Control method, program, and control system

JP7591744B2Active Publication Date: 2024-11-29PANASONIC INTELLECTUAL PROPERTY MANAGEMENT CO LTD
View PDF 6 Cites 0 Cited by

Patent Information

Application Number
JP2023545485
Authority / Receiving Office
JP · JP
Patent Type
Patents
Current Assignee / Owner
Priority Date
2021-08-31
Filing Date
2022-08-24
Publication Date
2024-11-29
Estimated Expiration
2042-08-24

AI Technical Summary

Benefits of technology

【0009】 本発明によれば、人の温冷感を推定して快適な環境温度を実現することができる空気調和機の制御方法及びプログラムが実現される。

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 0007591744000001
    Figure 0007591744000001
  • Figure 0007591744000002
    Figure 0007591744000002
  • Figure 0007591744000003
    Figure 0007591744000003
Patent Text Reader

Abstract

A computer for, inter alia, a control system (10) for executing a control method performs the following: acquiring (S11) a thermal image (100) showing a temperature distribution of a to-be-air conditioned space detected by a sensor (20) installed on the ceiling of the to-be-air conditioned space; detecting (S12) the outline of a person (1) present in the to-be-air conditioned space on the basis of the thermal image (100); identifying (S13) a human body region(100a), which is a region surrounded by the outline of the person (1), in the thermal image (100); calculating (S14) a human body temperature including the temperature of clothing (1b) of the person (1) on the basis of the temperature distribution inside the human body region (100a); calculating (S15) an ambient temperature of a region (peripheral region (100b)) other than the human body region (100a) in the thermal image (100); calculating (S16) a heat dissipation amount of the person (1) on the basis of a differential value between the human body temperature and the ambient temperature; estimating (S17) a warm / cold sensation of the person (1) on the basis of the heat dissipation amount; and controlling (S18) an air conditioner (30) on the basis of the warm / cold sensation.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical field]

[0001] The present invention relates to a control method, a program, and a control system for an air conditioner. [Background technology]

[0002] A technology is known that measures the temperature difference between the nose of an occupant and the temperature around the cheeks and / or forehead, measures the thermal state of the occupant, such as the occupant's sensation and comfort (hereinafter referred to as thermal sensation), based on the temperature difference, and controls an air conditioning device installed in the vehicle so that the occupant feels comfortable (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] JP 2017-197195 A Summary of the Invention [Problem to be solved by the invention]

[0004] However, the technology described in Patent Document 1 estimates a person's thermal sensation only from the temperature around the person's face, and it is difficult to say that it is able to estimate a person's thermal sensation and achieve a comfortable environmental temperature.

[0005] Therefore, the present invention provides an air conditioner control method, program, and control system that estimates a person's thermal sensation and achieves a comfortable environmental temperature. [Means for solving the problem]

[0006] A control method according to one aspect of the present invention is a control method for an air conditioner executed by a computer, in which the computer acquires a thermal image showing the temperature distribution in the air-conditioned space detected by a sensor installed on the ceiling of the air-conditioned space, detects the outline of a person present in the air-conditioned space based on the acquired thermal image, identifies a human body area in the thermal image that is the area surrounded by the detected outline of the person, calculates a human body temperature including the temperature of the person's clothing based on the temperature distribution within the identified human body area, calculates the ambient temperature of areas other than the human body area in the thermal image, calculates the amount of heat dissipation of the person based on the difference value between the human body temperature and the ambient temperature, estimates the person's thermal sensation based on the calculated amount of heat dissipation, and controls the air conditioner based on the estimated thermal sensation.

[0007] A program according to one aspect of the present invention is a program for causing a computer to execute the control method.

[0008] A control system according to one embodiment of the present invention comprises an acquisition unit that acquires a thermal image indicating the temperature distribution in the air-conditioned space detected by a sensor installed on the ceiling of the air-conditioned space, a detection unit that detects the outline of a person present in the air-conditioned space based on the acquired thermal image, an identification unit that identifies a human body area in the thermal image, which is the area surrounded by the detected outline of the person, a calculation unit that calculates a human body temperature including the temperature of the person's clothing based on the temperature distribution in the identified human body area, calculates the ambient temperature of areas other than the human body area in the thermal image, and calculates the amount of heat dissipation of the person based on the difference value between the human body temperature and the ambient temperature, an estimation unit that estimates the thermal sensation of the person based on the calculated amount of heat dissipation, and a control unit that controls an air conditioner installed in the air-conditioned space based on the estimated thermal sensation. Effect of the Invention

[0009] According to the present invention, there is provided a method and program for controlling an air conditioner that can estimate a person's thermal sensation and achieve a comfortable environmental temperature. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram illustrating an example of a functional configuration of a control system according to an embodiment. [Diagram 2] FIG. 2 is a diagram showing an air-conditioned space to which the operating system according to the embodiment is applied. [Diagram 3] FIG. 3 is a diagram for explaining a thermal image. [Figure 4] FIG. 4 is a flowchart of an operation example of the control system according to the embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, the embodiments will be described with reference to the drawings. Note that the embodiments described below are all comprehensive or specific examples. The numerical values, shapes, materials, components, component arrangement and connection forms, steps, and order of steps shown in the following embodiments are merely examples and are not intended to limit the present invention. Furthermore, among the components in the following embodiments, components that are not described in the independent claims will be described as optional components.

