Warm / cold feeling prediction device, program, and warm / cold feeling prediction method

Abstract

A warm / cold feeling prediction device (100) comprises: a personal region detection unit (104) that detects, from a thermal image indicating the temperature distribution in a predetermined space, a personal region that is narrower than the predetermined space and is a region of an individual present in the predetermined space; a local humidity estimation unit (106) that estimates a local humidity, which is the humidity of the personal region; and a warm / cold feeling prediction unit (107) that predicts the warm / cold feeling of the individual by using the local humidity.

Description

Thermal sensation prediction device, program, and thermal sensation prediction method

[0001] The present disclosure relates to a thermal sensation prediction device, a program, and a thermal sensation prediction method.

[0002] A thermal sensation estimation device is known that estimates thermal sensation, which is the sensation of a person feeling hot or cold, without the person reporting it and without directly placing a sensor on the human body. When the thermal sensation estimation device is installed in an air conditioning system, for example, the air volume can be controlled based on the estimated thermal sensation, thereby making the air conditioning system operate efficiently while keeping the person comfortable.

[0003] As a technology for estimating the thermal sensation of people crowded together, for example, there is an air conditioning system for railway vehicles described in Patent Document 1.

[0004] Patent Publication No. 5361816

[0005] In an enclosed space such as a room where many people gather, people feel hotter than usual due to heat radiation from the surfaces of the surrounding people's bodies and increased humidity caused by the sweating and exhalation of the surrounding people.

[0006] Conventional techniques estimate the thermal sensation of crowded people based on heat radiation from the surfaces of the surrounding people's bodies, but do not take into account the increase in humidity due to the sweating and exhalation of the surrounding people. Therefore, conventional techniques cannot accurately estimate the thermal sensation of crowded people. On the other hand, installing multiple humidity sensors to detect the increase in humidity due to the sweating and exhalation of the surrounding people is costly.

[0007] Therefore, one or more aspects of the present disclosure aim to enable accurate estimation of the thermal sensation of people crowded together without incurring additional costs.

[0008] A thermal sensation prediction device according to one aspect of the present disclosure is characterized by comprising: a personal area detection unit that detects a personal area, which is an area of ​​an individual present in a predetermined space that is smaller than the predetermined space, from a thermal image showing the distribution of temperature in the predetermined space; a local humidity estimation unit that estimates local humidity, which is the humidity in the personal area; and a thermal sensation prediction unit that predicts the thermal sensation of the individual using the local humidity.

[0009] A program according to one aspect of the present disclosure is characterized in that it causes a computer to function as a personal area detection unit that detects a personal area, which is an area of ​​an individual present in a predetermined space that is smaller than the predetermined space, from a thermal image showing the distribution of temperature in the predetermined space, a local humidity estimation unit that estimates local humidity, which is the humidity in the personal area, and a thermal sensation prediction unit that predicts the thermal sensation of the individual using the local humidity.

[0010] A thermal sensation prediction method according to one aspect of the present disclosure is characterized in that a personal area, which is an area of ​​an individual present in a predetermined space that is smaller than the predetermined space, is detected from a thermal image showing the distribution of temperature in the predetermined space, local humidity, which is the humidity in the personal area, is estimated, and the thermal sensation of the individual is predicted using the local humidity.

[0011] According to one or more aspects of the present disclosure, it is possible to accurately estimate the thermal sensation of people crowded together at low cost.

[0012] 1 is a block diagram showing an outline of the configuration of a thermal sensation prediction device according to a first embodiment. (A) and (B) are schematic diagrams for explaining the process of identifying a human region. (B) is a histogram of temperatures shown in a thermal image. (C) is a schematic diagram for explaining a human body region. (A) and (B) are schematic diagrams for explaining the process of identifying a personal region. (C) is a schematic diagram for explaining the process of identifying an individual and the number of people around the individual. (C) is a graph showing the relationship between a view factor and human density. (D) is a graph showing the relationship between an indoor temperature and the amount of human body water vapor generation. (D) is a graph showing the relationship between an indoor temperature and the amount of saturated water vapor. (A) and (B) are block diagrams showing an example of a hardware configuration. (D) is a flowchart showing the operation of a thermal sensation prediction device according to a first embodiment. (E) is a block diagram showing an outline of the configuration of a thermal sensation prediction device according to a second embodiment. (F) is a graph showing the relationship between an indoor temperature and the amount of personal water vapor generation for each physique. (F) is a flowchart showing the operation of a thermal sensation prediction device according to a second embodiment.

