This enables the vehicle maintenance system to perform vehicle inspection work taking into account the presence of passengers, and can suppress a decrease in vehicle inspection accuracy compared to conventional systems.

[0088] Furthermore, the inspection unit 304 compares and analyzes the image data acquired by the image data acquisition unit 303 with the reference image data, and if a difference is detected, it obtains a result that there is an abnormality in the vehicle, whereas if no difference is detected, it obtains a result that there is no abnormality in the vehicle. This allows the vehicle maintenance system to easily inspect vehicles.

[0089] The vehicle maintenance system also includes an inspection result notification unit 305 that notifies, as inspection results, data indicating the results obtained by the inspection unit 304. This allows the vehicle maintenance system to appropriately notify the inspection results.

[0090] Furthermore, when the inspection unit 304 determines that there is an abnormality in the vehicle, it identifies the location of the abnormality based on the comparison result between the image data and the reference image data, and the inspection result notification unit 305 notifies the inspection result including the location of the abnormality identified by the inspection unit 304. This allows the vehicle maintenance system to appropriately notify the location of the abnormality.

[0091] Embodiment 2 In the first embodiment, an example has been described in which the inspection unit 304 inspects a vehicle using image data optically captured inside the passenger compartment by the IP cameras 1 to 4. In the second embodiment, an example will be described in which a sensor is added in addition to the IP cameras 1 to 4, and the inspection unit 304 inspects a vehicle using sensor data acquired by the sensor in addition to the image data.

[0092] 11 is a diagram showing an example of the configuration of a vehicle maintenance system according to embodiment 2. In the vehicle maintenance system according to embodiment 2, sensors 6 to 9 are added to the system according to embodiment 1, and analysis device 30 is changed to analysis device 30b. Other configurations of the vehicle maintenance system according to embodiment 2 are the same as those of embodiment 1, so the same reference numerals are used and their description will be omitted.

[0093] The sensors 6-9 are composed of, for example, infrared sensors or distance sensors, and are attached to the IP cameras 1-4 installed in the passenger compartments of each car. The sensors 6-9 periodically sense the passenger compartments of each car and obtain data indicating the sensing results (hereinafter also referred to as "sensor data"). The sensors 6-9 also output the obtained sensor data to the recording device 5 as needed.

[0094] In addition to the reference image data and related information described in embodiment 1, the reference state DB32 stores data (hereinafter also referred to as "reference sensor data") showing the results of sensing the interior of each car's passenger compartment by sensors 6 to 9 when the vehicle is in the reference state.

[0095] Furthermore, the reference state DB 32 stores information related to the reference sensor data in association with the reference sensor data. The information related to the reference sensor data includes, for example, the train formation number of the vehicle from which the reference sensor data was acquired, the car number from which the reference sensor data was acquired, the number of the sensor that acquired the reference sensor data, and the date and time when the reference sensor data was saved in the reference state DB 32.

[0096] An example of the reference sensor data and related information stored in the reference state DB 32 is shown in Figure 12. As shown in Figure 12, the reference sensor data is saved under a file name such as "A001-1-1s.iri." Furthermore, according to the related information associated with this reference sensor data, it can be seen that this reference sensor data was acquired by sensor 6 installed in car No. 1 of a vehicle with the train formation number "A001," and that the data was saved in the reference state DB 32 at "2021 / 12 / 10 23:10."

[0097] Similarly, the reference sensor data is saved with a file name such as "A001-1-2s.iri." Furthermore, according to the related information associated with this reference sensor data, it is known that this reference sensor data was acquired by sensor 7 installed in car No. 1 of the vehicle with the train formation number "A001," and that the data was saved in the reference state DB 32 at "2021 / 12 / 3 23:00."

[0098] The reference sensor data and related information are created in advance by the administrator of the vehicle maintenance system based on the results of sensing by sensors 6 to 9 when the vehicle is in a reference state, and are stored in the reference state DB 32 using, for example, the analysis device 30b.

[0099] 12 does not show the reference image data and related information stored in the reference condition DB 32. However, the reference condition DB 32 can simultaneously store the reference image data and related information and the reference sensor data and related information.

[0100] The analysis device 30b inspects the vehicle based on image data captured by the IP cameras 1 to 4, sensor data from the sensors 6 to 9, reference image data and related information stored in the reference condition DB 32, and reference sensor data and related information stored in the reference condition DB 32.

[0101] By using the sensor data from sensors 6-9, analysis device 30b can identify abnormalities that would be difficult to analyze using only the image data captured by IP cameras 1-4. For example, if sensors 6-9 are configured with infrared sensors, sensors 6-9 can detect a wet seat. Therefore, in this case, analysis device 30b can identify a wet seat as an abnormality.

[0102] The sensors 6 to 9 may be attached to the IP cameras 1 to 4 or may be built into the IP cameras 1 to 4. The sensors 6 to 9 may also be installed separately from the IP cameras 1 to 4 in the vehicle and connected to a network within the vehicle.

[0103] Next, a functional block diagram of a vehicle maintenance system according to the second embodiment will be described with reference to FIG.

[0104] 11 is different from the functional block diagram of the vehicle maintenance system according to the first embodiment shown in Fig. 5 in that a sensor data acquisition unit 306 is added and the inspection unit 304 is changed to an inspection unit 304b. The other functional blocks shown in Fig. 11 are the same as the functional blocks shown in Fig. 5, and therefore the same reference numerals are used and their description will be omitted.

[0105] When the passenger presence estimation unit 302 estimates that there are no passengers in the vehicle at the time the vehicle state information is acquired, the sensor data acquisition unit 306 transmits to the recording device 5 a request to acquire the sensor data acquired by the sensors 6 to 9 of each car that is recorded in the recording device 5. Then, the sensor data acquisition unit 306 acquires the sensor data acquired by the sensors 6 to 9 of each car that is distributed from the recording device 5 as a response to this acquisition request. At this time, the sensor data acquisition unit 306 may acquire multiple pieces of sensor data from the recording device 5.