[0012] In addition, each drawing is a schematic diagram and is not necessarily a precise illustration. In each drawing, the same reference numerals are used for substantially the same configurations, and duplicated explanations may be omitted or simplified.

[0013] (Embodiment) [composition] First, the configuration of a control system according to an embodiment will be described. Fig. 1 is a block diagram showing an example of a functional configuration of a control system according to an embodiment. Fig. 2 is a diagram showing an air-conditioned space to which a control system according to an embodiment is applied.

[0014] The control system 10 extracts the contour of the person 1 present in the air-conditioned space based on a thermal image showing the temperature distribution in the air-conditioned space, identifies a human body region surrounded by the extracted contour of the person 1, calculates a difference value between the temperature in the identified human body region and other regions, and calculates the amount of heat dissipation of the person 1 based on the calculated difference value. The control system 10 then estimates the thermal sensation of the person 1 based on the calculated amount of heat dissipation of the person 1, and controls the air conditioner 30 based on the estimated thermal sensation. The air-conditioned space is, for example, an indoor space in a building. The building may be, for example, an office, a commercial facility, a public facility, an educational facility, or a house. As shown in FIG. 1, the control system 10 includes a sensor 20, an air conditioner 30, and a server device 40. The control system 10 may include a plurality of sensors 20.

[0015] [Sensor] The sensor 20 is installed, for example, on the ceiling of the space to be air conditioned, and generates a thermal image showing the temperature distribution when the space to be air conditioned is viewed from the ceiling side (in other words, from above).

[0016] 2, the sensor 20 may be directly installed on the ceiling of the space to be air-conditioned, or may be detachably connected to a power supply terminal of a lighting device (not shown) or a fire alarm (not shown) installed on the ceiling. In the latter case, the sensor 20 operates by receiving power from the lighting device or the fire alarm. The power supply terminal is, for example, a Universal Serial Bus (USB) terminal.

[0017] The sensor 20 is, for example, an infrared sensor. The infrared sensor is, for example, an infrared array sensor (in other words, a thermal image sensor) configured by an array of 8×8 infrared detection elements. In other words, a thermal image generated by the infrared sensor has 8×8 pixels. The thermal image shows the temperature distribution in the sensing range of the infrared sensor with a resolution of 8×8. FIG. 3 is a diagram for explaining a thermal image. FIG. 3(a) is a diagram that shows a schematic image of the air-conditioned space (more specifically, the sensing range of the sensor 20 in the air-conditioned space) detected by the sensor 20, taken by a camera. FIG. 3(b) is a diagram that shows a schematic thermal image detected by the sensor 20. Each of the 8×8 small areas in FIG. 3(a) represents a pixel included in the image, and each of the 8×8 small areas in FIG. 3(b) represents a pixel included in the thermal image 100. The numerical value in the pixel shown in FIG. 3(b) is a pixel value, specifically indicating a temperature. 3(b) is a human body region 100a, and the region other than the human body region 100a in the thermal image 100 is a surrounding region 100b. In the following embodiment, for the sake of simplicity, it is assumed that pixel value=temperature value.

[0018] The sensor 20 may be installed on a wall or the like to generate a thermal image showing the temperature distribution when the air conditioned space is viewed from the wall side (in other words, from the side). In this case, the control system 10 may obtain multiple thermal images showing the temperature distribution when the air conditioned space is viewed from above and from the side.

[0019] [Air conditioner] The air conditioner 30 controls the temperature (also called the environmental temperature) around the person 1 in the air conditioned space based on the thermal sensation of the person 1 in the air conditioned space. As shown in Fig. 1, the air conditioner 30 includes, for example, a communication unit 31, a control unit 32, a storage unit 33, a louver 36, a compressor 37, and a fan 38.

[0020] The communication unit 31 is a communication circuit (communication module) for the air conditioner 30 to communicate with the sensor 20 and the server device 40. The communication unit 31 may include a communication circuit (communication module) for communicating via the wide area communication network 5, and a communication circuit (communication module) for communicating via a local communication network (not shown). The communication unit 31 is, for example, a wireless communication circuit for performing wireless communication, but may also be a wired communication circuit for performing wired communication. There is no particular limitation on the communication standard for the communication performed by the communication unit 31.

[0021] The control unit 32 performs various information processing based on the thermal images acquired from the sensor 20, and controls the operation of the air conditioner 30. Specifically, the control unit 32 controls the louvers 36, the compressor 37, and the fan 38 based on the thermal sensation of the person 1 estimated by the estimation unit 32e. For example, when it is estimated that the person 1 feels hot, the control unit 32 performs control to orient the louvers 36 toward the person 1 and operate the compressor 37 and the fan 38 to generate cool air.

[0022] Specifically, the control unit 32 includes an acquisition unit 32a that acquires a thermal image received by the communication unit 31, a detection unit 32b that detects the contour of the person 1 including the skin 1a and clothing 1b of the person 1 in the thermal image, an identification unit 32c that identifies a human body region 100a (see FIG. 3B) that is an area surrounded by the contour of the person 1, a calculation unit 32d that calculates the amount of heat dissipation of the person 1 based on a difference value between the human body temperature and the ambient temperature, an estimation unit 32e that estimates the thermal sensation of the person 1 based on the amount of heat dissipation, and an output unit 32f that outputs the estimation result of the thermal sensation of the person 1 or a control condition determined based on the estimation result. The functions of the acquisition unit 32a, the detection unit 32b, the identification unit 32c, the calculation unit 32d, the estimation unit 32e, and the output unit 32f are realized by a processor or a microcomputer constituting the control unit 32 executing a computer program stored in the storage unit 33. The functions of the acquisition unit 32a, the detection unit 32b, the identification unit 32c, the calculation unit 32d, the estimation unit 32e, and the output unit 32f will be described in detail later in an operation example.