[0013] 1 is a block diagram showing a schematic configuration of a thermal sensation prediction device 100 according to embodiment 1. The thermal sensation prediction device 100 includes a thermal image acquisition unit 101, an indoor temperature acquisition unit 102, an indoor humidity acquisition unit 103, a personal area detection unit 104, a radiation temperature calculation unit 105, a local humidity estimation unit 106, and a thermal sensation prediction unit 107.

[0014] The thermal image acquisition unit 101 acquires a thermal image showing the temperature distribution in a predetermined space. It is assumed that a person is present in the space. For example, the thermal image acquisition unit 101 may acquire a thermal image from a thermal camera connected to a communication I / F (Interface) or connection I / F (not shown). If a thermal image is already stored in a storage unit (not shown), the thermal image acquisition unit 101 may acquire the thermal image from the storage unit. The thermal camera may be provided in an air conditioner (not shown) installed in the space where the person is present. The acquired thermal image is provided to the personal area detection unit 104.

[0015] The indoor temperature acquisition unit 102 acquires the temperature of a space where people are present as the indoor temperature. For example, the indoor temperature acquisition unit 102 may acquire the indoor temperature from a thermometer connected to a communication I / F or connection I / F (not shown), or, if the indoor temperature is already stored in a storage unit (not shown), may acquire the indoor temperature from the storage unit. The thermometer may be provided in an air conditioner (not shown) installed in the space where people are present. The indoor temperature is notified to the thermal sensation prediction unit 107.

[0016] The indoor humidity acquisition unit 103 acquires the humidity of a space where a person is present as the indoor humidity. For example, the indoor humidity acquisition unit 103 may acquire the indoor humidity from a hygrometer connected to a communication I / F or connection I / F (not shown), or, if the indoor humidity is already stored in a storage unit (not shown), may acquire the indoor humidity from the storage unit. The hygrometer may be provided in an air conditioner (not shown) installed in the space where a person is present. The indoor humidity is notified to the thermal sensation prediction unit 107.

[0017] The personal area detection unit 104 detects personal areas, which are areas of individual people present in a space, from the thermal image. For example, the personal area detection unit 104 detects personal areas, which are areas of individuals present in a predetermined space that is smaller than the predetermined space, from the thermal image.

[0018] Specifically, the personal area detection unit 104 first inputs the thermal image into a pre-trained model to identify the human area, which is the area of ​​a person included in the thermal image. The model used here is a model trained to be able to identify the human area from the thermal image using training data including the thermal image and ground truth data indicating the human area included in the thermal image.

[0019] For example, the personal area detection unit 104 inputs a thermal image 120 as shown in Figure 2(A) into a model, and identifies human areas 122A to 122F, which are rectangular areas that include each of the people 121A to 121F appearing in the thermal image 120, as shown in Figure 2(B).

[0020] Next, the personal area detection unit 104 calculates the average pixel value T ｍ Calculate.

[0021] Next, the personal area detection unit 104 generates a histogram of the temperatures shown in the thermal image and calculates the average value T ｍ A threshold temperature T is a temperature that is a threshold for distinguishing between a human body region including the θ Here, the personal area detection unit 104 uses a known binarization method such as Otsu's binarization method, mode method, P-tile method, or discriminant analysis method to determine the threshold temperature T θ It is sufficient to identify the following.

[0022] Specifically, as shown in Fig. 3, in the histogram of temperatures shown in the thermal image, the area corresponding to the human body area has a higher temperature than the area corresponding to the background area. ｍ A threshold temperature T for distinguishing between a human body region including the θ can be identified by using a known binarization method.

[0023] Then, the personal area detection unit 104 detects a threshold temperature T θ The area with a temperature higher than the threshold temperature T θ For example, the personal area detection unit 104 determines the area having the following temperature as the background area: θ In FIG. 4, the hatched area is the human body area.