[0106] Similar to the inspection unit 304 described in embodiment 1, the inspection unit 304b inspects the vehicle based on the image data acquired by the image data acquisition unit 303 and the reference image data stored in the reference state storage unit 320 (reference state DB32).

[0107] Furthermore, the inspection unit 304b inspects the vehicle based on the sensor data acquired by the sensor data acquisition unit 306 and the reference sensor data stored in the reference state storage unit 320 (reference state DB 32). Specifically, the inspection unit 304b compares and analyzes the sensor data and the reference sensor data using a background subtraction method to obtain an analysis result indicating whether or not there is an abnormality in the vehicle. Furthermore, if the inspection unit 304b obtains an analysis result indicating there is an abnormality in the vehicle, it identifies a difference between the sensor data and the reference sensor data as an abnormality location.

[0108] Here, an example of an inspection method by the inspection unit 304b will be described. Here, as an example, the inspection unit 304b compares and analyzes image data captured by the IP cameras 1 to 4 on the first car with reference image data stored in the reference state DB 32 using a background subtraction method, and also compares and analyzes sensor data acquired by the sensors 6 to 9 on the first car with reference sensor data stored in the reference state DB 32 using a background subtraction method.

[0109] It is assumed that the reference image data and related information stored in the reference state DB 32 are in the state shown in FIG. 4, and the reference sensor data and related information stored in the reference state DB 32 are in the state shown in FIG.

[0110] In this case, for example, the inspection unit 304b compares and analyzes, by background subtraction, the image data captured by the IP camera 1 in car No. 1 with the reference image data (file name "A001-1-1.jpg") captured by the IP camera 1 in car No. 1, which is stored in the reference condition DB 32. Then, the inspection unit 304b obtains an analysis result indicating whether or not there is an abnormality within the range captured by the IP camera 1 in the passenger compartment of car No. 1.

[0111] Furthermore, the inspection unit 304b compares and analyzes, by background subtraction, the sensor data acquired by the sensor 6 of car No. 1 with the reference sensor data (file name "A001-1-1s.iri") acquired by the sensor 6 of car No. 1 and stored in the reference state DB 32. As a result, the inspection unit 304b obtains an analysis result indicating whether or not there is an abnormality within the range sensed by the sensor 6 in the passenger compartment of car No. 1.

[0112] For example, in Fig. 14, reference numeral 1401 denotes a reference image taken by IP camera 1 on car No. 1. Reference numeral 1402 denotes an image taken by IP camera 1 on car No. 1. Image 1402 captures anomalies E1 in two locations. Reference numeral 1403 denotes the results of comparing and analyzing reference image 1401 and image 1402 by background subtraction.

[0113] In FIG. 14, reference numeral 1404 denotes reference sensor data acquired by sensor 6 on car No. 1. Reference numeral 1405 denotes sensor data acquired by sensor 6 on car No. 1. The sensor data 1405 includes an image of an abnormality E2 indicating a wet seat. Reference numeral 1406 denotes the results of comparing and analyzing the reference sensor data 1404 and the sensor data 1405 by background subtraction. Reference numeral 1407 denotes the results of overlaying the results 1403 and 1406.

[0114] As shown in FIG. 14, the inspection unit 304b performs comparison and analysis using background subtraction on the image 1402 and the sensor data 1405. Here, in a case where an abnormality E2 due to a wet seat has occurred, for example, the inspection unit 304b would have difficulty detecting the occurrence of the abnormality E2 by performing comparison and analysis using background subtraction using only the image data. On the other hand, the sensor 6 (infrared sensor) can detect the abnormality E2 due to the wet seat. Therefore, the inspection unit 304b can detect the abnormality E2 due to the wet seat and identify the location of the abnormality E2 by performing comparison and analysis using background subtraction using the sensor data.

[0115] As with the inspection unit 304 in the first embodiment, the inspection unit 304b may use not only background subtraction but also other methods such as machine learning as a comparison and analysis method.

[0116] Next, an example of the operation of the vehicle maintenance system according to the second embodiment will be described with reference to the flowchart shown in Fig. 15. Note that steps ST11 to ST12 shown in Fig. 15 are the same as steps ST1 to ST2 described as an example of the operation of the vehicle maintenance system according to the first embodiment shown in Fig. 9, and therefore will not be described again.

[0117] In step ST13, the image data acquisition unit 303 transmits to the recording device 5 a request to acquire image data captured by the IP cameras 1 to 4 of each car, which is recorded in the recording device 5. Then, the image data acquisition unit 303 acquires the image data captured by the IP cameras 1 to 4 of each car, which is distributed from the recording device 5 in response to this acquisition request.

[0118] Furthermore, in step ST13, the sensor data acquisition unit 306 transmits to the recording device 5 a request to acquire the sensor data acquired by the sensors 6 to 9 of each car, which is recorded in the recording device 5. Then, the sensor data acquisition unit 306 acquires the sensor data acquired by the sensors 6 to 9 of each car, which is distributed from the recording device 5 as a response to this acquisition request (step ST13).

[0119] Next, the inspection unit 304b inspects the vehicle, including the passenger compartments of each car, by comparing and analyzing the image data acquired by the image data acquisition unit 303 in step ST13 with the reference image data stored in the reference state DB32, and by comparing and analyzing the sensor data acquired by the sensor data acquisition unit 306 in step ST13 with the reference sensor data stored in the reference state DB32 (step ST14).

[0120] 15 are the same as steps ST5 to ST7 described as an example of the operation of the vehicle maintenance system according to the first embodiment shown in Fig. 9, and therefore will not be described again. In step ST16, when transmitting the inspection result to the monitor device 20 or the like, the inspection result notification unit 305 may also transmit information indicating which of the image data or the sensor data the inspection unit 304b used to detect the abnormality.

[0121] As described above, in the second embodiment, the inspection unit 304b inspects the vehicle using the sensor data acquired by the sensors 6 to 9 in addition to the image data captured by the IP cameras 1 to 4. As a result, in the vehicle maintenance system according to the second embodiment, when there is an abnormality that is difficult to detect by comparison and analysis using background subtraction of image data but can be detected by comparison and analysis using background subtraction of sensor data, the abnormality can be detected, and a wider variety of abnormalities can be detected than in the first embodiment.