[0023] The storage unit 33 is a storage device that stores dedicated application programs and the like to be executed by the control unit 32. The storage unit 33 may store a trained machine learning model (hereinafter also referred to as a trained model) and a database 35. The machine learning model is trained by a learning unit 44 of the server device 40. The control unit 32 stores the trained machine learning model transmitted from the server device 40 in the storage unit 33, thereby updating the trained machine learning model in the storage unit 33. The storage unit 33 is realized by, for example, a semiconductor memory.

[0024] The machine learning model may have a convolutional layer, and may be, for example, a convolutional neural network (CNN), but is not limited thereto. For example, as shown in FIG. 1, the machine learning model may be composed of a first machine learning model (for example, a trained model 34a) and a second machine learning model (for example, a trained model 34b). In this case, the first machine learning model (for example, the trained model 34a) may input a thermal image showing the temperature distribution of the air-conditioned space and output a super-resolution thermal image, and the second machine learning model (for example, the trained model 34b) may input the super-resolution thermal image output from the first machine learning model (for example, the trained model 34a) and output the outline of a person present in the air-conditioned space. In the example of FIG. 1, the two trained models 34a and 34b are stored in the storage unit 33, but one trained model in which the trained models 34a and 34b are linked may be stored in the storage unit 33.

[0025] The database 35 may store, for example, thermal sensations corresponding to the amount of heat radiation from a person during cooling operation and heating operation, and the control conditions of the air conditioner 30 corresponding to the thermal sensations, in association with each other.

[0026] [Server device] The server device 40 learns the machine learning model and updates the learned models 34a and 34b. In the example of Fig. 1, the server device 40 is a cloud computer installed outside a building having a space to be air-conditioned, but may be an edge computer installed inside the building. The server device 40 includes a communication unit 41, a control unit 42, a storage unit 43, and a learning unit 44.

[0027] The communication unit 41 is a communication module (communication circuit) for the server device 40 to communicate with the air conditioner 30. The communication unit 41 transmits, for example, trained models 45a and 45b to the air conditioner 30. In the example of FIG. 1, the communication unit 41 is a communication circuit (communication module) for communicating via the wide area communication network 5, but may also include a communication circuit (communication module) for communicating via a local communication network (not shown). The communication performed by the communication unit 41 may be wireless communication or wired communication. There is no particular limitation on the communication standard used for the communication.

[0028] The control unit 42 performs various information processes in the server device 40. The control unit 42 is specifically realized by a processor or a microcomputer.

[0029] The storage unit 43 is a storage device that stores the machine learning models (trained models 45a, 45b) trained by the learning unit 44, teacher data 46 used for learning, and computer programs executed by the control unit 42. Specifically, the storage unit 43 is realized by a semiconductor memory, a hard disk drive (HDD), or the like.

[0030] The learning unit 44 performs machine learning using the teacher data 46. The learning unit 44 uses machine learning to generate a machine learning model (so-called trained model) that receives a thermal image showing the temperature distribution in the air-conditioned space as an input and outputs the outline of a person present in the air-conditioned space. As described above, the machine learning model may be, for example, a CNN, and may be composed of a first machine learning model and a second machine learning model. In this case, the learning unit 44 may use machine learning to generate a first machine learning model (for example, trained model 45a) that receives a thermal image showing the temperature distribution in the air-conditioned space and outputs a super-resolution thermal image, and a second machine learning model (for example, trained model 45b) that receives the super-resolution thermal image and outputs the outline of a person present in the air-conditioned space. The trained models 45a and 45b include trained parameters adjusted by machine learning. The generated trained models 45a and 45b are stored in the memory unit 43, for example, and are transmitted to the air conditioner 30 via the communication unit 41. As a result, the air conditioner 30 updates the trained model 34a in the storage unit 33 using the acquired trained model 45a, and updates the trained model 34b using the trained model 45b. The learning unit 44 is realized by, for example, a processor executing a program stored in the storage unit 43.

[0031] The teacher data 46 has a thermal image showing the temperature distribution of the air-conditioned space as input data, and a contour of a person present in the air-conditioned space as output data. Specifically, the teacher data 46 may be a data set including a pair of a thermal image as input data and a contour of a person as output data, or may include a plurality of data sets (e.g., first teacher data and second teacher data). For example, the first teacher data is teacher data used for learning a first machine learning model (e.g., trained model 45a), and is a data set including, for example, a pair of input data of the teacher data 46 (i.e., a thermal image) and a super-resolution thermal image obtained by super-resolution of the thermal image as output data. For example, the second teacher data is teacher data used for learning a second machine learning model (e.g., trained model 45b), and is a data set including a pair of a super-resolution thermal image as input data and an output data of the teacher data 46 (i.e., a contour of a person) as output data.

[0032] 1 has been described as an example in which a trained machine learning model is generated by the server device 40 and the generated trained machine learning model is transmitted to the air conditioner 30 to update the machine learning model, but this is not limiting. For example, the air conditioner 30 may have the configuration of the server device 40. Also, for example, the server device 40 may acquire a thermal image detected by the sensor 20, estimate the thermal sensation of the person 1, and output the estimated result of the thermal sensation of the person 1 to the control of the air conditioner 30.