[0024] Next, the personal area detection unit 104 identifies pixels in the human body area that have local maximum values ​​as face pixels. This is because, in a human being, the face has the highest temperature, and therefore, one local maximum value in the human body area is considered to correspond to one person. For example, as shown in FIG. 5A, the personal area detection unit 104 can identify face pixels 123A-123F by identifying local maximum values ​​in the human body area in the vertical or horizontal direction, in other words, in the x or y direction shown in FIG. 5A. Each of the face pixels 123A-123F may be a single pixel or a collection of multiple pixels.

[0025] The personal area detection unit 104 then identifies the position of the individual in the thermal image from the positions of each of the face pixels 123A-123F in the thermal image, and identifies the personal area that includes each of the face pixels 123A-123F. For example, as shown in FIG. 5B, the personal area detection unit 104 identifies the personal areas 124A-124F by using frames that become larger the closer the positions of the face pixels 123A-123F are to the thermal camera. Note that in the thermal camera that captures the thermal image, each pixel is assumed to be associated in advance through experiments or the like with the position of the individual when that pixel matches a face pixel. Hereinafter, one personal area to be processed out of one or more personal areas detected by the personal area detection unit 104 is also referred to as a target personal area.

[0026] The radiation temperature calculation unit 105 calculates the radiation temperature, which is the temperature radiated from the surroundings, for each personal area detected by the personal area detection unit 104. The radiation temperature is greatly influenced by the body temperatures of the surrounding people when the density of people is high, and is greatly influenced by the floor temperature when the density is low. Therefore, the radiation temperature Tr is calculated using the following formula (1): Tr=F １ (x) x T １ +F ２ (x) x T ２ (1)

[0027] Here, F １ is the view factor between people, F ２ is the shape factor between the person and the floor, T １ is the average temperature of the target personal area, which is one of the personal areas to be calculated, T ２ is the indoor temperature, and x is the density of people around the target individual area. １ is calculated as the average value of the temperatures indicated by all pixels included in the target individual area.

[0028] 6, when the target personal area is personal area 124C, x is determined as the number of people present within predetermined range 125 around personal area 124C. In the example of FIG. 6, four personal areas 124B, 124C, 124D, and 124E exist within predetermined range 125, so the value of x is "4."

[0029] Here, the predetermined range is assumed to be a square centered on the target individual area. The field of view range of the thermal image is assumed to be known, and pixels in the thermal image are assumed to be associated with positions within the field of view in advance. Therefore, once the target individual area is identified, the radiation temperature calculation unit 105 can identify the position of the square centered on the target individual area. Note that, as shown in FIG. 6, the predetermined range 125 is assumed to be a range set on the floor surface within the field of view.

[0030] For example, as shown in FIG. １ and F ２A function, graph, or table showing the relationship between F and x is stored in advance in a storage unit (not shown), and the radiation temperature calculation unit 105 calculates F according to the function, graph, or table. １ and F ２ is determined and the radiation temperature Tr is calculated. １ The larger x is, the larger F ２ The larger x is, the smaller the radiant temperature becomes. The acquired radiant temperature is provided to the thermal sensation predicting unit 107.

[0031] The local humidity estimation unit 106 estimates the local humidity, which is the humidity of each personal area detected by the personal area detection unit 104. The local humidity is the local humidity in the personal area. The local humidity is the humidity actually felt by the subject, who is the person identified in the target personal area, and is the humidity in a local space, which is a predetermined range of space around the subject. The local space is a space smaller than the space in which indoor humidity is measured, and is assumed to be a very narrow space that can contain the subject. Here, the local space is assumed to coincide with the target personal area. The estimated local humidity is provided to the thermal sensation prediction unit 107. Examples of methods for estimating local humidity include the following first estimation method and second estimation method.

[0032] First, a first estimation method for estimating local humidity will be described. In this first estimation method, the local humidity estimation unit 106 estimates local humidity by inputting the indoor humidity, which is the humidity of a predetermined space, and a portion of a thermal image including a personal area for which local humidity is to be calculated, into a pre-trained model. The model here is trained using training data in which the input data is the humidity of a target person's space, which is the space in which the target person exists, and a thermal image including at least the target person in the target person's space, and the correct data is the humidity measured within a predetermined range around the target person that is smaller than the target person's space.