[0122] In the above description, an example has been described in which the sensors 6 to 9 are attached to the IP cameras 1 to 4. However, the installation locations of the sensors 6 to 9 in the second embodiment are not limited to this. For example, the sensors 6 to 9 may be installed outside the passenger compartment together with the IP cameras 1 to 4 in order to sense the pantograph and underfloor of the vehicle. In this case, the inspection unit 304b can also inspect the pantograph and devices installed under the floor of the vehicle by using image data captured by the IP cameras 1 to 4 and sensor data acquired by the sensors 6 to 9.

[0123] In the above description, an example has been described in which infrared sensors are used as the sensors 6 to 9. However, the sensors 6 to 9 are not limited to this, and various sensors such as a distance sensor, a microphone (audio sensor), a CO2 sensor, and an acceleration sensor may be used. In this case, the output log data and the like may be regarded as sensor data, and the device mounted on the vehicle itself may be used as a sensor.

[0124] In the above description, the sensor data acquiring unit 306 transmits a request to the recording device 5 to acquire the sensor data acquired by the sensors 6 to 9 of each car, and acquires the sensor data distributed from the recording device 5 in response to this acquisition request. However, the sensor data acquiring unit 306 is not limited to this, and may be configured to acquire the sensor data directly from the sensors 6 to 9.

[0125] In the above description, an example has been described in which the analysis device 30b includes the vehicle state information acquisition unit 301, the passenger presence / absence estimation unit 302, the image data acquisition unit 303, the sensor data acquisition unit 306, the inspection unit 304b, and the inspection result notification unit 305. However, the above units are not limited to this, and the IP cameras 1 to 4, the sensors 6 to 9, the recording device 5, and the like may be incorporated into devices other than the analysis device 30b.

[0126] As described above, according to the second embodiment, the vehicle maintenance system includes a sensor data acquisition unit 306 that acquires sensor data indicating the results of sensing the vehicle by sensors 6 to 9 when the passenger presence estimation unit 302 estimates that there are no passengers in the vehicle, and the inspection unit 304b inspects the vehicle based on the image data, reference image data, the sensor data acquired by the sensor data acquisition unit 306, and the reference sensor data indicating the results of sensing the vehicle by sensors 6 to 9 when there are no abnormalities in the vehicle beforehand. As a result, the vehicle maintenance system according to the second embodiment can detect a wider variety of abnormalities than the first embodiment, in addition to the effects of the first embodiment.

[0127] Embodiment 3 In the first and second embodiments, examples of inspecting a vehicle using data obtained from a vehicle consisting of a single formation have been described. In the third embodiment, an example of inspecting a vehicle using data obtained from a vehicle consisting of multiple formations will be described.

[0128] Fig. 16 is a diagram showing an example of the configuration of a rolling stock maintenance system according to embodiment 3. As shown in Fig. 16, the rolling stock maintenance system according to embodiment 3 includes rolling stock consisting of a plurality of formations (formation A001, formation A002, ..., formation A00n, formation B00n). The rolling stock of each formation is connected to an analysis base 60 and a ground base so as to be able to communicate with each other via a wireless device 40 or the like.

[0129] In the following explanation, each of the multiple train formations will be represented as, for example, "A001," "A002," and "B00n." Furthermore, if the train formation is "A001," "A" indicates the train formation type and "001" indicates the train formation number. Therefore, train formations "A001" and "A002" have the same train formation type, and train formations "A001" and "B00n" have different train formation types.

[0130] In the following explanation, it is assumed that the interiors of the passenger compartments of vehicles with the same train type are the same. For example, since the train types "A001" and "A002" are the same, "A," the interiors of the passenger compartments of the vehicles are the same. Furthermore, since the train types "A001" and "B00n" are different, the interiors of the passenger compartments of the vehicles are also different.

[0131] In the following explanation, it is assumed that if the train set type or train set number is different, the manufacturing date of the vehicle will also be different. For example, train set "A001" and train set "A002" have different train set numbers, so the manufacturing date of the vehicle will also be different. Furthermore, train set "A001" and train set "B00n" have different train set types, so the manufacturing date of the vehicle will also be different.

[0132] Similar to the vehicles described in embodiment 1, formation A001 is equipped with IP cameras 1-4, a recording device 5, a vehicle state information management device 10, a monitoring device 20, an analysis device 30c, a vehicle state DB 31, and a radio device 40.

[0133] The train sets A002 to A00n have the same train type as train set A001, and the configuration of the devices and the like installed in the cars is also the same as that of train set A001. That is, like train set A001, train sets A002 to A00n are equipped with IP cameras 1 to 4, a recording device 5, a vehicle state information management device 10, a monitoring device 20, an analysis device 30c, a vehicle state DB 31, and a radio device 40. Note that, in FIG. 16, for the sake of simplicity, only the radio device 40 is shown for train sets A002 to A00n.

[0134] As described above, the compositions A001 to A00n have the same configuration of devices installed in the cars, and also the same interior decorations, etc. However, the compositions A001 to A00n have different manufacturing dates for the cars.

[0135] Although the train set B00n is a different train set type from train set A001, the configuration of the installed devices is the same as that of train set A001. That is, like train set A001, train set B00n is also equipped with IP cameras 1-4, a recording device 5, a vehicle state information management device 10, a monitoring device 20, an analysis device 30c, a vehicle state DB 31, and a radio device 40. Note that, in FIG. 16, for the sake of simplicity, only the radio device 40 is shown for train set B00n.

[0136] Of the components installed in each of the above-mentioned formations, the components other than the analysis device 30c are the same as those explained in the first embodiment, and therefore will not be explained again.