[0033] [Example] Next, a description will be given of an example of the operation of the control system 10. FIG.

[0034] The control unit 32 of the air conditioner 30 acquires a thermal image 100 (e.g., FIG. 3(b)) indicating the temperature distribution in the air-conditioned space (more specifically, the space within the sensing range of the sensor 20 in the air-conditioned space) detected by the sensor 20 (S11). More specifically, the acquisition unit 32a of the air conditioner 30 acquires the thermal image 100 from the sensor 20 via the communication unit 31. At this time, the acquisition unit 32a may store the acquired thermal image 100 in the memory unit 33.

[0035] Next, the detection unit 32b detects the contour of the person 1 present in the air-conditioned space based on the thermal image 100 acquired by the acquisition unit 32a in step S11 (S12). For example, the detection unit 32b divides the thermal image 100 into two or more temperature distribution regions, each of which has the same characteristics and is made up of a plurality of adjacent pixels, and detects the contour of the person 1 including the clothing 1b (see FIG. 2) of the person 1 based on the two or more divided temperature distribution regions. For example, the detection unit 32b may perform segmentation on the thermal image 100 using a machine learning model to divide the thermal image 100 into a region in which the person 1 is captured and a region in which the person 1 is not captured. As described above, the machine learning model may have a convolutional layer, and may be, for example, a convolutional neural network (CNN), but is not limited thereto. Note that the machine learning model (for example, the trained model 34b in FIG. 1) may be applied to a super-resolution thermal image obtained by super-resolution-enhancing the thermal image 100.

[0036] The detection unit 32b may generate a super-resolution thermal image having a higher resolution than the thermal image 100 by interpolating at least one pixel between adjacent pixels for a plurality of pixels in the thermal image 100, and detect the contour of the person 1 based on the generated super-resolution thermal image. For example, the detection unit 32b may super-resolution the thermal image 100 by using a method in which, in generating the super-resolution thermal image, an average value of pixel values ​​(i.e., temperature values) of adjacent pixels in the thermal image 100 is calculated, and a new pixel having a pixel value equivalent to the average value is inserted between adjacent pixels. In addition, for example, the detection unit 32b may super-resolution the thermal image 100 using a trained machine learning model (for example, the trained model 34a in FIG. 1). The trained model 34a may be, for example, a Generative Adversarial Network for Super-Resolurion (SRGAN), but is not limited thereto.

[0037] In step S12, the detection unit 32b may detect the outline of the person 1 from the thermal image 100, for example, using a trained machine learning model stored in the memory unit 33 (for example, the trained models 34a and 34b in FIG. 1).

[0038] Next, the identifier 32c identifies, within the thermal image 100, a human body region 100a (see FIG. 3(b)), which is a region surrounded by the outline of the person 1 detected by the detector 32b in step S12 (S13).

[0039] Next, the calculation unit 32d calculates the human body temperature including the temperature of the clothing 1b of the person 1 based on the temperature distribution in the human body region 100a identified by the identification unit 32c in step S13 (S14). Specifically, in step S14, the calculation unit 32d calculates the average value of the surface temperature of the skin 1a (see FIG. 2) of the person 1 and the surface temperature of the clothing 1b of the person 1 as the human body temperature. For example, the calculation unit 32d may calculate the average value of the pixel values ​​(i.e., temperature values) of the human body region 100a as the human body temperature.

[0040] Next, the calculation unit 32d calculates the ambient temperature of the person 1 based on the temperature distribution in an area other than the human body area 100a (the so-called surrounding area 100b) (S15). For example, in step S15, the calculation unit 32d may calculate the average value of the pixel values ​​(i.e., the temperature value) of the surrounding area 100b as the ambient temperature. Note that steps S14 and S15 may be performed in parallel.

[0041] Next, the calculation unit 32d calculates the amount of heat dissipation from the person 1 based on the difference between the human body temperature and the ambient temperature (S16). Specifically, the calculation unit 32d calculates the difference between the human body temperature and the ambient temperature calculated in steps S14 and S15, and calculates the amount of heat dissipation from the person 1 based on the calculated difference.

[0042] Next, the estimation unit 32e estimates the thermal sensation of the person 1 based on the amount of heat radiation of the person 1 calculated by the calculation unit 32d in step S16 (S17). For example, the estimation unit 32e estimates that the person 1 feels cold when the amount of heat radiation is greater than the threshold, and estimates that the person 1 feels hot when the amount of heat radiation is smaller than the threshold. The threshold may include a threshold for cooling operation and a threshold for heating operation. The thermal sensation of the person 1 corresponds to the amount of heat radiation of the person 1, but the thermal sensation of the person 1 may differ depending on the season even if the amount of heat radiation is the same. Therefore, by estimating the thermal sensation of the person 1 using the threshold for cooling operation and the threshold for heating operation, the thermal sensation of the person 1 can be more accurately estimated.

[0043] Next, the control unit 32 controls the air conditioner 30 based on the thermal sensation of the person 1 estimated by the estimation unit 32e in step S17 (S18). For example, when the estimation unit 32e estimates that the person 1 feels cold in step S17, the control unit 32 controls the air conditioner 30 to increase the ambient temperature, and when the estimation unit 32e estimates that the person 1 feels hot, the control unit 32 controls the air conditioner 30 to decrease the ambient temperature. The control unit 32 may determine control conditions for cooling or heating operation corresponding to the thermal sensation of the person 1 based on the database 35, and control the operations of the louvers 36, the compressor 37, and the fan 38 based on the determined control conditions. The database 35 has been described above, so a description thereof will be omitted here.

