TWI769791B - Air-conditioning control system based on human comfort - Google Patents

Abstract

An air-conditioning control system based on human comfort is applied to an air conditioner in a space. The air-conditioning control system comprises an environmental sensing unit, a camera device, and a computing device. The environmental sensing unit is to sense the environmental condition of the space and outputs sensing data accordingly. The camera device is to capture and output a real-time image of the space, wherein the real-time image includes person(s) in the image. The computing device stores a posture recognition model, a clothing recognition model, and a program package for comfort estimation. The real-time image is transmitted to the posture recognition model and the clothing recognition model. Based on recognition results from the models and the sensing data, the program package generates a comfort reference value. The computing device controls the operation of the air conditioner according to the comfort reference value.

Description

Air Conditioning Control System Based on Personnel Comfort

The present invention relates to an air-conditioning control system, in particular to an air-conditioning control system based on personnel comfort.

A conventional air conditioner (such as an air conditioner) installed in an indoor space has a temperature detector, the temperature detector is used to detect an indoor temperature, and a user can operate a remote control to output a temperature setting value to the air conditioner equipment, the controller of the air-conditioning equipment can determine whether the indoor temperature is greater than the temperature set value. When the controller determines that the indoor temperature is greater than the temperature set value, the controller can increase the operating power of the compressor motor of the air conditioner to enhance the cooling effect; on the contrary, when the controller determines that the indoor temperature is lower than For the temperature setting value, the controller can reduce the operating power of the compressor motor of the air conditioner to save electricity.

It can be seen that the conventional air conditioner only unilaterally decides whether to strengthen the cooling effect according to the level of the indoor temperature. However, the indoor temperature cannot effectively reflect the body temperature of the indoor occupants, resulting in the conventional air conditioner in enhancing or weakening the cooling effect. The timing of the effect is often unable to satisfy the indoor personnel, so that the indoor personnel feel too cold or the cooling effect is insufficient, causing troubles in use.

In view of this, the main purpose of the present invention is to provide an air-conditioning control system based on the comfort of the personnel, in order to overcome the shortcomings of conventional air-conditioning equipment that often cannot satisfy indoor personnel when the cooling effect is enhanced or weakened.

The air-conditioning control system based on personnel comfort of the present invention supplies at least one air-conditioning device in a space, and the air-conditioning control system includes: an environment sensing unit, detecting the environment of the space and correspondingly outputting an environment sensing data; a camera device for capturing and outputting a real-time image of the space, the real-time image including at least one portrait; and a computing device, the signal is connected to the environment sensing unit, the camera device and the at least one air conditioner, the computing device stores a human body posture recognition model, a clothing recognition model and a human body comfort program package; the computing device stores the real-time The image is input to the human body posture recognition model and the clothing recognition model, and according to the recognition information and the environment sensing data, the calculation of the human comfort level program suite is performed to obtain a comfort level evaluation value, and the computing device is based on the comfort level evaluation value. The value controls the operating state of the at least one air conditioner.

To sum up, the present invention implements the calculation of the human comfort level program kit according to the identification information of the body posture recognition model and the clothing recognition model in addition to the environmental sensing data, so as to obtain the comfort level evaluation value. The evaluation value of the degree of comfort is related to the comfort degree that the person feels at the moment, and it can be seen that the evaluation value of the degree of comfort is one of the basis for controlling the operation state of the air conditioner in the present invention. Compared with the conventional air-conditioning equipment, the present invention does not unilaterally decide whether to strengthen the cooling effect according to the indoor temperature, but comprehensively considers personnel posture, clothing and environmental sensing data to control the at least one air conditioner. The operating state of the equipment, under this comprehensive consideration, can relatively make the personnel feel more comfortable.

The air-conditioning control system of the present invention is supplied for at least one air-conditioning device in a space, which can be an indoor space, such as a classroom, an office, a conference room, an auditorium, a room, a living room, etc., but not limited to the foregoing examples. The number of the at least one air conditioner may be one or more, and the number and installation location may be determined by the size of the space or the number of squares. The at least one air conditioner may be an air conditioner, a heater, or an air conditioner or heater. In the embodiment of the present invention, the at least one air conditioner is described by taking an air conditioner as an example.