[0137] Similar to the analysis device 30 in the first embodiment, the analysis device 30c acquires vehicle state information periodically distributed from the vehicle state information management device 10. Then, the analysis device 30c compares the acquired vehicle state information with the unoccupied state information stored in the vehicle state DB 31, and estimates, based on the result of the comparison, whether or not the timing at which the vehicle state information was acquired was a timing at which no passengers were in the vehicle.

[0138] If the analysis device 30c estimates that the timing at which the vehicle status information was acquired means that no passengers were in the vehicle, the analysis device 30c acquires image data captured by the IP cameras 1 to 4 in each vehicle from the recording device 5. The analysis device 30c then transmits the acquired image data to the analysis site 60 via the wireless device 40 and a network device 66, which will be described later.

[0139] The analysis base 60 is configured, for example, by a cloud, a data center, etc. At the analysis base 60, a reference state group DB 63, a schedule information DB 64, an analysis server 65, and network devices 66 are installed.

[0140] The network device 66 performs wireless communication between the radio device 40 installed in each train formation and the radio device 41 installed at a ground base.

[0141] The reference condition group DB 63 stores reference image data and related information in a plurality of components.

[0142] An example of reference image data and related information stored in the reference state group DB 63 is shown in Figure 17. As shown in Figure 17, the reference image data is saved with a file name such as "A001-1-1_210610-2234.jpg." Furthermore, according to the related information associated with this reference image, it can be seen that this reference image was taken by IP camera 1 installed in car No. 1 of a vehicle with the formation number "A001," and that the data was saved in the reference state group DB 63 at "2021 / 6 / 10 22:34."

[0143] Similarly, the reference image data is saved with a file name such as "A001-1-1_210912-0610.jpg." Furthermore, according to the related information associated with this reference image data, it can be seen that this reference image was taken by IP camera 1 installed in car No. 1 of a vehicle with the formation number "A001," and that the data was saved in reference state group DB63 at "2021 / 9 / 12 06:10."

[0144] Similarly, the reference image data is saved with a file name such as "A002-1-1_210913-0551.jpg." Furthermore, according to the related information associated with this reference image data, it can be seen that this reference image was taken by IP camera 1 installed in car No. 1 of a vehicle with the formation number "A002," and that the data was saved in reference state group DB63 at "2021 / 9 / 13 05:51."

[0145] Similarly, the reference image data and related information for the rolling stock whose formation is "A003" and the rolling stock whose formation is "B001" are also stored in the reference state group DB 63 in the same manner as above.

[0146] The formation information DB 64 stores formation information for each formation.

[0147] An example of the train formation information stored in the train formation information DB 64 is shown in Fig. 18. As shown in Fig. 18, the train formation information includes, for example, an IP address set in the radio device 40 installed in each train formation, the train formation number, the train formation type, the manufacturing date of the vehicle, and the date and time of the most recent maintenance of the vehicle.

[0148] For example, the first piece of formation information shown in Figure 18 indicates that the manufacturing date of a vehicle with formation number "A001" and formation type "A" is "2020 / 3 / 10" and the most recent maintenance date and time is "2021 / 12 / 10 23:10". The first piece of formation information also indicates that the IP address set in the wireless device 40 installed in the vehicle is "192.168.1.1".

[0149] Similarly, the second piece of train formation information indicates that the train formation number is "A002", the train formation type is "A", the manufacturing date is "2020 / 6 / 11", and the most recent maintenance date and time is "2021 / 12 / 11 23:12". The second piece of train formation information also indicates that the IP address set in the wireless device 40 installed in the train formation is "192.168.1.2".

[0150] Similarly, for the train formation number "A003" and the train formation number "B001", train formation information including the train formation type, manufacturing date, most recent maintenance date and time, and the IP address of the wireless device 40 is stored.

[0151] The analysis server 65 receives the image data transmitted from each analysis device 30c via the wireless device 40 and the network device 66. Then, the analysis server 65 inspects the vehicles for each formation based on the received image data, the reference image data and related information stored in the reference condition group DB 63, and the formation information stored in the formation information DB 64.

[0152] The analysis server 65 also transmits the inspection results to the monitor device 20 installed in the formation corresponding to the inspection results via the network device 66 and the radio device 40. The analysis server 65 also transmits the inspection results to the terminal 50 via the network device 66 and the radio device 41. The analysis method used by the analysis server 65 will be described in detail later.

[0153] Next, a functional block diagram of a vehicle maintenance system according to the third embodiment will be described with reference to FIG.

[0154] The functional block diagram shown in Figure 19 adds a communication unit 311, a formation identification unit 312, a reference image data selection unit 313, and a formation information storage unit 340 to the functional block diagram of the vehicle maintenance system related to embodiment 1 shown in Figure 5.

[0155] 19 is different from the functional block diagram of the vehicle maintenance system according to the first embodiment shown in Fig. 5 in that the reference state storage unit 320 is changed to a reference state group storage unit 330, the inspection unit 304 is changed to an inspection unit 304c, and the inspection result notification unit 305 is changed to an inspection result notification unit 305c. The other functional blocks shown in Fig. 19 are the same as the functional blocks shown in Fig. 5, and therefore the same reference numerals are used and their description will be omitted.

[0156] Here, an example will be described in which the vehicle state information acquisition unit 301, the passenger presence / absence estimation unit 302, the image data acquisition unit 303, and the communication unit 311 are included in the analysis device 30c installed in each train set. Note that, although Fig. 19 will be described taking the analysis device 30c installed in train set A001 as an example, the analysis devices 30c installed in the other train sets are similar to the analysis device 30c installed in train set A001.

[0157] Here, an example will be described in which the formation identification unit 312, the reference image data selection unit 313, the inspection unit 304c, and the inspection result notification unit 305c are included in the analysis server 65. Here, the reference state group storage unit 330 is configured, for example, by the reference state group DB 63. The formation information storage unit 340 is configured, for example, by the formation information DB 64.

[0158] The communication unit 311 transmits the image data acquired by the image data acquisition unit 303 to the formation identification unit 312 via the radio device 40 and the network device 66.