[0044] At this time, the control unit 32 may control the air conditioner 30 based on the position of the human body region 100a and the thermal sensation of the person 1. Specifically, the calculation unit 32d calculates the position of the person 1 in the air-conditioned space based on the position of the human body region in the thermal image. For example, the storage unit 33 may store table information indicating the correspondence between the positions of pixels in the thermal image and coordinates in the air-conditioned space, and the calculation unit 32d may use the table information to calculate the position (specifically, coordinates) of the person 1 in the air-conditioned space from the position (pixel position) where the person 1 is detected in the thermal image 100.

[0045] In addition, when the thermal image shows the temperature distribution in the air-conditioned space when viewed from above (i.e., the ceiling side) as in this embodiment, the coordinates of person 1 calculated by calculation unit 32d are two-dimensional coordinates when the air-conditioned space is viewed from above.

[0046] As described above, the control system 10 can estimate the thermal sensation of the person 1 present in the air-conditioned space and control the air conditioner 30 based on the estimated thermal sensation. Therefore, the control system 10 can estimate the thermal sensation of the person 1 and achieve a comfortable environmental temperature.

[0047] [Variations] The control unit 32 controls the air conditioner 30 based on the estimation result (the thermal sensation of the person 1) estimated by the estimation unit 32e, and may output the estimation result. The output unit 32f may output the result estimated by the estimation unit 32e to an information terminal (not shown) carried by the person 1. This allows the person 1 to check the estimation result displayed on the display unit of the information terminal, so that if the thermal sensation actually felt by the person 1 (hereinafter referred to as the actual thermal sensation) differs from the estimation result, the person 1 can input the actual thermal sensation into the information terminal to correct the estimation result. Both the estimation result and the corrected estimation result are stored in the memory unit 33.

[0048] The detection unit 32b may perform information processing on the thermal image based on a rule-based algorithm that does not use a machine learning model. For example, the detection unit 32b may perform processing to detect pixels whose pixel values ​​are maximum values ​​among a plurality of pixels included in the thermal image. Here, the pixel whose pixel value is the maximum value means a pixel whose pixel value is the maximum value in a two-dimensional arrangement of pixels. In other words, the pixel whose pixel value is the maximum value means a pixel whose pixel value is higher than that of surrounding pixels when pixel values ​​at the same time are compared in a two-dimensional arrangement of pixels. There may be a plurality of pixels whose pixel values ​​are maximum values ​​in one thermal image.

[0049] Here, assuming that there is no heat source with a higher temperature than a person in the air-conditioned space, even if the pixel value is a maximum value, if the pixel value is smaller than a predetermined value (i.e., the temperature is low), it is considered that no person is present at the position corresponding to the pixel. Therefore, when the detection unit 32b detects a pixel whose pixel value is a maximum value and whose pixel value is equal to or greater than a predetermined value (e.g., 30°C or more), it may infer that a person 1 is present in the air-conditioned space, and detect the outline of the person 1 based on the detected pixels.

[0050] In addition, the calculation unit 32d can calculate the position (i.e., coordinates) of person 1 in the air-conditioned space by applying the above-mentioned table information to the position of a pixel whose pixel value is a maximum value and whose pixel value is equal to or greater than a predetermined value (e.g., 30°C or greater).

[0051] Another example of information processing based on a rule-based algorithm is processing to detect changes over time in the pixel value (temperature) of each of a plurality of pixels included in a thermal image. Assuming that there is no heat source other than a person in the air-conditioned space, and if there is no person in the air-conditioned space, the pixel value (temperature) of each of a plurality of pixels included in the thermal image changes slowly over time. In this state, when a person 1 enters the air-conditioned space, the pixel value of the pixel in the portion of the thermal image in which the person 1 is reflected changes (increases) suddenly. By detecting changes over time in the pixel value of each of a plurality of pixels, the detection unit 32b may, for example, estimate that a person 1 is present in the air-conditioned space when the pixel value increases suddenly, and perform processing to detect the outline of the person 1.

[0052] [Effects, etc.] As described above, the control method is a method for controlling the air conditioner 30 executed by a computer such as the control system 10, in which the computer acquires a thermal image 100 showing the temperature distribution in the air-conditioned space detected by the sensor 20 installed on the ceiling of the air-conditioned space (step S11 in FIG. 4), detects the contour of a person 1 present in the air-conditioned space based on the acquired thermal image 100 (step S12), and identifies a human body region 100a, which is an area surrounded by the contour of the detected person 1, within the thermal image 100 ( Step S13), calculate the human body temperature including the temperature of the clothing 1b of the person 1 based on the temperature distribution in the identified human body region 100a (Step S14), calculate the ambient temperature of the region other than the human body region 100a (surrounding region 100b) in the thermal image 100 (Step S15), calculate the amount of heat dissipation of the person 1 based on the difference between the human body temperature and the ambient temperature (Step S16), estimate the thermal sensation of the person 1 based on the calculated amount of heat dissipation (Step S17), and control the air conditioner 30 based on the estimated thermal sensation (Step S18).

