Estimation device

Abstract

To provide an estimation device for easily estimating the thermal sensation of a person in a building on the basis of a skin area and skeleton information.SOLUTION: An acquisition unit 101 that acquires image information from a camera 300 that captures an image including a person who is present in a building, a skin region unit 103 that extracts a skin region of the person who is present in the building on the basis of the image information, a skeleton information unit 104 that obtains skeleton information of the person who is present in the building on the basis of the image information, and a coordinate unit 105 that obtains forearm coordinates including coordinates of an elbow and coordinates of a wrist of a forearm of the person who is present in the building and a plurality of discrete intermediate coordinates between the elbow and the wrist on the basis of the skeleton information; A sleeve roll-up estimation unit 106 that estimates a sleeve roll-up amount on the basis of the skin region and the forearm coordinates, and a change unit 108 that changes a predetermined set value in control of the air conditioning device 500 on the basis of the estimated sleeve roll-up amount.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to an estimation device. [Background technology]

[0002] A known method for automatically adjusting the settings of air conditioning equipment installed in an office is to use a camera installed in the office to photograph the amount of clothing worn by the occupants, estimate the thermal sensation of the occupants from the amount of clothing worn by the occupants, and control the air conditioning equipment based on the estimated thermal sensation of the occupants (see, for example, Patent Document 1). (For example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2023-138228 Summary of the Invention [Problem to be solved by the invention]

[0004] When performing the above-mentioned control, the thermal sensation of the occupants is estimated based on the result of estimating whether the sleeves of the clothes are long or short. However, the above-mentioned method has a problem in that the estimation accuracy of the thermal sensation may be low in some cases.

[0005] The present invention has been made to solve the above-mentioned problems, and aims to provide an estimation device that can easily estimate the thermal sensation of an occupant based on skin area and skeletal information. [Means for solving the problem]

[0006] In order to achieve the above object, the present invention provides the following means. An estimation device according to one aspect of the present invention is an estimation device that estimates the temperature environment of a target space, and is characterized by comprising: an acquisition unit that acquires image information from a camera that captures images including occupants; a skin area unit that extracts the skin area of ​​the occupants based on the image information; a skeletal information unit that determines skeletal information of the occupants based on the image information; a coordinate unit that determines forearm coordinates including the elbow coordinate, wrist coordinate, and a plurality of discrete intermediate coordinates between the elbow and the wrist of the occupant's forearm based on the skeletal information; a sleeve roll estimation unit that estimates the amount of sleeve roll based on the skin area and the forearm coordinates; and a change unit that changes a predetermined setting value for controlling an air conditioning unit based on the estimated amount of sleeve roll.

[0007] According to the estimation device of the first aspect of the present invention, the skin area and skeletal information of an inhabitant are acquired. Based on the acquired skeletal information, multiple coordinates from the elbow to the wrist can be determined. From the determined multiple coordinates and the acquired skin area, the extent to which the inhabitant's sleeves are rolled up can be estimated. Based on the estimated extent to which the sleeves are rolled up, the setting values ​​of the air conditioning device can be changed.

[0008] Specifically, the multiple coordinates are preferably five coordinates from the wrist to the elbow. The five coordinates are preferably as follows: The first is the coordinate of the wrist. The second is the coordinate of the elbow. The third is the coordinate of the midpoint between the wrist and the elbow (hereinafter also referred to as the second midpoint). The fourth is the coordinate of the midpoint between the wrist and the second midpoint (hereinafter also referred to as the first midpoint). The fifth is the coordinate of the midpoint between the elbow and the second midpoint (hereinafter also referred to as the third midpoint). Note that the number of coordinates may be more or less than five, and may be coordinate points other than those mentioned above.

[0009] In the first aspect of the invention, it is preferable that the sleeve rolling estimation unit estimates the amount of sleeve rolling based on whether or not each coordinate included in the forearm coordinates is included in the skin area.

[0010] In this way, the sleeve-rolledness estimation unit can estimate how much the occupant's sleeves are rolled up by determining whether the extracted range of the skin region includes at least one coordinate. In the first aspect of the invention, it is preferable that the change unit changes the set value to a lower value as the estimated amount of rolling up of the sleeves increases.

[0011] In this way, the change unit can change the setting value of the air conditioner based on the amount of rolled-up sleeves. Specifically, when the amount of rolled-up sleeves is large, the change unit can control to lower the setting value of the air conditioner. Note that a large amount of rolled-up sleeves refers to when multiple forearm coordinates are included in the skin region area.