[0159] The formation identification unit 312 receives the image data transmitted from the communication unit 311. Then, based on the transmission information accompanying the received image data, the formation identification unit 312 identifies the IP address of the wireless device 40 through which the image data was transmitted.

[0160] Next, the formation identification unit 312 refers to the formation information storage unit 340 (formation information DB 64) based on the identified IP address, and identifies the formation number of the vehicle in which the wireless device 40 to which the IP address is set is installed. In other words, the formation identification unit 312 identifies the formation number of the vehicle that is the transmission source of the image data.

[0161] Next, the reference image data selection unit 313 identifies the formation type from the formation number identified by the formation identification unit 312. Then, the reference image data selection unit 313 refers to the reference condition group storage unit 330 (reference condition group DB 63) and selects two or more reference image data that meet selection conditions from the reference image data stored in the reference condition group DB 63. The selection conditions will be described later.

[0162] Next, the inspection unit 304c inspects the vehicle based on each of the two or more reference image data selected by the reference image data selection unit 313 and the image data transmitted from the communication unit 311. Specifically, the inspection unit 304c compares and analyzes the image data and each of the reference image data using a background subtraction method to obtain an analysis result indicating whether or not there is an abnormality in the vehicle. Furthermore, if the inspection unit 304c obtains an analysis result indicating there is an abnormality in the vehicle, it identifies differences between the image data and each of the reference image data as abnormal locations.

[0163] The inspection result notifying unit 305c notifies the inspection result by transmitting the analysis result obtained by the inspection unit 304c as the inspection result of the vehicle to the monitor device 20 installed in the formation corresponding to the inspection result via the network device 66 and the wireless device 40. Furthermore, the inspection result notifying unit 305c transmits the obtained analysis result as the inspection result of the vehicle to the terminal 50 via the network device 66 and the wireless device 41.

[0164] Here, an example of an inspection method by the inspection unit 304c included in the analysis server 65 will be described. Here, as an example, it is assumed that the reference image data and related information stored in the reference state group DB 63 are in the state shown in Fig. 17. It is also assumed that the organization information stored in the organization information DB 64 is in the state shown in Fig. 18.

[0165] First, the formation identification unit 312 refers to the image data received from the communication unit 311 via the wireless device 40 and the network device 66, and identifies the IP address of the wireless device 40 through which the image data was transmitted. Here, it is assumed that the formation identification unit 312 identifies "192.168.1.2" as the IP address, for example.

[0166] Next, the formation identification unit 312 refers to the formation information DB 64 based on the identified IP address "192.168.1.2" and identifies the formation number of the vehicle in which the wireless device 40 to which the IP address is set is installed. In this example, as shown in FIG. 18, the formation number corresponding to "192.168.1.2" is "A002." Therefore, the formation identification unit 312 identifies that the formation number of the vehicle that transmitted the image data is "A002."

[0167] Next, the reference image data selection unit 313 identifies the formation type "A" from the formation number "A002" identified by the formation identification unit 312. Then, the reference image data selection unit 313 refers to the reference condition group DB63 and selects two or more reference image data that meet the selection conditions from the reference image data stored in the reference condition group DB63.

[0168] Here, the selection conditions include, for example, (1) the reference image data is taken of a train set having the same train set type as the train set type "A" identified by the train set identification unit 312, and (2) the reference image data is the same in terms of the car number and camera number used to capture the image indicated by the image data received by the train set identification unit 312 (acquired by the image data acquisition unit 303). The reference image data selection unit 313 selects reference image data that meets the above selection conditions for each train set number. Note that "the camera number is the same" means that the position of the IP camera in the car is the same.

[0169] For example, suppose the train set number at which the image indicated by the image data was taken is "A002," the car where the image was taken is "car 1," and the camera number that took the image is "camera 1." In this case, the reference image data selection unit 313 selects "A002-1-1_210913-0551.jpg" and "A002-1-1_211211-2312.jpg," which have a train set number of "A002," a train set type of "A," a car number of "car 1," and a camera number of "camera 1," as reference image data that meets the selection conditions. In other words, the reference image data selection unit 313 selects two reference image data corresponding to the train set number "A002."

[0170] Furthermore, the reference image data selection unit 313 selects, as reference image data that meets the selection conditions, "A001-1-1_210610-2234.jpg," "A001-1-1_210912-0610.jpg," and "A001-1-1_211210-2310.jpg," which have a train formation number of "A001," a train formation type of "A," a car number of "car 1," and a camera number of "camera 1." In other words, the reference image data selection unit 313 selects three pieces of reference image data that correspond to the train formation number "A001."

[0171] Similarly, the reference image data selection unit 313 selects "A003-1-1_211209-2234.jpg," which has a train formation number of "A003," a train formation type of "A," a car number of "car 1," and a camera number of "camera 1," as reference image data that meets the selection conditions. In other words, the reference image data selection unit 313 selects one reference image data corresponding to the train formation number "A003."

[0172] 17, the reference state group DB63 also stores reference image data whose train formation number is "B001" and whose train formation type is "B." However, since this reference image data does not meet the above selection conditions, the reference image data selection unit 313 does not select this reference image data.

[0173] Next, the inspection unit 304c compares and analyzes each of the two or more reference image data selected by the reference image data selection unit 313 with the image data transmitted from the communication unit 311 of the train formation number "A002" using background difference, and obtains an analysis result indicating whether or not there is an abnormality in the vehicle of the train formation number "A002".

[0174] 20 shows an example of comparison and analysis by the inspection unit 304c using background subtraction. Here, as an example, an example will be described in which the inspection unit 304c compares and analyzes the image data with each of the reference image data corresponding to the train set number "A001", the reference image data corresponding to the train set number "A002", and the reference image data corresponding to the train set number "A003", which have been selected by the reference image data selection unit 313, using background subtraction.

[0175] For example, in Figure 20, the top row shows an example of comparison and analysis between reference image 2001 corresponding to train formation number "A001" and image 2002, the middle row shows an example of comparison and analysis between reference image 2004 corresponding to train formation number "A002" and image 2002, and the bottom row shows an example of comparison and analysis between reference image 2006 corresponding to train formation number "A003" and image 2002.