[0053] This control method can realize an environmental temperature that the person 1 feels comfortable at by estimating the thermal sensation of the person 1 present in the air-conditioned space.

[0054] Also, for example, when the amount of heat dissipation is greater than a threshold, the computer infers that person 1 feels cold and controls the air conditioner 30 to raise the ambient temperature based on the estimated thermal sensation, and when the amount of heat dissipation is less than the threshold, the computer infers that person 1 feels hot and controls the air conditioner 30 to lower the ambient temperature based on the estimated thermal sensation.

[0055] Such a control method can estimate the thermal sensation of the person 1 from the amount of heat radiation of the person 1 based on a threshold value.

[0056] Furthermore, for example, in detecting the contour of person 1, the computer divides the thermal image 100 into two or more temperature distribution regions (e.g., a human body region 100a and a surrounding region 100b) that have the same characteristics in at least a portion of the thermal image 100 and that are each made up of a plurality of adjacent pixels, and detects the contour of person 1 including the clothing 1b of person 1 based on the two or more divided temperature distribution regions.

[0057] Such a control method can divide a thermal image into a region corresponding to the person 1 (human body region 100a) and a region corresponding to other than the person 1 (surrounding region 100b) by dividing a plurality of adjacent pixels having the same characteristics in at least a part of the thermal image into one temperature distribution region. This makes it possible for the control method to detect the outline of the person 1 based on the pixel values ​​of the plurality of pixels included in the region corresponding to the person 1.

[0058] Also, for example, the computer generates a super-resolution thermal image having a higher resolution than the thermal image 100 by interpolating at least one pixel between adjacent pixels for multiple pixels in the thermal image 100, and detects the contour of person 1 based on the super-resolution thermal image.

[0059] Such a control method detects the contour of the person 1 with a finer resolution than the thermal image 100, and therefore the contour of the person 1 can be detected with higher accuracy.

[0060] Furthermore, for example, the computer detects the outline of the person 1 from the thermal image 100 using a trained machine learning model (for example, trained models 34a and 34b).

[0061] Such a control method can quickly and accurately detect the contour of the person 1 from the thermal image 100 by using a machine learning model.

[0062] Furthermore, for example, the above machine learning model is trained using training data 46, and the training data 46 has a thermal image as input data and a human outline as output data.

[0063] Such a control method can detect the outline of the person 1 by inputting the thermal image 100 into a trained machine learning model.

[0064] Also, for example, the above machine learning model is composed of a first machine learning model (e.g., trained model 34a) and a second machine learning model (e.g., trained model 34b), and the first machine learning model is trained using first teacher data including a pair of input data of the teacher data 46 as input data and a super-resolution thermal image of the input data as output data, and the second machine learning model is trained using second teacher data including a pair of the super-resolution thermal image as input data and the output data of the teacher data 46 as output data.

[0065] Such a control method can detect the contour of the person 1 from the thermal image 100 using a first machine learning model (e.g., trained model 34a) that receives the thermal image 100 as input and outputs a super-resolution thermal image of the thermal image 100, and a second machine learning model (e.g., trained model 34b) that receives the super-resolution thermal image as input and outputs the contour of the person. Therefore, the control method can detect the contour of the person 1 from a super-resolution thermal image that has a finer resolution than the thermal image 100, and therefore can detect the contour of the person 1 with higher accuracy.

[0066] The program is a program for causing the computer to execute any one of the above control methods.

[0067] Such a program can realize an environmental temperature that the person 1 feels comfortable at by estimating the thermal sensation of the person 1 present in the air-conditioned space.

[0068] The control system 10 also includes an acquisition unit 32a that acquires a thermal image 100 indicating the temperature distribution in the air-conditioned space detected by a sensor 20 installed on the ceiling of the air-conditioned space, a detection unit 32b that detects the outline of a person 1 present in the air-conditioned space based on the acquired thermal image 100, an identification unit 32c that identifies a human body region 100a, which is an area surrounded by the outline of the detected person 1, in the thermal image 100, a calculation unit 32d that calculates a human body temperature including the temperature of clothing 1b of the person 1 based on the temperature distribution in the identified human body region 100a, calculates the ambient temperature of an area other than the human body region 100a in the thermal image 100 (so-called surrounding region 100b), and calculates the amount of heat dissipation of the person 1 based on the difference value between the human body temperature and the ambient temperature, an estimation unit 32e that estimates the thermal sensation of the person 1 based on the calculated amount of heat dissipation, and a control unit 32 that controls the air conditioner 30 installed in the air-conditioned space based on the estimated thermal sensation.

[0069] Such a control system 10 can realize an environmental temperature that the person 1 feels comfortable at by estimating the thermal sensation of the person 1 present in the air-conditioned space.

[0070] (Other embodiments) Although the control system and the control method according to the embodiment have been described above, the present invention is not limited to the above embodiment.

[0071] In the above embodiment, the control system is realized by a plurality of devices, but may be realized as a single device. For example, the control system may be realized as a single device corresponding to a server device. When the control system is realized by a plurality of devices, each component of the control system may be allocated to the plurality of devices in any manner.

[0072] In the above embodiment, the processes executed by a specific processing unit may be executed by another processing unit. The order of multiple processes may be changed, or multiple processes may be executed in parallel.

[0073] In the above embodiment, each component may be realized by executing a software program suitable for each component. Each component may be realized by a program execution unit such as a CPU or a processor reading and executing a software program recorded on a recording medium such as a hard disk or a semiconductor memory.