[0012] In the first aspect of the above invention, it is preferable that a clothing estimation unit is further provided that estimates the type of clothing worn by the occupant based on the image information, and that the change unit changes the setting value based on the estimated amount of sleeve roll-up and the estimated type of clothing.

[0013] In this way, by providing a clothing estimation unit, it is possible to estimate the type of clothing worn by the occupants, and more specifically, the sleeve length of the clothing worn by the occupants. [Effects of the Invention]

[0014] According to the estimation device of the present invention, it is possible to estimate the extent to which the sleeves are rolled up based on the skin region and skeletal information, and to estimate the thermal sensation of the person in the building based on the estimated extent to which the sleeves are rolled up. [Brief explanation of the drawings]

[0015] [Figure 1] 1 is a block diagram illustrating a configuration of an estimation device according to a first embodiment of the present invention. [Figure 2] 10A and 10B are schematic diagrams for explaining processing by a skeleton information unit. [Figure 3] FIG. 10 is a schematic diagram for explaining processing by a coordinate unit. [Figure 4] FIG. 2 is a schematic diagram for explaining the configuration of a camera. [Figure 5] 3 is a flowchart illustrating processing performed by the estimation device according to the first embodiment of the present invention. [Figure 6] FIG. 4 is a block diagram illustrating the configuration of an estimation device according to a second embodiment of the present invention. [Figure 7] 10 is a flowchart illustrating processing performed by an estimation device according to a second embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0016] [First embodiment] Configuration description An estimation device 100 according to a first embodiment of the present invention will be described with reference to Fig. 1 to Fig. 5. The estimation device 100 of this embodiment is a device for estimating the thermal sensation of an occupant in a building. It is also possible to control an air conditioning device 500 based on the estimated environmental temperature.

[0017] The air conditioner 500 is configured to take in indoor air of a building, cool it, and supply the cooled air to the indoors of the building. Specifically, the air is cooled by circulating a refrigerant between the air and an outdoor unit (not shown) to exchange heat between the air and the refrigerant. The outdoor unit is configured so that the refrigerant removes heat from the air and releases it to the outside air.

[0018] The air conditioner 500 is provided with a heat exchanger (not shown) that cools the intake air by heat exchange, and an air conditioner fan (not shown) that sends out the heat-exchanged air. The heat exchanger and the air conditioner fan have known configurations.

[0019] The air conditioner 500 is preferably installed in an office. However, the air conditioner 500 may be installed in buildings other than offices. For example, it may be installed in a house or a school classroom. Furthermore, the air conditioner 500 may be installed in a means of transportation other than a building.

[0020] The air conditioner 500 has a configuration that allows it to be connected so as to be able to communicate with the estimation device 100. Specifically, it is preferable that the air conditioner 500 be controlled based on a signal transmitted from the estimation device 100.

[0021] An inhabitant is a person who is in a room that is the target of air conditioning. The estimation device 100 is an information processing device such as a server or computer having a CPU (central processing unit), ROM, RAM, input / output interface, etc. The estimation device 100 is connected to a plurality of air conditioners 500 and cameras 300 so that information can be communicated therewith. The number of air conditioners 500 and cameras 300 connected to the estimation device 100 may be one or more.

[0022] As shown in FIG. 1 , the estimation device 100 has stored in the ROM or the like a program that causes the CPU, ROM, RAM, and input / output interface to cooperate with each other to function as at least an acquisition unit 101, a memory unit 102, a skin region unit 103, a skeleton information unit 104, a coordinate unit 105, a rolled-up sleeve estimation unit 106, a thermal sensation estimation unit 107, a change unit 108, and a control unit 109.

[0023] The acquisition unit 101 is connected so as to be able to communicate and acquire image information captured by a camera 300 (described later). The acquisition unit 101 has a function of acquiring image information including images of people present in the building captured by the camera 300.

[0024] The image information preferably includes at least the skin area and skeletal information of the occupants, which will be described later. The storage unit 102 is an information storage medium and has a function of storing information for estimating the thermal sensation of occupants. The stored information preferably includes information acquired by the acquisition unit 101. The storage unit 102 may be a flash memory such as an SD memory card, or may be a recording medium of another type.