[0176] Here, in the image 2002, for example, only three straps T are shown, although there should actually be four straps T. In other words, the image 2002 shows an abnormality in that there are not enough straps T.

[0177] However, for some reason, an image containing the abnormality of a lack of strap T, similar to the above, is mistakenly stored in the reference condition group DB 63 as the reference image 2001. In this case, by comparing the reference image 2001 with the image 2002, the abnormality of a lack of strap T cannot be detected, as indicated by the reference symbol 2003.

[0178] On the other hand, the reference image 2004 and the reference image 2006 show four straps T correctly attached. Therefore, by comparing the reference image 2004 with the image 2002 and by comparing the reference image 2006 with the image 2002, it is possible to detect an abnormality, namely, a lack of straps T, as shown by the reference numerals 2005 and 2007.

[0179] In this case, the inspection unit 304c may obtain the final analysis result 2008 by, for example, majority vote. In the above example, since there was one result of "no abnormality" and two results of "abnormality detected," "abnormality detected" is set as the final analysis result 2008. Alternatively, the inspection unit 304c may set the final analysis result 2008 to "abnormality detected" if there is at least one result of "abnormality detected," for example.

[0180] In this way, the inspection unit 304c compares and analyzes each of the two or more reference image data selected by the reference image data selection unit 313 with the image data transmitted from the communication unit 311 using background subtraction, and obtains a final analysis result based on each comparison result.

[0181] That is, the inspection unit 304c inspects the vehicle not only using the reference image data of the train set to be identified as an anomaly, but also using the reference image data of multiple train sets of the same train set type for comparison and analysis. This enables the inspection unit 304c to detect anomalies that are difficult to identify by comparison and analysis with reference image data obtained from a single train set. For example, even if one of two or more reference image data contains an anomaly, normal analysis is possible by using the results of comparison with the other reference image data.

[0182] Furthermore, even in cases where it is difficult to distinguish between age-related deterioration and an abnormality, such as when only a train with a specific train number has rapidly deteriorating seats and floors, the inspection unit 304c can accurately detect abnormalities by comparing and analyzing the reference image data of trains with other train numbers, further improving the accuracy of the analysis.

[0183] In addition, the above-mentioned selection conditions may include selecting reference image data in which the number of days elapsed from the date of manufacture of the vehicle on which the reference image indicated by the reference image data was taken to the date on which the reference image data was saved in the reference condition group DB63 is closest to the number of days elapsed from the date of manufacture of the vehicle on which the image indicated by the image data was taken to the date on which the image data acquisition unit 303 acquired the image data.

[0184] For example, similarly to the above, the formation identification unit 312 identifies that the formation number of the vehicle that transmitted the image data is "A002." Also, assume that the image data acquisition unit 303 acquires the image data on "2021 / 12 / 20," and the formation identification unit 312 receives the image data on the same day, "2021 / 12 / 20."

[0185] In this case, the reference image data selection unit 313 refers to the train formation information DB 64 and determines that the manufacturing date of the train formation with the train formation number "A002" is "2020 / 6 / 11." Then, it calculates "1 year and 6 months" as the number of days that have elapsed from this manufacturing date "2020 / 6 / 11" to the image data acquisition date "2021 / 12 / 20."

[0186] Then, the reference image data selection unit 313 selects the reference image data for which the number of days elapsed from the manufacturing date of the composition to the storage date of the reference image data is closest to the above-mentioned "1 year and 6 months."

[0187] For example, the manufacturing date of train set number "A002" is "2020 / 6 / 11." Therefore, the reference image data selection unit 313 selects the reference image data whose storage date is closest to "2021 / 12 / 11," which is 1 year and 6 months after the manufacturing date "2020 / 6 / 11." In the above example, for train set number "A002," the reference image data selection unit 313 selects "A002-1-1_211211-2312.jpg," which has a storage date of "2021 / 12 / 11," as the reference image data.

[0188] Similarly, the reference image data selection unit 313 references the formation information DB 64 and determines that the manufacturing date of formation number "A001" is "2020 / 3 / 10". Therefore, for formation number "A001", the reference image data selection unit 313 selects the reference image data whose storage date is closest to "2021 / 9 / 10", 1 year and 6 months after the manufacturing date "2020 / 3 / 10". In the above example, for formation number "A001", the reference image data selection unit 313 selects "A001-1-1_210912-0610.jpg", which has a storage date of "2021 / 9 / 12", as the reference image data.

[0189] Similarly, the reference image data selection unit 313 references the formation information DB 64 and determines that the manufacturing date of formation number "A003" is "2021 / 9 / 9." However, since the date the image data was received was "2021 / 12 / 20," just three months after the manufacturing date, there is no corresponding reference image data for formation number "A003." Therefore, the reference image data selection unit 313 does not select reference image data for formation number "A003."

[0190] In this way, by adding a condition related to the manufacturing date of the train set to the selection conditions, the reference image data selection unit 313 can select reference image data that is closest to the state of the vehicle at the time the image data acquisition unit 303 acquired the image data.

[0191] Next, an example of the operation of the vehicle maintenance system according to the third embodiment will be described with reference to the flowchart shown in Fig. 21. Note that steps ST21 to ST23 shown in Fig. 21 are the same as steps ST1 to ST3 described as an example of the operation of the vehicle maintenance system according to the first embodiment shown in Fig. 9, and therefore will not be described again.

[0192] In step ST24, the communication unit 311 transmits the image data acquired by the image data acquisition unit 303 to the analysis server 65 via the wireless device 40 and the network device 66 (step ST24).

[0193] Next, the reference image data selection unit 313 of the analysis server 65 refers to the reference condition group DB 63 and selects two or more reference image data that meet the selection conditions from the reference image data stored in the reference condition group DB 63 (step ST25). The selection conditions and selection method are as described above, so a repeated explanation will be omitted.