[0074] Furthermore, each component may be realized by hardware. For example, each component may be a circuit (or an integrated circuit). These circuits may form a single circuit as a whole, or each may be a separate circuit. Furthermore, each of these circuits may be a general-purpose circuit, or a dedicated circuit.

[0075] In addition, the general or specific aspects of the present invention may be realized as a system, an apparatus, a method, an integrated circuit, a computer program, or a recording medium such as a computer-readable CD-ROM. Also, the present invention may be realized as any combination of a system, an apparatus, a method, an integrated circuit, a computer program, and a recording medium. For example, the present invention may be realized as a control method executed by a computer such as a control system. Also, the present invention may be realized as a program for causing a computer to execute the control method, or as a non-transitory computer-readable recording medium on which such a program is stored.

[0076] In addition, the present invention also includes forms obtained by applying various modifications to each embodiment that a person skilled in the art may think of, or forms realized by arbitrarily combining the components and functions of each embodiment without departing from the spirit of the present invention.

[0077] (Additional Note) Hereinafter, examples of inventions that can be obtained from the disclosure of this specification will be given, and the effects and the like that can be obtained from the inventions will be described.

[0078] [Invention 1] A method for controlling an air conditioner executed by a computer, comprising: The computer includes: A thermal image showing a temperature distribution in the space to be air-conditioned detected by a sensor installed on a ceiling of the space to be air-conditioned is acquired; Detecting an outline of a person present in the air-conditioned space based on the acquired thermal image; Identifying a human body region within the thermal image, the human body region being an area surrounded by an outline of the detected human; Calculating a human body temperature including the temperature of clothing of the person based on a temperature distribution in the identified human body region; Calculating an ambient temperature of an area other than the human body area in the thermal image; Calculating the amount of heat dissipation from the person based on a difference between the human body temperature and the ambient temperature; Estimating a thermal sensation of the person based on the calculated amount of heat loss; controlling the air conditioner based on the estimated thermal sensation; Control methods.

[0079] [Effects of Invention 1] This control method can realize an environmental temperature that people feel comfortable at by estimating the thermal sensation of people in the air-conditioned space.

[0080] [Invention 2] The computer includes: When the amount of heat radiation is greater than a threshold, it is estimated that the person feels cold, and the air conditioner is controlled to increase the ambient temperature based on the estimated thermal sensation; When the amount of heat radiation is smaller than the threshold, it is estimated that the person feels hot, and the air conditioner is controlled to lower the ambient temperature based on the estimated thermal sensation. A control method according to claim 1.

[0081] [Effects of Invention 2] Such a control method can estimate a person's thermal sensation from the amount of heat radiation from the person based on a threshold value.

[0082] [Invention 3] The computer, in detecting the contour of the person, Dividing the thermal image into two or more temperature distribution regions, each of which has the same characteristic and is adjacent to the other pixel, as one temperature distribution region; detecting an outline of the person including clothing of the person based on the two or more divided temperature distribution regions; A control method according to claim 1 or 2.

[0083] [Effects of Invention 3] Such a control method can divide a thermal image into an area corresponding to a person (a so-called human body area) and an area corresponding to other than the person 1 (a so-called surrounding area) by dividing a plurality of adjacent pixels having the same characteristics in at least a part of the thermal image into one temperature distribution area. This makes it possible for the control method to detect the outline of a person based on the pixel values ​​of a plurality of pixels included in the area corresponding to a person.

[0084] [Invention 4] The computer includes: generating a super-resolution thermal image having a higher resolution than the thermal image by interpolating at least one pixel between adjacent pixels for a plurality of pixels in the thermal image; Detecting an outline of the person based on the super-resolution thermal image. The control method according to any one of Inventions 1 to 3.

[0085] [Effects of Invention 4] This control method detects the contours of a person with a finer resolution than a thermal image, and therefore can detect the contours of a person with greater accuracy.

[0086] [Invention 5] The computer detects the outline of the person from the thermal image using a trained machine learning model. The control method according to any one of Inventions 1 to 4.

[0087] [Effects of Invention 5] This type of control method uses a machine learning model to quickly and accurately detect human contours from thermal images.

[0088] [Invention 6] The machine learning model is trained using training data, The teacher data has the thermal image as input data and the outline of the person as output data. A control method according to claim 5.

[0089] [Effects of Invention 6] This control method can detect the contours of a person by inputting a thermal image into a trained machine learning model.

[0090] [Invention 7] The machine learning model is composed of a first machine learning model and a second machine learning model, The first machine learning model is trained using first teacher data including a pair of the input data of the teacher data and a super-resolution thermal image of the input data; The second machine learning model is trained using second teacher data including a pair of the super-resolution thermal image and the output data of the teacher data. A control method according to claim 6.

[0091] [Effects of Invention 7] Such a control method can detect the contour of a person from a thermal image using a first machine learning model that takes a thermal image as an input and outputs a super-resolution thermal image of the thermal image, and a second machine learning model that takes the super-resolution thermal image as an input and outputs a contour of a person. Therefore, the control method can detect the contour of a person from a super-resolution thermal image, which has a finer resolution than a thermal image, and therefore can detect the contour of a person with higher accuracy.

[0092] [Invention 8] A method for causing a computer to execute the control method according to any one of claims 1 to 7. program.