[0033] Specifically, in the first estimation method, the local humidity estimation unit 106 inputs a local image, which is an image of a predetermined range including the target individual area from the thermal image, and the indoor humidity as input data into a pre-trained model, and obtains the estimated local humidity as the output of the model.

[0034] The model here is a model trained using the local image and the indoor humidity as input data and the local humidity as training data. Note that the local image preferably includes, for example, a personal area determined to exist within a predetermined range identified when calculating the density x.

[0035] Next, a second estimation method for estimating local humidity will be described. In the second estimation method, the local humidity estimation unit 106 determines the human body water vapor generation rate, which is the amount of water vapor emitted by an individual corresponding to a personal area, based on the indoor temperature, which is a predetermined temperature of the space, so that the higher the indoor temperature, the greater the human body water vapor generation rate. The local humidity estimation unit 106 calculates the local water vapor generation rate by multiplying the human body water vapor generation rate by a value obtained by adding 1 to the number of people surrounding the individual in the predetermined space. The local humidity estimation unit 106 calculates the local water vapor amount by adding the average water vapor amount determined by the indoor temperature to the local water vapor generation rate. The local humidity estimation unit 106 then estimates the local humidity by dividing the local water vapor amount by the saturated water vapor amount determined by the indoor temperature.

[0036] Here, in the second estimation method, the local humidity estimation unit 106 calculates the local water vapor amount, which is the amount of water vapor in the target individual area, from the number of people around the target individual area, including the individual corresponding to the target individual area, and the human body water vapor generation amount, which is the amount of water vapor emitted by the person, and estimates the local humidity.

[0037] Specifically, the local humidity estimation unit 106 first calculates the local water vapor generation rate, which is the amount of water vapor emitted by the target personal area and people around the target personal area, using the following equation (2): ３ ) = local number of people × human body water vapor generation (g / m ３ ) (2)

[0038] Here, the local number of people may be the same as or different from the density x. Also, a function, graph, or table showing the relationship between the amount of human body water vapor generated and the indoor temperature, such as that shown in Fig. 8, is assumed to be stored in advance in a storage unit (not shown), and the local humidity estimation unit 106 determines the amount of human body water vapor generated from the indoor temperature according to the function, graph, or table. As shown in Fig. 8, the amount of human body water vapor generated increases as the indoor temperature increases.

[0039] Next, the local humidity estimation unit 106 calculates the average water vapor amount using the following formula (3): ３ ) = saturated water vapor amount (g / m ３ ) x indoor humidity (%) ÷ 100 (3)

[0040] The amount of saturated water vapor increases as the temperature increases, as shown in Figure 9. It is assumed that a function, graph, or table showing the relationship shown in Figure 9 is stored in advance in a storage unit (not shown). The temperature here is the indoor temperature.

[0041] Next, the local humidity estimation unit 106 calculates the local water vapor amount, which is the amount of local water vapor in the target personal area and around the target personal area, using the following equation (4): ３ ) = local water vapor generation rate (g / m ３ ) + average water vapor content (g / m ３ ) (4)

[0042] Then, the local humidity estimation unit calculates the local humidity using the following equation (5): Local humidity (%) = Local water vapor amount (g / m ３ ) ÷ saturated water vapor amount (g / m ３ ) x 100 (5)

[0043] The thermal sensation prediction unit 107 predicts the thermal sensation of the corresponding individual using the corresponding local humidity for each personal area. Here, the thermal sensation prediction unit 107 calculates a thermal sensation index value (PMV) for each personal area. Typically, the thermal sensation index value is calculated using a known formula using the indoor temperature, radiation temperature, humidity, air volume, amount of clothing, and metabolic rate. Here, the thermal sensation prediction unit 107 uses the indoor temperature acquired by the indoor temperature acquisition unit 102 and the radiation temperature calculated by the radiation temperature calculation unit 105 as the indoor temperature and radiation temperature for calculating the thermal sensation index value. Then, the thermal sensation prediction unit 107 uses the local humidity estimated by the local humidity estimation unit 106 as the humidity for calculating the thermal sensation index value.