[0025] The skin region unit 103 has a function of extracting the skin region of an inhabitant based on the image information acquired by the acquisition unit 101. Specifically, the skin region unit 103 preferably extracts the skin region of an inhabitant by using a segmentation model.

[0026] The skin region is the exposed part of the skin of the occupant. In this embodiment, the skin region is the exposed part of the skin from the hand to the elbow (hereinafter also referred to as the forearm) of the occupant.

[0027] The segmentation model is a model that uses machine learning to divide image information into several objects. In this embodiment, it is preferable to divide the image into a skin region and other regions. Note that a known learning method can be used as the machine learning method.

[0028] 2, the skeletal information unit 104 has a function of acquiring skeletal information of the occupant based on the image information acquired by the acquisition unit 101. Specifically, it is preferable that the skin region unit 103 acquires the skeletal information of the occupant by using a posture estimation model.

[0029] The posture estimation model is a model that can detect a posture of a person by connecting the joint points from image information that includes a person using machine learning, in real time from the image information. In this embodiment, it is preferable to detect a posture of a person by connecting the joint points around the arms. Note that a known learning method can be used for machine learning.

[0030] The coordinate unit 105 can calculate a plurality of forearm coordinates based on the skeletal information. Specifically, it is preferable that the coordinate unit 105 calculates a plurality of forearm coordinates based on the skeletal reliability calculated from the skeletal information.

[0031] In this embodiment, as shown in FIG. 3 , the multiple forearm coordinates are five coordinates from the wrist to the elbow. The first is wrist coordinate A. The second is elbow coordinate E. The third is the coordinate of a point midway between wrist coordinate A and elbow coordinate E (hereinafter also referred to as second intermediate point C). The fourth is the coordinate of a point midway between wrist coordinate A and second intermediate point C (hereinafter also referred to as first intermediate point B). The fifth is the coordinate of a point midway between elbow coordinate E and second intermediate point C (hereinafter also referred to as third intermediate point D). In other words, the forearm coordinates are arranged in the following order: wrist coordinate A, first intermediate point B, second intermediate point C, third intermediate point D, and elbow coordinate E. Note that four or more coordinates are preferable, but more or less than five may be used, and coordinate points other than those mentioned above may also be used.

[0032] Skeleton reliability is a value that indicates the accuracy of skeletal information. The closer the value of skeleton reliability is to 1, the higher the reliability, and the closer it is to 0, the lower the reliability. Skeleton reliability is calculated for each forearm coordinate.

[0033] The sleeve roll-up estimation unit 106 has a function of estimating the extent to which the inhabitant's sleeves are rolled up based on the skin area and the forearm coordinates. Specifically, the sleeve roll-up estimation unit 106 can estimate the extent to which the inhabitant's sleeves are rolled up by determining whether the forearm coordinates are included in the exposed skin area.

[0034] In this embodiment, the amount of rolling up the sleeves is preferably expressed by a score ranging from 0 to 4 points. If the skin area includes only the coordinates of the wrist, the amount of rolling up the sleeves is 0 points. If the skin area includes from wrist coordinate A to first intermediate point B, the amount of rolling up the sleeves is 1 point. If the skin area includes from wrist coordinate A to second intermediate point C, the amount of rolling up the sleeves is 2 points. If the skin area includes from wrist coordinate A to third intermediate point D, the amount of rolling up the sleeves is 3 points. If the skin area includes from wrist coordinate A to elbow coordinate E, the amount of rolling up the sleeves is 4 points. In other words, the amount of rolling up the sleeves increases as the score increases.

[0035] The thermal sensation estimation unit 107 has a function of estimating the thermal sensation of the occupants based on the extent to which the occupants' sleeves are rolled up. In this embodiment, the thermal sensation estimation unit 107 can estimate the thermal sensation of the occupants based on the score of the extent to which the occupants' sleeves are rolled up.

[0036] Specifically, the thermal sensation estimation unit 107 determines that the indoor temperature is comfortable for the occupant when the number of rolled sleeves is 0. As the number of rolled sleeves increases, the thermal sensation estimation unit 107 determines that the indoor temperature is hot for the occupant.

[0037] The change unit 108 has a function of transmitting a signal to change the setting value of the air conditioner 500 based on the estimated thermal sensation of the occupants. A known technique may be used as a specific means for transmitting the signal.

