Seat Wetness Detection System

Abstract

To provide a seat wetting detection system that can reduce the work time and the physical burden required to detect wetting of a seat.SOLUTION: The present disclosure relates to a seat wetting detection system for detecting wetting of a seat. The seat wetting detection system includes: an image acquisition unit for acquiring a thermographic image that shows the temperature distribution of a space including seats and a visible light image of the space; an extraction unit for extracting a seat region from the visible light image; and a determination unit for determining wetting of a seat on the basis of the temperature distribution of the seat region extracted by the extraction unit.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a wet seat detection system. [Background technology]

[0002] Seats equipped with cushions are arranged in vehicles (e.g., railroad cars, buses, airplanes, etc.) and facilities (e.g., movie theaters, concert halls, etc.) (see Patent Document 1). Since an unspecified number of users take turns sitting on these seats, the seats may get wet with liquids such as beverages.

[0003] Therefore, seat cleaners must check whether the seats are wet between passenger changes and perform cleaning work if they find any. One method for detecting wetness is to detect moisture by stroking the seat with a broom equipped with electrodes, for example. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2014-87557 A Summary of the Invention [Problem to be solved by the invention]

[0005] The above-mentioned wetness detection method requires the repeated brushing of the detection area, such as the seat surface, with a broom for each seat. This takes time to complete the detection of all the seats. In addition, the physical burden on the worker performing the detection work is large.

[0006] One aspect of the present disclosure aims to provide a seat wetness detection system that can reduce the work time and physical burden involved in detecting seat wetness. [Means for solving the problem]

[0007] One aspect of the present disclosure is a wet seat detection system for detecting wetness of a seat, the wet seat detection system including an image acquisition unit configured to acquire a thermographic image showing a temperature distribution of a space including a seat and a visible light image of the space, an extraction unit configured to extract a seat area from the visible light image, and a determination unit configured to determine wetness of the seat based on the temperature distribution of the seat area extracted by the extraction unit.

[0008] According to this configuration, the wetness of the seat can be detected by acquiring a thermographic image and a visible light image including the seat. Therefore, the work time for wetness detection can be shortened. Furthermore, since physical contact with the seat is not required, the physical burden on the worker is reduced.

[0009] In one aspect of the present disclosure, the extraction unit may be configured to extract the seat area using a trained model generated by machine learning using a plurality of images with designated seat areas as training data. With this configuration, by using images including the seats to be detected as training data, it becomes possible to extract the seat area according to the shape, pattern, etc. of the seats to be detected.

[0010] In one aspect of the present disclosure, the determination unit may be configured to determine a wet spot on the seat. With this configuration, it is possible to grasp a spot that needs a recovery operation for the wet spot, thereby improving the efficiency of the cleaning operation.

[0011] In one aspect of the present disclosure, the extraction unit may be configured to exclude a specific area other than the seat surface from the seat area when extracting the seat area. With such a configuration, it is possible to prevent a part of the seat that does not need to be determined as wet from being detected as a wet area. As a result, the efficiency of the wetness detection work is improved. [Brief description of the drawings]

[0012] [Figure 1] FIG. 1 is a block diagram showing a schematic configuration of a wet seat detection system according to an embodiment. [Diagram 2]2A is an example of a thermographic image acquired by the image acquisition unit of FIG. 1, FIG. 2B is an example of a seat area extracted by the extraction unit of FIG. 1, and FIG. 2C is an example of a wet area determined by the determination unit of FIG. 1. [Diagram 3] FIG. 3 is a flow diagram illustrating an outline of the processing executed by the processing unit in FIG. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0013] Hereinafter, embodiments to which the present disclosure is applied will be described with reference to the drawings. [1. First embodiment] [1-1. Configuration] 1 detects wetness of a seat placed in a vehicle or facility. The wet seat detection system 1 includes an image acquisition unit 2, a processing unit 3, and a notification unit 4.

[0014] Examples of vehicles in which seats are arranged include public transportation facilities that transport a large number of passengers, such as railway cars, buses, airplanes, etc. Examples of facilities in which seats are arranged include facilities where a large number of passengers watch or experience content, such as movie theaters, concert halls, and theme parks.

[0015] Among these, in railway vehicles, the passenger compartment space in which the seats are arranged is narrow, and the seats need to be cleaned at short intervals during turnaround operations, so there is a particular demand for reducing the work time and physical burden involved in detecting wet seats in railway vehicles.