[0044] The thermal sensation prediction unit 107 uses a predetermined value or a value corresponding to the air volume set in an air conditioner (not shown) installed in the space where people are present as the air volume for calculating the thermal sensation index value. Furthermore, a clothing amount value is predetermined for each season, and the thermal sensation prediction unit 107 uses a value corresponding to the season for which the thermal sensation index value is calculated as the clothing amount for calculating the thermal sensation index value. Furthermore, since crowds usually do not move, the thermal sensation prediction unit 107 uses a predetermined value corresponding to standing still as the metabolic rate for calculating the thermal sensation index value.

[0045] As shown in FIG. 10A , some or all of the above-described thermal image acquisition unit 101, indoor temperature acquisition unit 102, indoor humidity acquisition unit 103, personal area detection unit 104, radiation temperature calculation unit 105, local humidity estimation unit 106, and thermal sensation prediction unit 107 can be configured by a memory 10 and a processor 11 such as a CPU (Central Processing Unit) that executes a program stored in the memory 10. In other words, the thermal sensation prediction device 100 can be realized by a computer such as a PC (Personal Computer). Such a program may be provided via a network or may be recorded on a recording medium. That is, such a program may be provided as, for example, a computer program product.

[0046] 10B , a processing circuit 12 such as a single circuit, a composite circuit, a program-driven processor, a program-driven parallel processor, an ASIC (Application Specific Integrated Circuit), or an FPGA (Field Programmable Gate Array) can be used to configure part or all of the thermal image acquisition unit 101, the indoor temperature acquisition unit 102, the indoor humidity acquisition unit 103, the personal area detection unit 104, the radiation temperature calculation unit 105, the local humidity estimation unit 106, and the thermal sensation prediction unit 107. As described above, the thermal image acquisition unit 101, the indoor temperature acquisition unit 102, the indoor humidity acquisition unit 103, the personal area detection unit 104, the radiation temperature calculation unit 105, the local humidity estimation unit 106, and the thermal sensation prediction unit 107 can be realized by a processing circuit network.

[0047] 11 is a flowchart showing the operation of the thermal sensation prediction device 100 according to Embodiment 1. Here, the description will be given assuming that the thermal image acquisition unit 101 has already acquired a thermal image, the indoor temperature acquisition unit 102 has already acquired the indoor temperature, and the indoor humidity acquisition unit 103 has already acquired the indoor humidity.

[0048] First, the personal area detection unit 104 detects the personal areas of each person present in the space from the thermal image (S10).

[0049] The radiation temperature calculation unit 105 calculates the radiation temperature for each personal area detected in step S10 (S11).

[0050] The local humidity estimation unit 106 estimates the local humidity for each personal area detected in step S10 (S12).

[0051] The thermal sensation predicting unit 107 acquires the indoor temperature, the air volume, the metabolic rate, and the amount of clothing in order to calculate the thermal sensation index value (S13).

[0052] Then, the thermal sensation prediction unit 107 calculates a thermal sensation index value for each personal area using the radiation temperature calculated in step S11, the local humidity estimated in step S12, and the indoor temperature, air volume, metabolic rate, and amount of clothing acquired in step S13 (S14).

[0053] As described above, according to the first embodiment, the thermal sensation of people in a crowd can be accurately predicted by taking into consideration the increase in humidity due to sweat and exhalation of surrounding people. Furthermore, by using the thermal sensation accurately predicted in this way to control, for example, an air conditioner, the air conditioner can be operated efficiently while keeping the crowd comfortable.

[0054] 12 is a block diagram showing a schematic configuration of a thermal sensation prediction device 200 according to embodiment 2. The thermal sensation prediction device 200 includes a thermal image acquisition unit 101, an indoor temperature acquisition unit 102, an indoor humidity acquisition unit 103, a personal area detection unit 104, a radiation temperature calculation unit 105, a local humidity estimation unit 206, a thermal sensation prediction unit 107, and a personal water vapor generation amount estimation unit 208.

[0055] The thermal image acquisition unit 101, the indoor temperature acquisition unit 102, the indoor humidity acquisition unit 103, the personal area detection unit 104, the radiation temperature calculation unit 105 and the thermal sensation prediction unit 107 of the thermal sensation prediction device 200 of embodiment 2 are similar to the thermal image acquisition unit 101, the indoor temperature acquisition unit 102, the indoor humidity acquisition unit 103, the personal area detection unit 104, the radiation temperature calculation unit 105 and the thermal sensation prediction unit 107 of the thermal sensation prediction device 100 of embodiment 1.