Please refer to FIG. 1 , an embodiment of the air conditioning control system of the present invention includes an environment sensing unit 10 , a camera device 20 and a computing device 30 , the configuration of which can be referred to as an example in FIG. 2 , the space S is provided with a podium 40 , a seat The chair 41 and an air conditioner 42, the air outlet of the air conditioner 42 faces the interior of the space S.

The environment sensing unit 10 detects the environment of the space S and outputs an environment sensing data 100 correspondingly. The environment sensing unit 10 may include a dry bulb thermometer 11 , a black bulb thermometer 12 , a relative humidity meter 13 and an air At least one of the flow meters 14, the dry bulb thermometer 11, the black bulb thermometer 12, the relative humidity meter 13 and the air flow meter 14 can be electronic sensing devices, which can be arranged at appropriate positions in the space as required (eg ceiling, wall or floor). Alternatively, the dry bulb thermometer 11 , the black bulb thermometer 12 and the relative humidity meter 13 may be an integrated measuring instrument.

The dry bulb thermometer 11 is used to output a dry bulb temperature signal tdb, which reflects the temperature of the space S sensed by the dry bulb thermometer 11, and its unit can be Celsius (°C) or Fahrenheit (°F); the black bulb thermometer 12 is used for outputting a dry bulb temperature signal tdb. Output a black bulb temperature signal tr, which reflects the temperature of the space S sensed by the black bulb thermometer 12, and its unit can be Celsius (°C) or Fahrenheit (°F); the relative hygrometer 13 outputs a relative humidity signal rh, which reflects the relative humidity of the space S sensed by the relative hygrometer 13, which can be a percentage (%) value; the air velocity meter 14 is for outputting an air velocity signal vr, which reflects the sensed by the air velocity meter 14. The measured air circulation state of the space can be measured in meters per second (m/s). For example, the air velocity meter 14 can be installed at doors and windows. Correspondingly, the environmental sensing data 100 output by the environmental sensing unit 10 may include at least one of the dry bulb temperature signal tdb, the black bulb temperature signal tr, the relative humidity signal rh and the air velocity signal vr.

The camera device 20 is disposed in the space S and shoots toward the space S, the camera device 20 outputs a real-time image 200 of the space S, wherein, when there is at least one person in the space S and the at least one person enters the space S For the shooting range of the camera device 20, the real-time image 200 may include at least one portrait corresponding to the at least one person. The camera device 20 can be, for example, a digital camera, a web camera, a motion camera, or a smartphone or tablet computer with a shooting function.

The computing device 30 is signal-connected to the environment sensing unit 10 , the camera device 20 and the at least one air conditioner 42 for mutual signal and data transmission, so the computing device 30 can receive real-time reception from the environment sensing unit 10 For the environmental sensing data 100, the computing device 30 can receive the real-time image 200 from the camera device 20, and the computing device 30 can output a control signal CS to the at least one air conditioner 42 to control the operation of the at least one air conditioner 42 The state, for example, starts the air conditioner 42 and discharges cool air, or stops the operation and does not discharge cool air.

The signal connection means of the computing device 30 to the environment sensing unit 10 , the camera device 20 and the at least one air conditioner 42 can be wired connection means or wireless connection means. Taking wired connection means as an example, the computing device 30 can be connected to the camera device 20 through a physical transmission line, and the physical signal line can be, for example, a high-definition multimedia interface (HDMI) transmission line or a universal serial bus (USB) transmission line; the The computing device 30 can be connected to the electronic sensing device in the environment sensing unit 10 through the physical signal line; the computing device 30 can be connected to the control contacts/pins of the control circuit boards of the air conditioners 42 through the physical signal line . On the other hand, the wireless connection means can use the short-range wireless communication technology of the Internet of Things. For example, the computing device 30 can be connected to the camera device 20 , the air conditioners 42 and/or through the WiFi communication technology or the Bluetooth communication technology. Or the electronic sensing device in the environmental sensing unit 10, and through the wireless connection means, the engineering of physical wiring in the space S can be avoided. In the example shown in FIG. 2, the space S is provided with a WiFi device 50 for implementing the wireless connection of each device.