[0038] As shown in Fig. 4, at least one camera 300 is installed in a room so as to capture images of people present in the room. Furthermore, camera 300 is configured to be able to transmit captured image information to estimation device 100.

[0039] The camera 300 may be a camera 300 having the functions of an existing camera 300 in a building, may be incorporated into the estimation device 100, or may be configured independently. In this embodiment, a case will be described in which the camera 300 is configured independently of the estimation device 100.

[0040] Furthermore, camera 300 is preferably installed in an upper space so that the room to be photographed and the people present in the room can be viewed from above. Specifically, camera 300 is preferably installed on the ceiling of the room to be photographed. However, camera 300 may also be installed somewhere other than the ceiling of the room to be photographed.

[0041] Description of action Next, the operation of the estimation device 100 configured as described above will be described. First, the air conditioner 500 will be described, second, the camera 300 will be described, third, the processing of the estimation device 100 will be described, and fourth, the learning method of the learning model will be described.

[0042] First, the air conditioner 500 rotates the air conditioning fan to draw air from the office room into the air conditioner 500. The air that is drawn in includes air that has had heat radiated from people in the building and air that has been exhausted from electronic devices installed in the office.

[0043] The drawn-in air is cooled in a heat exchanger (not shown). Specifically, the temperature of the drawn-in air is lowered as heat is absorbed by the refrigerant circulating between the air and an outdoor unit (not shown). The refrigerant that has absorbed heat then releases the heat to the outside air in the outdoor unit. The refrigerant that has released the heat then absorbs heat from the drawn-in air again in the heat exchanger.

[0044] The air cooled in the heat exchanger is sent into the office room by an air conditioner fan (not shown). The air sent into the office room is conditioned based on the PMV value to ensure comfort for the occupants. The air sent into the office room is then drawn back into the air conditioner 500.

[0045] The air to be cooled is preferably air that brings the ambient temperature in the room to the set temperature estimated by the estimation device 100. Next, the control of camera 300 will be described. When camera 300 is powered on, it continues to capture images of the room. When an inhabitant is captured, camera 300 transmits image information of the inhabitant to estimation device 100. In other words, every time an inhabitant is captured, camera 300 transmits image information of the inhabitant to estimation device 100.

[0046] Next, the control in the estimation device 100 will be described with reference to FIG. When processing in the estimation device 100 starts, the acquisition unit 101 performs processing to acquire image information from the camera 300 at any timing (S1). The image information may include one person present in the building, or may include multiple people present in the building.

[0047] The image information acquired by the acquisition unit 101 is stored in the storage unit 102. Based on the stored image information, the skin region unit 103 performs a process of extracting the skin region of the occupant (S2). Specifically, the process of extracting the skin region of the occupant's forearm is performed.

[0048] Once the skin region is extracted, the skeleton information unit 104 performs processing to acquire skeleton information of the occupants (S3). When the skeletal information is acquired, the coordinate unit 105 calculates the forearm coordinates based on the skeletal information of the person (S4). Specifically, a process is performed to calculate the coordinate A of the wrist, the first intermediate point B, the second intermediate point C, the third intermediate point D, and the coordinate E of the elbow.

[0049] Once the forearm coordinates have been calculated, the sleeve rolled-up estimation unit 106 performs processing to estimate the degree to which the sleeves of the occupants are rolled up, based on the skin area and the forearm coordinates (S5). Specifically, the sleeve rolled-up estimation unit 106 estimates a score for the degree to which the sleeves are rolled up for each occupant included in the image information.

[0050] Once the extent of sleeve rolling has been estimated, the thermal sensation estimation unit 107 performs processing to estimate the thermal sensation of the occupant (S6). Specifically, the thermal sensation estimation unit 107 determines that the higher the score of the extent of sleeve rolling, the hotter the occupant feels. If the score of the extent of sleeve rolling is 0, it determines that the occupant's thermal sensation is appropriate.

[0051] Once the thermal sensation of the occupants has been calculated, the change unit 108 estimates the set temperature of the air conditioner 500 based on the estimated thermal sensation of the occupants (S7). Once the set temperature has been estimated, the control unit 109 performs processing to transmit a signal to change the set temperature of the air conditioner 500 based on the estimated set temperature (S8).

[0052] Next, we will explain the machine learning of the learning model. In this embodiment, an example will be described in which machine learning of a learning model is performed in an information processing device different from the estimation device 100. The learning model that has undergone machine learning is stored in the storage unit 102 before processing by the estimation device 100.