[0056] The personal water vapor generation rate estimation unit 208 estimates the personal water vapor generation rate, which is the rate of water vapor generated by the body of an individual included in the thermal image. For example, the personal water vapor generation rate estimation unit 208 estimates the personal water vapor generation rate so that the larger the individual's physique, the higher the personal water vapor generation rate. Here, the personal water vapor generation rate estimation unit 208 determines that the larger the human region in the thermal image, the larger the physique of the individual. For example, the personal water vapor generation rate estimation unit 208 identifies the physique of the individual included in the thermal image according to the total number of pixels in the human body region identified by the personal region detection unit 104, and estimates the personal water vapor generation rate so that the larger the physique of the identified individual, the higher the personal water vapor generation rate.

[0057] The personal water vapor generation rate estimation unit 208 compares the total number of pixels in the human body region identified by the personal region detection unit 104 with a first threshold and a second threshold to classify the physique of the individual included in the thermal image into "small physique," "medium physique," and "large physique." Here, it is assumed that the first threshold is less than the second threshold. Specifically, if the total number of pixels in the human body region is equal to or less than the first threshold, the personal water vapor generation rate estimation unit 208 classifies the physique of the individual included in the thermal image into "small physique." Furthermore, if the first threshold is less than the total number of pixels in the human body region and less than the second threshold, the personal water vapor generation rate estimation unit 208 classifies the physique of the individual included in the thermal image into "medium physique." Furthermore, if the second threshold is equal to or less than the total number of pixels in the human body region, the personal water vapor generation rate estimation unit 208 classifies the physique of the individual included in the thermal image into "large physique."

[0058] 13, a function, graph, or table showing the relationship between the personal water vapor generation rate for each physique and the indoor temperature is stored in advance in a storage unit (not shown), and the personal water vapor generation rate estimation unit 208 determines the personal water vapor generation rate according to the physique classified from the indoor temperature in accordance with the function, graph, or table. Note that, as shown in FIG. 13, the personal water vapor generation rate for each physique increases as the indoor temperature increases.

[0059] The local humidity estimation unit 206 estimates the local humidity in each personal area detected by the personal area detection unit 104. In the second embodiment, the local humidity estimation unit 206 estimates the local humidity using the second estimation method described above. However, in the second embodiment, the local humidity estimation unit 206 calculates the amount of local water vapor generation using the following equation (6) instead of equation (2). The local water vapor generation amount (g / m ３ ) = local number of people × personal water vapor generation (g / m ３ ) (6)

[0060] The personal water vapor generation rate estimator 208 described above can also be configured, for example, as shown in Fig. 10(A), with a memory 10 and a processor 11 such as a CPU that executes a program stored in the memory 10. The personal water vapor generation rate estimator 208 can also be configured, for example, with a processing circuit 12 as shown in Fig. 10(B). As described above, the personal water vapor generation rate estimator 208 can also be realized by a processing circuit network.

[0061] Fig. 14 is a flowchart showing the operation of the thermal sensation prediction device 200 according to embodiment 2. Here, the description will be made assuming that the thermal image acquisition unit 101 has already acquired a thermal image, the indoor temperature acquisition unit 102 has already acquired the indoor temperature, and the indoor humidity acquisition unit 103 has already acquired the indoor humidity. Furthermore, among the steps included in the flowchart shown in Fig. 14, steps that perform the same processing as the steps included in the flowchart shown in Fig. 11 are assigned the same reference numerals as those shown in Fig. 11.

[0062] The processes of steps S10 and S11 in Fig. 14 are the same as the processes of steps S10 and S11 in Fig. 11. However, after step S11 in Fig. 14, the process proceeds to step S20.

[0063] In step S20, the personal water vapor generation rate estimation unit 208 estimates the personal water vapor generation rate, which is the rate of water vapor generated by the human body of an individual included in the thermal image.

[0064] Next, the local humidity estimation unit 206 estimates the local humidity for each personal area detected in step S10 (S21). Here, the local humidity estimation unit 206 estimates the local humidity using the personal water vapor generation rate estimated in step S20 by the second estimation method described above. Then, the process proceeds to step S13.