[0016] <Image acquisition section> The image acquisition unit 2 is configured to simultaneously acquire a thermographic image showing the temperature distribution in a space including the seats, and a visible light image (i.e., a photographic image) of the space. The visible light image has the same angle of view as the thermographic image. Multiple seats may be included in one capture (i.e., one thermographic image and one visible light image).

[0017] A photographing device equipped with a known thermography camera that uses analysis of infrared radiation and a visible light camera can be used as the image acquisition unit 2. The image acquisition unit 2 can be, for example, a portable type attached to a mobile terminal used by a worker, or a fixed type installed in the space where the seats are arranged (for example, the ceiling, etc.).

[0018] The image acquiring unit 2 may acquire a visible light image by extracting the visible light image from the thermography image.

[0019] Fig. 2A shows an example of a thermographic image acquired by the image acquisition unit 2. The image in Fig. 2A was captured by the image acquisition unit 2 of a seat made up of three connected seats arranged in a passenger compartment of a railway vehicle.

[0020] In the image of Figure 2A, dark (i.e. black) areas represent low temperatures, and light (i.e. white) areas represent high temperatures. Note that in actual thermography images, the temperature is shown in stages in color. The example image of Figure 2A is a color thermography image binarized to black and white.

[0021] <Processing section> The processing unit 3 is configured to execute a process of determining whether or not the seat is wet based on the thermographic image and the visible light image acquired by the image acquisition unit 2. The processing unit 3 has an extraction unit 31, a determination unit 32, and a storage unit 33.

[0022] The processing unit 3 is configured by, for example, a computer including a processor, a storage medium such as a RAM or a ROM, and an input / output unit. As the processing unit 3, for example, a mobile terminal such as a smartphone can be used.

[0023] (Extraction part) The extraction unit 31 is configured to extract a seat area from the visible light image. The "seat area" is an area for which wetness is detected, and includes, for example, a seat cushion and backrest, but does not include structures other than the seat, such as a floor, a wall, a window, or a ceiling.

[0024] In addition, when extracting the seat area, the extraction unit 31 excludes a specific area other than the seat surface from the seat area. The specific area is a predetermined area, and examples of the specific area include an armrest, a replaceable seat cover, and the like.

[0025] Fig. 2B shows an example of a seating area S extracted by the extraction unit 31 from a visible light image having the same angle of view as the thermographic image in Fig. 2A. In Fig. 2B, the cushions and backrests of three seats are extracted as the seating area S. On the other hand, in Fig. 2B, the floor and walls of the passenger compartment in which the seats are arranged, and the two armrests of the seats are excluded from the seating area S.

[0026] Specifically, the extraction unit 31 extracts the seat area by using a trained model generated by machine learning using a plurality of images with designated seat areas as training data (i.e., labeled data). The trained model is stored in the storage unit 33.

[0027] The trained model used by the extraction unit 31 is constructed by machine learning using a plurality of images including at least one seat, which are input data, and seat areas, which are labels for the input data.

[0028] The image used as input data may be a visible light image acquired by the image acquiring unit 2, or may be a visible light image captured by an image capturing device other than the image acquiring unit 2. Furthermore, the labeled data may include the same visible light image with a different orientation (e.g., rotated by 90 degrees).

[0029] The trained model is a classifier (i.e., a classification model) constructed by supervised machine learning, and is configured, for example, by a multilayer neural network. Note that a model other than a neural network may also be used as the trained model.

[0030] In the learning step of generating a trained model, a large amount of labeled data is analyzed by a machine learning device (not shown). The machine learning device learns features for classifying the range of seating areas from the large amount of labeled data, and constructs a trained model.

[0031] The trained model may be generated for each attribute of the seat to be detected (e.g., size, shape, pattern, space in which the seat is arranged, etc.). A general-purpose trained model that can handle multiple types of seats may also be generated. The machine learning device may be incorporated in the processing unit 3.

[0032] (Judgment Department) The determination unit 32 is configured to determine the wetness of the seat based on the temperature distribution of the seat area extracted by the extraction unit 31. Specifically, the determination unit 32 is configured to determine the wet spot on the seat and output the determined wet spot to the notification unit 4.

[0033] The wetness determination procedure by the determination unit 32 is as follows. First, the determination unit 32 calculates the average temperature of the seat area. Next, the determination unit 32 determines a part of the seat area whose temperature is lower than the average temperature by a predetermined threshold or more as a wet part W. The threshold can be, for example, 3°C.

[0034] Fig. 2C shows an example of a wet spot W determined by the determination unit 32 from the thermographic image of Fig. 2A. In Fig. 2C, there is a low-temperature (i.e., black-colored) area on the seat surface of the center seat, and this area is determined as the wet spot W by the determination unit 32.