[0053] Furthermore, after control is performed by the estimation device 100, a model that has undergone further machine learning may be stored in the storage unit 102. In this case, the previously stored learning model is replaced with the learning model that has undergone further machine learning.

[0054] Note that machine learning of the learning models may be performed in different information processing devices as described above, or may be performed in the estimation device 100. When machine learning is performed in the estimation device 100, a machine learning unit that performs machine learning is provided in the estimation device 100. Alternatively, machine learning for one of the learning models may be performed in different information processing devices, and machine learning for the other may be performed in the estimation device 100.

[0055] In the machine learning of the learning model, when a person in a building moves, training data is used to determine that the person is the same person before and after the movement. Specifically, training data that determines the person is the same person based on the distance the person moved estimated from previous and next frames of the image may be used, or training data that determines the person is the same person based on similarity may be used.

[0056] Regarding the specific machine learning in the learning model, known supervised learning can be used, and the specific content of the calculation process in supervised learning is not limited. Also, regarding the method of creating the teacher data for the learning model, known creation methods can be used, and the specific creation method is not limited.

[0057] Description of effects The estimation device 100 configured as described above can acquire skeletal information and skin area. Furthermore, multiple forearm coordinates from the elbow to the wrist can be obtained based on the skeletal information. The extent to which the occupant's sleeves are rolled up can be estimated from the multiple forearm coordinates obtained and the acquired skin area. The estimation device 100 can estimate the occupant's thermal sensation based on the estimated extent to which the sleeves are rolled up. By performing the above-described processing and calculating the setting values ​​of the air conditioning device 500, it is possible to estimate the occupant's thermal sensation with high accuracy.

[0058] In this way, the sleeve-rolledness estimation unit 106 can estimate the extent to which the occupant's sleeves are rolled up by determining whether the extracted range of the skin region includes at least one or more forearm coordinates.

[0059] In this way, the changing unit 108 can change the setting value of the air conditioner 500 based on the amount of sleeves rolled up. Specifically, when the amount of sleeves rolled up is large, the thermal sensation estimating unit 107 determines that the thermal sensation of the occupant is hot. Based on the determined thermal sensation, the changing unit 108 can perform control to lower the setting value of the air conditioner 500.

[0060] <Second embodiment> Next, a second embodiment of the present invention will be described with reference to Fig. 6 and Fig. 7. The basic configuration of estimation device 100A of this embodiment is basically similar to that of estimation device 100 of the first embodiment, except that a clothing estimation unit 110A is further provided. In this embodiment, only the configuration that differs from the first embodiment will be described, and a description of the same configuration will be omitted.

[0061] As shown in FIG. 6, the estimation device 100A has stored in the ROM or the like a program that causes the CPU, ROM, RAM, and input / output interface to cooperate with each other to function as at least an acquisition unit 101, a memory unit 102, a skin region unit 103, a skeleton information unit 104, a coordinate unit 105, a rolled-up sleeve estimation unit 106, a thermal sensation estimation unit 107A, a change unit 108, a control unit 109, and an clothing estimation unit 110A.

[0062] The thermal sensation estimation unit 107A has a function of estimating the thermal sensation of the occupants based on the extent to which the occupants' sleeves are rolled up and the type of clothing worn by the occupants (described later). Specifically, the thermal sensation estimation unit 107A estimates the thermal sensation of the occupants based on the score of the sleeve length and the extent to which the sleeves are rolled up of the clothing worn by the occupants.

[0063] The clothing estimation unit 110A has a function of estimating the type of clothing worn by the occupant. In this embodiment, the clothing estimation unit 110A has a function of estimating the type of sleeve length of the clothing worn by the occupant. Furthermore, the clothing estimation unit 110A has a function of estimating the type of clothing worn by the occupant by determining whether the clothing is short-sleeved clothing, long-sleeved clothing, or short-sleeved clothing, or whether it is three-quarter-length clothing or eight-quarter-length clothing, etc. Note that the clothing estimation unit 110A may also determine clothing with sleeve lengths other than three-quarter-length clothing or eight-quarter-length clothing, or whether it is short-sleeved clothing, long-sleeved clothing, or short-sleeved clothing.