The computing device 30 implements the computation of a human body posture recognition model M(met), a clothing recognition model M(col) and a human body comfort program package P(comfort), and controls the operation of the air conditioner according to the computation results, The human body posture recognition model M(met) and the clothing recognition model M(col) are image recognition models. Please refer to FIG. 3 . The computing device 30 inputs the real-time image 200 into the human body posture recognition model M(met). ) to obtain a human body posture information X1, that is, the human body posture information X1 is the output information of the human body posture recognition model M(met), and input the real-time image 200 into the clothing recognition model M(col) to obtain a clothing The information X2, that is, the clothing information X2 is the output information of the clothing recognition model M(col).

The computing device 30 can be a desktop computer, a notebook computer or a cloud computing device. The computing device 30 includes a storage unit 31 for storing the human body posture recognition model M(met), the clothing recognition model Model M(col), the human comfort program package P(comfort), and program data related to controlling the air conditioning equipment. The storage unit 31 can be, for example, a conventional hard disk (HDD), a solid-state disk (SSD), a memory card or a memory. The computing device 30 includes a central processing unit (CPU) or a graphics processing unit (GPU), so the program data can be executed by the central processing unit (CPU) or the graphics processing unit (GPU) to implement the human posture The identification model M(met), the clothing identification model M(col) and the human comfort level program package P(comfort) are calculated, and the air conditioning equipment 42 can be implemented according to the execution result of the human body comfort level program package P(comfort) operation control.

The aforementioned program data can be created in the Python programming language, but is not limited to the Python programming language, that is, the human posture recognition model M(met), the clothing recognition model M(col), and the human comfort program suite The program data of P(comfort) can use the existing technology. For example, the program data of the human body posture recognition model M(met) and the clothing recognition model M(col) can be downloaded from the Github website platform, wherein the human body posture recognition model M( met) can use OpenPose open source code (reference URL: https://github.com/CMU-Perceptual-Computing-Lab/openpose). The clothing recognition model M(col) can use OpenCV open source code (Reference URL: https://github.com/EdjeElectronics/TensorFlow-Object-Detection-API-Tutorial-Train-Multiple-Objects-Windows-10). The program data of the human comfort program package P(comfort) can use the pythermalcomfort public program package, and the version used in the present invention is 1.3.6 (refer to the website: https://pypi.org/project/pythermalcomfort/). Therefore, the human posture recognition model M(met), the clothing recognition model M(col) and the human comfort program package P(comfort) are publicly available information, and the program structure and operation thereof will not be described in detail here. principle.

After downloading and storing the human body posture recognition model M(met) and the clothing recognition model M(col) in the storage unit 31, each model can be trained, and examples of training image samples can be referred to FIGS. 4A to 4E, wherein, FIG. 4A is a training picture sample 61 in which a person presents a “standing posture” and wears “long clothes and trousers”, FIG. 4B is a training picture sample 62 in which a person presents a “standing posture” and wears “long clothes and shorts”, and FIG. 4C is a training picture sample 62 in which the personnel presents “ A training image sample 63 of standing posture and wearing “shorts and trousers”, FIG. 4D is a training picture sample 64 of a person showing a “standing posture” and wearing “shorts and shorts”, and FIG. Shorts and shorts” training image sample 65. It should be noted that there are various forms of training image samples, and FIGS. 4A to 4E are examples for reference only, not limited thereto. For example, the personnel in other training image samples may show squatting posture, walking and running. , or wearing long skirts, short skirts...etc. Therefore, through the training of the model, the recognition accuracy of the human body posture recognition model M(met) and the clothing recognition model M(col) can be improved. In the actual application of the present invention, the human body posture recognition model M(met) and the clothing recognition model M(col) are models that have completed training respectively.