[0194] Next, the inspection unit 304c compares and analyzes each of the two or more reference image data selected by the reference image data selection unit 313 with the image data transmitted from the communication unit 311 by using background subtraction, and inspects the vehicle, including the interior of each car (step ST26). Furthermore, if the inspection unit 304c obtains an analysis result indicating that there is an abnormality in the vehicle, it identifies differences between the image data transmitted from the communication unit 311 and each of the reference image data as abnormal locations.

[0195] Next, the inspection result notification unit 305c checks whether the analysis result obtained by the inspection unit 304c in step ST26 indicates that there is an abnormality in the vehicle (step ST27).

[0196] As a result, if the analysis result indicates that there is an abnormality in the vehicle (step ST27; Yes), the inspection result notifying unit 305c transmits analysis result data indicating the presence of the abnormality as the inspection result to the monitor device 20 installed in the vehicle that transmitted the image data via the network device 66 and the wireless device 40. The inspection result notifying unit 305c also transmits the analysis result data as the inspection result to the terminal 50 via the network device 66 and the wireless device 40. At this time, the inspection result notifying unit 305c transmits the abnormal location identified by the inspection unit 304c together with the inspection result (step ST28).

[0197] On the other hand, if the analysis result indicates that there is no abnormality in the vehicle (step ST27; No), the inspection result notification unit 305c transmits analysis result data indicating that there is no abnormality in the vehicle as the inspection result to the monitor device 20 installed in the vehicle that transmitted the image data via the network device 66 and the wireless device 40. Furthermore, the inspection result notification unit 305c transmits the analysis result data as the inspection result to the terminal 50 via the network device 66 and the wireless device 40 (step ST29).

[0198] The third embodiment may be combined with the second embodiment so that the inspection unit 304c performs comparison and analysis using sensor data in addition to image data. In this case, the communication unit 311 transmits the image data acquired by the image data acquisition unit 303 and the sensor data acquired by the sensor data acquisition unit 306 to the analysis server 65 via the wireless device 40 and the network device 66.

[0199] In addition, in the third embodiment, an example has been described in which an abnormality in a vehicle is detected as a final inspection result. However, the third embodiment is not limited to this, and can also be applied to detecting an abnormality in equipment such as a deviation in the angle of view of the IP cameras 1 to 4, lens damage, and focus deviation as a final inspection result.

[0200] Furthermore, in the third embodiment, an example has been described in which the reference image data selection unit 313 adds, as a selection condition, that the number of days elapsed from the manufacturing date of the train set to the date the reference image data is saved is closest to the number of days elapsed from the manufacturing date of the train set that is the source of the image data to the date the image data was acquired. However, the reference image data selection unit 313 is not limited to this, and may add, as a selection condition, that the number of days elapsed from the most recent maintenance date and time of the train set to the date the reference image data is saved is closest to the number of days elapsed from the most recent maintenance date and time of the train set that is the source of the image data to the date the image data was acquired.

[0201] In the third embodiment, the example has been described in which the formation identification unit 312, the reference image data selection unit 313, the inspection unit 304c, and the inspection result notification unit 305c are included in the analysis server 65. However, the present invention is not limited to this, and each of the above units may be incorporated into another server or the like installed at the analysis base 60, for example.

[0202] As described above, according to the third embodiment, the vehicle maintenance system is a railway vehicle, and includes a reference image data selection unit 313 that selects two or more reference image data that meet predetermined selection conditions from a reference condition group DB 63 in which a plurality of reference image data is stored for each vehicle formation type and each formation number, and an inspection unit 304c inspects the vehicle based on the two or more reference image data selected by the reference image data selection unit 313 and the image data acquired by the image data acquisition unit 303. As a result, in addition to the effects of the first embodiment, the vehicle maintenance system according to the third embodiment performs inspection using two or more reference image data, thereby improving inspection accuracy compared to the first embodiment.

[0203] Furthermore, the inspection unit 304c compares and analyzes each of the two or more reference image data selected by the reference image data selection unit 313 with the image data acquired by the image data acquisition unit 303, and obtains a final analysis result based on two or more results obtained by the comparison and analysis. As a result, even if one of the two or more reference image data contains an abnormality, the vehicle maintenance system can perform a normal analysis by using the result of comparison with the other reference image data.

[0204] Furthermore, the predetermined selection conditions include selecting, for each train formation number, reference image data that indicates a reference image taken of a vehicle of the same train formation type as the vehicle on which the image indicated by the image data was taken, and in which the car number of the vehicle on which the reference image was taken and the position of the imaging device that took the reference image are the same as the car number of the vehicle on which the image indicated by the image data was taken and the position of the imaging device that took the image. This allows the vehicle maintenance system to appropriately select two or more reference image data.

[0205] The predetermined selection conditions also include selecting reference image data in which the number of days elapsed from the manufacturing date of the vehicle when the reference image indicated by the reference image data was taken to the date when the reference image data was saved in the reference condition group DB 63 is closest to the number of days elapsed from the manufacturing date of the vehicle when the image indicated by the image data was taken to the date when the image data acquisition unit 303 acquired the image data. This allows the vehicle maintenance system to select reference image data in a state that is closest to the state of the vehicle at the time when the image data acquisition unit 303 acquired the image data.