[0035] 2A, the floor and walls around the seats also have low temperatures, but the floor and walls are not included in the seating area S. Therefore, the determination unit 32 does not determine the floor and walls as wet areas W. In addition, the armrest disposed between the two seats is also not included in the seating area S, and is therefore excluded from the wetness determination target.

[0036] <Notification section> The notification unit 4 notifies the result of the determination by the determination unit 32 (i.e., the presence or absence of a wet spot and its location) by image and / or sound. As the notification unit 4, for example, an output device (e.g., a display, a speaker, etc.) of a mobile terminal such as a smartphone can be used.

[0037] [1-2. Processing] Hereinafter, an example of processing executed by the processing unit 3 of the seat wetness detection system 1 will be described with reference to the flow diagram of FIG.

[0038] In this process, the processing unit 3 first resizes the visible light image to a size appropriate for extracting the seat area (step S10).

[0039] After resizing, the processing unit 3 extracts the seat area using the trained model (step S20). After extracting the seat area, the processing unit 3 extracts the temperature in the seat area by overlaying the seat area on the thermographic image (step S30).

[0040] After extracting the temperature, the processing unit 3 determines that a portion whose temperature is lower than the average temperature of the seat area by a threshold value or more is a wet portion (step S40).

[0041] [1-3. Effects] According to the embodiment described above in detail, the following effects can be obtained. (1a) By acquiring a thermographic image and a visible light image including the seat, the wetness of the seat can be detected. Therefore, the work time for wetness detection is shortened. Furthermore, since physical contact with the seat is not required, the physical burden on the worker is reduced.

[0042] (1b) Since the extraction unit 31 extracts the seat area using a trained model, by using an image including the seat to be detected as training data, it becomes possible to extract the seat area according to the shape, pattern, etc. of the seat to be detected.

[0043] (1c) The determining unit 32 determines the wet locations, making it possible to grasp the locations that require recovery work to be performed to correct the wetness, thereby improving the efficiency of the cleaning work.

[0044] (1d) By excluding the specific region from the seat region by the extraction unit 31, it is possible to prevent a portion of the seat that does not need to be determined as wet from being detected as a wet portion, thereby improving the efficiency of the wetness detection process.

[0045] 2. Other embodiments Although the embodiments of the present disclosure have been described above, it goes without saying that the present disclosure is not limited to the above-described embodiments and can take various forms.

[0046] (2a) In the seat wetness detection system 1 of the above embodiment, the extraction unit does not necessarily need to use a trained model generated by machine learning using multiple images with designated seat areas as teacher data. For example, the extraction unit may use a trained model generated by unsupervised learning. In addition, the extraction unit may extract seat areas using an algorithm other than the trained model.

[0047] (2b) In the seat wetness detection system 1 of the above embodiment, the determination unit does not necessarily have to determine the wet spot on the seat. For example, the determination unit may determine only the presence or absence of wetness on the seat.

[0048] (2c) In the seat wetness detection system 1 of the above embodiment, the extraction unit does not necessarily have to exclude the specific area from the seat area. In other words, the extraction unit may always include the entire seat in the seat area.

[0049] (2d) The function of one component in the above embodiments may be distributed among multiple components, or the functions of multiple components may be integrated into one component. In addition, a part of the configuration of the above embodiments may be omitted. In addition, at least a part of the configuration of the above embodiments may be added to or substituted for the configuration of another of the above embodiments. Note that all aspects included in the technical idea identified from the wording of the claims are embodiments of the present disclosure. [Explanation of symbols]

[0050] 1...seat wetness detection system, 2...image acquisition unit, 3...processing unit, 4...notification unit, 31...extraction section, 32...judgment section, 33...storage section.

Claims

1. A seat wetness detection system for detecting seat wetness, comprising: an image acquisition unit configured to acquire a thermographic image showing a temperature distribution in a space including the seat and a visible light image of the space; an extractor configured to extract a seating area from the visible light image; a determination unit configured to determine wetness of the seat based on the temperature distribution of the seat area extracted by the extraction unit; Equipped with The extraction unit is configured to exclude a specific area other than a seat surface of the seat from the seat area when extracting the seat area. Wet seat detection system.

2. 2. The seat wetness detection system of claim 1, The seat wetness detection system is configured such that the extraction unit extracts the seat area using a trained model generated by machine learning using a plurality of images in which the seat area is specified as training data.

3. 3. A seat wetness detection system according to claim 1 or 2, comprising: A seat wetness detection system, wherein the determination unit is configured to determine a wet spot on the seat.

Images