In the above, the human body posture information X1 can reflect the posture of the person in the space, such as sitting posture, standing posture, walking posture or running posture, etc.; the clothing information X2 can reflect the clothing of the personnel in the space, such as long (sleeves) Clothes, short (sleeve) clothes, trousers (skirt) or shorts (skirt)...etc. It can be seen that the human body temperature is closely related to the human body posture information X1 and the clothing information X2. For example, a person running is warmer than a person sitting, or a person wearing a short-sleeved shirt is colder than a person wearing long-sleeved clothing. . As shown in FIG. 6, a comfort evaluation value Y is calculated according to the human body posture information X1 and the clothing information X2. Therefore, the computing device 30 of the present invention controls the at least one comfort level according to the comfort evaluation value Y. The operating state of the air conditioner 42 so as to make the person feel comfortable will be explained in detail as follows.

The human body posture information X1 includes at least one posture recognition result Xa and at least one personnel quantity value Xb, and a posture recognition result Xa corresponds to a personnel quantity value Xb; the clothing information X2 includes at least one clothing recognition result Xc and at least one clothing quantity value Xd, a piece of clothing identification result Xc corresponds to a piece of clothing quantity value Xd. For example, FIG. 5 is one of the real-time images 200 captured by the computing device 30 from the camera 20 (FIG. 5 only shows a schematic diagram of a portrait, and the background scene may be omitted). The human gesture recognition model M(met) is based on the The at least one gesture recognition result Xa identified by the real-time image 200 includes "standing posture", and the at least one person number value Xb includes "2", that is, it is recognized that there are two standing persons in the real-time image 200 . On the other hand, the at least one clothing recognition result Xc recognized by the clothing recognition model M(col) according to the real-time image 200 includes "long clothing", "short clothing", "long pants" and "shorts", corresponding to , the at least one clothing quantity value Xd includes "1", "1", "1" and "1", that is, it is recognized that the real-time image 200 contains a long coat, a short coat, a trousers and A pair of shorts. As mentioned above, the identification results of Figure 5 are arranged in the following table: Attitude recognition results Personnel value Clothing recognition results Clothing quantity value Live image standing 2 long coat 1 short coat 1 trousers 1 shorts 1

Please refer to the table below. The storage unit 31 can store a metabolic comparison table Table_met, and the metabolic comparison table Table_met includes a plurality of posture states and a metabolic reference value corresponding to each posture state.

posture state metabolic reference value sitting position 1 standing 1.2 walk 2.0

Please refer to the table below, the storage unit 31 can store a clothing comparison table Table_col, and the clothing comparison table Table_col includes a plurality of clothing states and a clothing reference value corresponding to each clothing state.

clothing status clothing reference value shorts 0.06 trousers 0.24 short coat 0.17 long coat 0.34

It should be noted that the posture that the human body posture recognition model M(met) can recognize and the content of the metabolism comparison table Table_met are not limited to “standing posture”, “sitting posture”, and “walking”. The clothing recognition model M(col ) identifiable clothing and the content of the clothing comparison table Table_col is not limited to "long clothing", "short clothing", "long pants" and "shorts", the present invention only provides examples for illustration.

Specifically, the content of the metabolism comparison table Table_met and the clothing comparison table Table_col can be based on existing public information, with reference to the American Society of Heating, Refrigerating, and Air Condition Engineers (ASHRAE) Standard No. 55-2010 "ANSI/ASHRAE Standard 55-2010: Thermal Environmental Conditions for Human Occupancy" (as attached), available at http://arco-hvac.ir/wp-content/uploads/2015/11/ASHRAE-55-2010. pdf. Among them, Table A1 (TABLE A1: Metabolic Rates for Typical Tasks) in ANSI/ASHRAE Standard 55-2010 can be used as the attitude state (corresponding to the Activity column of the table A1) and the metabolic reference value ( Reference basis for the Met Units column of this Table A1). Table B2 (TABLE B2: Garment Insulation) in ANSI/ASHRAE Standard 55-2010 can be used as the clothing status (corresponding to the Garment Description column of this table B2) and the clothing reference value (corresponding to the table B2) in the clothing comparison table Table_col. the I _clu ,col field). Therefore, the user can create and store the contents of the metabolism comparison table Table_met and the clothing comparison table Table_col according to ANSI/ASHRAE Standard 55-2010 in advance and store them in the storage unit 31 for reading during operation.