[0206] In addition, the present disclosure allows for free combination of the respective embodiments, modification of any of the components of the respective embodiments, or omission of any of the components of the respective embodiments. [Industrial Applicability]

[0207] The vehicle maintenance system according to the present disclosure is capable of suppressing a decline in vehicle inspection accuracy compared to conventional systems, and is suitable for use in vehicle maintenance, etc. [Explanation of symbols]

[0208] 1 to 4 IP cameras (imaging devices), 5 recording devices, 6 to 9 sensors, 10 vehicle state information management device, 20 monitoring devices, 30, 30b, 30c analysis devices, 31 vehicle state DB, 32 reference state DB, 40 to 41 wireless devices, 50 terminal, 51 processing circuit, 52 CPU, 53 memory, 60 analysis base, 63 reference state group DB (storage unit), 64 formation information DB, 65 analysis server, 66 network equipment, 301 vehicle state information acquisition unit, 302 passenger presence / absence estimation unit, 303 image data acquisition unit, 304, 304b, 304c inspection unit, 305, 305c inspection result notification unit, 306 sensor data acquisition unit, 310 vehicle state storage unit, 311 communication unit, 312 formation identification unit, 313 reference image data selection unit, 320 reference state storage unit, 330 Reference state group memory unit, 340 train formation information memory unit, 601 reference image, 602 image, 603 image showing inspection results, 1401 reference image, 1402 image, 1403, 1406, 1407 images showing inspection results, 1404 reference sensor data, 1405 sensor data, 2001, 2004, 2006 reference image, 2002 image, 2003, 2005, 2007, 2008 images showing inspection results, E, E1, E2 abnormality, K lintel, T strap

Claims

1. a vehicle state information acquisition unit that acquires vehicle state information indicating a state of the vehicle, the vehicle state information being distributed while the vehicle is traveling and including information indicating a state in which the vehicle is traveling; a passenger presence / absence estimation unit that estimates the presence or absence of a passenger in the vehicle while it is traveling by comparing the vehicle state information acquired by the vehicle state information acquisition unit with preset information indicating a time when the vehicle is free of passengers while it is traveling; an image data acquisition unit that acquires image data representing an image of the vehicle captured by an imaging device when the passenger presence estimation unit estimates that there are no passengers in the vehicle; an inspection unit that inspects the vehicle based on the image data acquired by the image data acquisition unit and reference image data that indicates an image of the vehicle that was previously captured by the imaging device when there was no abnormality in the vehicle; A vehicle maintenance system equipped with

2. The inspection unit The image data acquired by the image data acquisition unit is compared and analyzed with the reference image data, and if a difference is extracted, a result is obtained that there is an abnormality in the vehicle, whereas if no difference is extracted, a result is obtained that there is no abnormality in the vehicle.

2. The vehicle maintenance system according to claim 1.

3. an inspection result notification unit that notifies data indicating the results obtained by the inspection unit as inspection results; 3. The vehicle maintenance system according to claim 1 or 2.

4. The inspection unit If a result indicating that an abnormality exists in the vehicle is obtained, a location of the abnormality is identified based on a comparison result between the image data and the reference image data; The inspection result notification unit The location of the abnormality identified by the inspection unit is included in the inspection result and notified.

4. The vehicle maintenance system according to claim 3.

5. a sensor data acquisition unit that acquires sensor data indicating a result of sensing the vehicle by a sensor when the passenger presence estimation unit estimates that there are no passengers in the vehicle; The inspection unit The vehicle is inspected based on the image data, the reference image data, the sensor data acquired by the sensor data acquisition unit, and reference sensor data indicating the results of sensing the vehicle by the sensor in a state where there is no abnormality in the vehicle.

5. The vehicle maintenance system according to claim 1, wherein the vehicle maintenance system comprises: a vehicle maintenance system for performing maintenance on a vehicle;

6. a vehicle state information acquisition unit that acquires vehicle state information indicating a state of the vehicle; a passenger presence / absence estimation unit that estimates the presence or absence of a passenger in the vehicle by comparing the vehicle state information acquired by the vehicle state information acquisition unit with preset information indicating a time when no passenger is present in the vehicle; an image data acquisition unit that acquires image data representing an image of the vehicle captured by an imaging device when the passenger presence estimation unit estimates that there are no passengers in the vehicle; an inspection unit that inspects the vehicle based on the image data acquired by the image data acquisition unit and reference image data that indicates an image of the vehicle that was previously captured by the imaging device when there was no abnormality in the vehicle; Equipped with the vehicle is a rail vehicle, a reference image data selection unit that selects two or more pieces of reference image data that meet predetermined selection conditions from a storage unit in which a plurality of pieces of reference image data are stored for each train formation type and each train formation number of the vehicle; The inspection unit A vehicle maintenance system that inspects the vehicle based on two or more reference image data selected by the reference image data selection unit and image data acquired by the image data acquisition unit.

7. The inspection unit The two or more reference image data selected by the reference image data selection unit are compared and analyzed with the image data acquired by the image data acquisition unit, and a final analysis result is obtained based on the two or more results obtained by the comparison and analysis.

7. The vehicle maintenance system according to claim 6.

8. The predetermined selection conditions are: The reference image data indicates a reference image taken in a vehicle of the same train formation type as the vehicle in which the image indicated by the image data was taken, and the car number of the vehicle in which the reference image was taken and the position of the imaging device that took the reference image are the same as the car number of the vehicle in which the image indicated by the image data was taken and the position of the imaging device that took the image.

8. The vehicle maintenance system according to claim 6 or 7.

9. The predetermined selection conditions are: selecting reference image data in which the number of days elapsed from the manufacturing date of the vehicle on which the reference image indicated by the reference image data was taken to the date on which the reference image data was stored in the storage unit is closest to the number of days elapsed from the manufacturing date of the vehicle on which the image indicated by the image data was taken to the date on which the image data acquisition unit acquired the image data.

9. The vehicle maintenance system according to claim 8.

10. A vehicle maintenance method by a vehicle maintenance system, comprising: a step in which a vehicle state information acquisition unit acquires vehicle state information indicating a state of the vehicle, the vehicle state information being distributed while the vehicle is traveling and including information indicating a state in which the vehicle is traveling; a passenger presence / absence estimation unit comparing the vehicle state information acquired by the vehicle state information acquisition unit with preset information indicating a time when the vehicle is traveling and no passengers are present, and estimating the presence / absence of a passenger in the traveling vehicle; an image data acquisition unit acquiring image data representing an image of the vehicle captured by an imaging device when the passenger presence / absence estimation unit estimates that there are no passengers in the vehicle; an inspection unit inspecting the vehicle based on the image data acquired by the image data acquisition unit and reference image data indicating an image of the vehicle previously captured by the imaging device when there was no abnormality in the vehicle; A vehicle maintenance method comprising:

Images