[0065] The processes in steps S13 and S14 in FIG. 14 are similar to the processes in steps S13 and S14 in FIG.

[0066] First, the personal area detection unit 104 detects the personal areas of each person present in the space from the thermal image (S10).

[0067] The radiation temperature calculation unit 105 calculates the radiation temperature for each personal area detected in step S10 (S11).

[0068] The local humidity estimation unit 106 estimates the local humidity for each personal area detected in step S10 (S12).

[0069] The thermal sensation predicting unit 107 acquires the indoor temperature, the air volume, the metabolic rate, and the amount of clothing in order to calculate the thermal sensation index value (S13).

[0070] Then, the thermal sensation prediction unit 107 calculates a thermal sensation index value for each personal area using the radiation temperature calculated in step S11, the local humidity estimated in step S12, and the indoor temperature, air volume, metabolic rate, and amount of clothing acquired in step S13 (S14).

[0071] As described above, according to the second embodiment, the amount of water vapor emitted by each person in the vicinity can be accurately estimated based on the person's physique. Therefore, the local humidity at the location of the subject can also be accurately estimated. As a result, the thermal sensation of the subject can be more accurately estimated.

[0072] 100, 200 Thermal sensation prediction device, 101 Thermal image acquisition unit, 102 Indoor temperature acquisition unit, 103 Indoor humidity acquisition unit, 104 Personal area detection unit, 105 Radiant temperature calculation unit, 106, 206 Local humidity estimation unit, 107 Thermal sensation prediction unit, 208 Personal water vapor generation amount estimation unit.

Claims

1. A thermal sensation prediction device comprising: a personal area detection unit that detects a personal area that is an area of ​​an individual present in a predetermined space and that is smaller than the predetermined space from a thermal image showing the temperature distribution in the predetermined space; a local humidity estimation unit that estimates the local humidity that is the humidity in the personal area; and a thermal sensation prediction unit that predicts the thermal sensation of the individual using the local humidity.

2. The thermal sensation prediction device of claim 1, wherein the local humidity estimation unit receives input data from the humidity of a subject space in which a subject exists and a thermal image including at least the subject in the subject space, and estimates the local humidity by inputting the humidity of the predetermined space and a portion of the thermal image including the individual that includes at least the individual area into a model trained using training data in which the correct answer data is humidity measured within a predetermined range around the subject that is smaller than the subject space.

3. The thermal sensation prediction device according to claim 1, characterized in that the local humidity estimation unit determines a human body water vapor generation rate, which is the amount of water vapor emitted by the individual, based on the temperature of the predetermined space, such that the higher the temperature of the predetermined space, the greater the amount of water vapor emitted by the individual; calculates the local water vapor generation rate by multiplying the human body water vapor generation rate by a value obtained by adding 1 to the number of people around the individual in the predetermined space; calculates the local water vapor amount by adding the local water vapor generation rate to an average water vapor amount determined by the temperature of the predetermined space; and estimates the local humidity by dividing the local water vapor amount by the saturated water vapor amount determined by the temperature of the predetermined space.

4. The thermal sensation prediction device according to claim 3, further comprising a personal water vapor generation amount estimation unit that estimates the amount of water vapor generated by the human body so that the larger the physique of the individual, the greater the amount of water vapor generated by the human body.

5. The thermal sensation prediction device according to claim 4, wherein the personal water vapor generation rate estimation unit determines that the larger the area of ​​the person in the thermal image, the larger the physique of the individual.

6. A program causing a computer to function as: a personal area detection unit that detects a personal area that is an area of ​​an individual present in a predetermined space and that is smaller than the predetermined space from a thermal image showing the distribution of temperature in the predetermined space; a local humidity estimation unit that estimates the local humidity that is the humidity in the personal area; and a thermal sensation prediction unit that predicts the thermal sensation of the individual using the local humidity.

7. A method for predicting thermal sensation, comprising: detecting a personal area, which is an area of ​​an individual present in a predetermined space and is smaller than the predetermined space, from a thermal image showing the distribution of temperature in the predetermined space; estimating local humidity, which is the humidity in the personal area; and predicting the thermal sensation of the individual using the local humidity.

Images