[0064] Description of action Next, the processing of the estimation device 100A configured as described above will be described with reference to Fig. 7. Note that the description of the processing up to the calculation of the skin region by the air conditioner 500, the camera 300, and the estimation device 100, which is the same as the operations of the first embodiment, will be omitted.

[0065] Once the forearm coordinates are calculated, clothing estimation unit 110A performs a process of estimating the type of clothing worn by the occupant (S11). Specifically, the type of sleeve length of the clothing worn by the occupant is estimated.

[0066] Once the type of clothing is determined, the skeleton information unit 104 performs processing to acquire skeleton information of the person in the building (S12). When the skeletal information is acquired, the coordinate unit 105 calculates the forearm coordinates based on the skeletal information of the person in the building (S13). Specifically, a process is performed to calculate the coordinate A of the wrist, the first intermediate point B, the second intermediate point C, the third intermediate point D, and the coordinate E of the elbow.

[0067] Once the forearm coordinates have been calculated, the sleeve rolled-up estimation unit 106 performs processing to estimate the degree to which the sleeves of the occupants are rolled up, based on the skin area and the forearm coordinates (S14). Specifically, the sleeve rolled-up estimation unit 106 estimates a score for the degree to which the sleeves are rolled up for each occupant included in the image information.

[0068] Once the extent to which the sleeves are rolled up has been estimated, the thermal sensation estimation unit 107 performs processing to estimate the thermal sensation of the occupant based on the extent to which the sleeves are rolled up and information about the clothing (S15). Once the thermal sensation of the occupants has been calculated, the change unit 108 estimates the set temperature of the air conditioner 500 based on the estimated thermal sensation of the occupants (S16). Once the set temperature has been estimated, the control unit 109 performs processing to transmit a signal to change the set temperature of the air conditioner 500 based on the estimated set temperature (S17).

[0069] Description of effects In this way, the estimation device 100 can change the setting value of the air conditioner 500 for each sleeve length of clothing. Specifically, the thermal sensation estimation unit 107 can estimate the thermal sensation of the occupant based on at least the type of clothing worn by the occupant estimated by the clothing estimation unit 110A. The setting value of the air conditioner 500 can be changed based on the estimated thermal sensation. Therefore, the thermal sensation of the occupant can be estimated with higher accuracy than in the first embodiment.

[0070] In this way, clothing estimation unit 110A can estimate the clothing type of clothing with sleeve lengths that do not fall into either the long-sleeve or short-sleeve categories, in addition to the clothing types of long-sleeve and short-sleeve clothing.

[0071] The technical scope of the present invention is not limited to the above-described embodiments, and various modifications can be made without departing from the spirit of the present invention. For example, the present invention is not limited to applications of the above-described embodiments, and may be applied to embodiments in which these embodiments are appropriately combined, and is not particularly limited. [Explanation of symbols]

[0072] 100, 100A...estimation device, 101...acquisition unit, 102...memory unit, 103...skin area unit, 104...skeletal information unit, 105...coordinate unit, 106...rolled-sleeve estimation unit, 107, 107A...thermal sensation estimation unit, 108...change unit, 109...control unit, 110A...clothing estimation unit, 300...camera, 500...air conditioning unit.

Claims

1. An estimation device for estimating a temperature environment in a target space, an acquisition unit that acquires image information from a camera that captures images including people in the building; a skin area unit that extracts a skin area of ​​the person based on the image information; a skeletal information unit that obtains skeletal information of the occupants based on the image information; a coordinate unit that calculates a forearm coordinate including an elbow coordinate, a wrist coordinate, and a plurality of discrete intermediate coordinates between the elbow and the wrist of the person based on the skeletal information; a sleeve roll-up estimation unit that estimates a sleeve roll-up amount based on the skin area and the forearm coordinates; and a change unit that changes a predetermined setting value in control of an air conditioner based on the estimated amount of sleeves rolled up.

2. 2. The estimation device according to claim 1, wherein the sleeve roll-up estimation unit estimates the amount of sleeve roll-up based on whether or not each coordinate included in the forearm coordinates is included in the skin area.

3. 2. The estimation device according to claim 1, wherein the change unit changes the set value to a lower value as the estimated amount of rolled-up sleeves increases.

4. a clothing estimation unit that estimates the type of clothing worn by the person in the building based on the image information; 2. The estimation device according to claim 1, wherein the change unit changes the set value based on the estimated amount of rolled-up sleeves and the estimated type of clothing.

Images