[0093] [Effects of Invention 8] Such a program can realize an environmental temperature that people feel comfortable at by estimating the thermal sensation of people in the air-conditioned space.

[0094] [Invention 9] an acquisition unit that acquires a thermal image showing a temperature distribution in the space to be air-conditioned, the thermal image being detected by a sensor installed on a ceiling of the space to be air-conditioned; a detection unit that detects an outline of a person present in the air conditioned space based on the acquired thermal image; an identification unit that identifies a human body region, which is a region surrounded by an outline of the detected person, in the thermal image; a calculation unit that calculates a human body temperature including the temperature of the person's clothing based on the temperature distribution in the identified human body region, calculates an ambient temperature of an area other than the human body region in the thermal image, and calculates an amount of heat dissipation from the person based on a difference value between the human body temperature and the ambient temperature; an estimation unit that estimates a thermal sensation of the person based on the calculated heat dissipation amount; a control unit that controls an air conditioner installed in the air conditioned space based on the estimated thermal sensation; Equipped with Control system.

[0095] [Effects of Invention 9] Such a control system can realize an environmental temperature that people feel comfortable at by estimating the thermal sensation of people in the air-conditioned space. [Explanation of symbols]

[0096] 1 person 1b Clothes 10. Control System 20 Sensors 30 Air conditioner 32 Control section 32a Acquisition part 32b Detection section 32c Specific part 32d Calculation section 32e Estimation part 34a Pre-trained model 34b Pre-trained model 45a Pre-trained model 45b Pre-trained model 46 Teacher Data 100 Thermal Images 100a human body area 100b Surrounding area

Claims

1. A method for controlling an air conditioner executed by a computer, comprising: The computer includes: A thermal image showing a temperature distribution in the space to be air-conditioned detected by a sensor installed on a ceiling of the space to be air-conditioned is acquired; Detecting an outline of a person present in the air-conditioned space based on the acquired thermal image; Identifying a human body region within the thermal image, the human body region being an area surrounded by an outline of the detected human; Calculating a human body temperature including the temperature of clothing of the person based on a temperature distribution in the identified human body region; Calculating an ambient temperature of an area other than the human body area in the thermal image; Calculating the amount of heat dissipation from the person based on a difference between the human body temperature and the ambient temperature; Estimating a thermal sensation of the person based on the calculated amount of heat loss; Controlling the air conditioner based on the estimated thermal sensation; Detecting the outline of the person from the thermal image using a trained machine learning model; The machine learning model is trained using training data, The teacher data includes the thermal image as input data and the outline of the person as output data, The machine learning model is composed of a first machine learning model and a second machine learning model, The first machine learning model is trained using first teacher data including a pair of the input data of the teacher data and a super-resolution thermal image of the input data; The second machine learning model is trained using second teacher data including a pair of the super-resolution thermal image and the output data of the teacher data. Control methods.

2. The computer includes: When the amount of heat radiation is greater than a threshold, it is estimated that the person feels cold, and the air conditioner is controlled to increase the ambient temperature based on the estimated thermal sensation; When the amount of heat radiation is smaller than the threshold, it is estimated that the person feels hot, and the air conditioner is controlled to lower the ambient temperature based on the estimated thermal sensation. The control method according to claim 1 .

3. The computer, in detecting the contour of the person, Dividing the thermal image into two or more temperature distribution regions, each of which has the same characteristics and is adjacent to the other pixels in at least a portion of the thermal image, as one temperature distribution region; detecting an outline of the person including clothing of the person based on the two or more divided temperature distribution regions; The control method according to claim 1 or 2.

4. The computer includes: generating a super-resolution thermal image having a higher resolution than the thermal image by interpolating at least one pixel between adjacent pixels for a plurality of pixels in the thermal image; Detecting an outline of the person based on the super-resolution thermal image. The control method according to claim 1 or 2.

5. A method for causing a computer to execute the control method according to claim 1 or 2. program.

6. an acquisition unit that acquires a thermal image showing a temperature distribution in the space to be air-conditioned, the thermal image being detected by a sensor installed on a ceiling of the space to be air-conditioned; a detection unit that detects an outline of a person present in the air conditioned space based on the acquired thermal image; an identification unit that identifies a human body region, which is a region surrounded by an outline of the detected person, in the thermal image; a calculation unit that calculates a human body temperature including the temperature of the person's clothing based on the temperature distribution in the identified human body region, calculates an ambient temperature of an area other than the human body region in the thermal image, and calculates an amount of heat dissipation from the person based on a difference value between the human body temperature and the ambient temperature; an estimation unit that estimates a thermal sensation of the person based on the calculated heat dissipation amount; a control unit that controls an air conditioner installed in the air conditioned space based on the estimated thermal sensation; Equipped with The detection unit detects an outline of the person from the thermal image using a trained machine learning model, The machine learning model is trained using training data, The teacher data includes the thermal image as input data and the outline of the person as output data, The machine learning model is composed of a first machine learning model and a second machine learning model, The first machine learning model is trained using first teacher data including a pair of the input data of the teacher data and a super-resolution thermal image of the input data; The second machine learning model is trained using second teacher data including a pair of the super-resolution thermal image and the output data of the teacher data. Control system.

Citation Information

Patent Citations

  • Air conditioning device

    JP2013124833A

  • Air conditioner

    JP2015052435A

  • Detection method, detection device and control method

    JP2016208408A

  • Sheet adjustment method

    JP2017197195A

  • Air conditioner

    JP2017203576A