Please refer to FIG. 6 , after obtaining the identification information of the human body posture identification model M(met) and the clothing identification model M(col) respectively, the computing device 30 can compare the metabolism comparison table Table_met by means of table look-up. At least one metabolic reference value Xa_met corresponding to the at least one gesture recognition result Xa is obtained, and at least one clothing reference value Xc_col corresponding to the at least one clothing recognition result Xc is obtained by comparing in the clothing comparison table Table_col.

Therefore, the computing device 30 can perform the calculation of the human comfort program package P(comfort) according to the environmental sensing data 100, the number of persons value Xb, the metabolic reference value Xa_met, the number of clothes value Xd and the clothing reference value Xc_col to obtain the The comfort evaluation value Y, that is, the comfort evaluation value Y is the output information of the human comfort level program package P(comfort). Mean Vote, PMV) program kit, hereinafter referred to as PMV program kit (PMV); an embodiment of the human comfort program kit P (comfort) may be a Standard Effective Temperature (SET) program kit (SET), It is called SET program suite later, and they are explained as follows.

Please refer to FIG. 6 , when the computing device 30 implements the calculation of the PMV program suite (PMV), in addition to the number of personnel value Xb, the reference value of metabolism Xa_met, the value of the number of clothes Xd and the reference value of clothing Xc_col, the dry ball The temperature signal tdb, the black bulb temperature signal tr, the relative humidity signal rh and the air velocity signal vr are provided to the PMV program package (PMV) together, so that the PMV program package (PMV) can be based on the number of people value Xb, metabolism The reference value Xa_met, the clothing quantity value Xd, the clothing reference value Xc_col, the dry bulb temperature signal tdb, the black bulb temperature signal tr, the relative humidity signal rh and the air velocity signal vr calculate and output the comfort evaluation value Y, the PMV program package ( The comfort evaluation value Y output by PMV) is a numerical value.

The storage unit 31 stores an adjustable upper limit value TH_u and/or a lower limit value TH_1, the upper limit value TH_u and the lower limit value can be a value, for example, the upper limit value TH_u can be 0.5, the lower limit value TH_l can be -0.5. After the computing device 30 obtains the comfort level evaluation value Y through the PMV program package (PMV), the computing device 30 determines whether the comfort level evaluation value Y is greater than the upper limit TH_u. When the computing device 30 determines the comfort level If the evaluation value Y is greater than the upper limit value TH_u, that is, Y>TH_u, it means that the person in the space S may feel hot and uncomfortable. The air conditioner 42 operates to discharge cold air; on the other hand, when the computing device 30 determines that the comfort evaluation value Y is less than the lower limit value TH_1, that is, Y<TH_1, it means that the person in the space S may feel cold. The control command CS outputted by the computing device 30 to the air conditioner 42 is a stop command, which stops the operation of the air conditioner 42 to stop discharging cold air.

When the computing device 30 implements the operation of the SET program suite (SET), similar to the PMV program suite (PMV), the SET program suite (SET) can calculate the number of people Xb, the metabolic reference value Xa_met, the number of clothes Xd, The clothing reference value Xc_col, the dry bulb temperature signal tdb, the black bulb temperature signal tr, the relative humidity signal rh and the air velocity signal vr calculate and output the comfort evaluation value Y, the comfort evaluation output by the SET program kit (SET) The value Y is a temperature value, such as a Celsius temperature value (°C) or a Fahrenheit temperature value (°F). The storage unit 31 stores an adjustable temperature setting value Tset, for example, the temperature setting value Tset may be 27 degrees Celsius. After the computing device 30 obtains the comfort level evaluation value Y through the SET program kit (SET), the computing device 30 determines whether the comfort level evaluation value Y is greater than the temperature setting value Tset, when the computing device 30 determines the comfort level If the evaluation value Y is greater than the temperature set value Tset, that is, Y>Tset, it means that the person in the space S may feel warmer. 42 operates to exhaust cold air.

The aforementioned example takes one air conditioner 42 as an example. In other embodiments, when there are a plurality of air conditioners in the space, the computing device can also be connected to the signals of the plurality of air conditioners, and output an activation command to the plurality of air conditioners. A set of air conditioners, so that the plurality of air conditioners can be started or stopped synchronously. In addition to synchronously controlling a plurality of air conditioners, the present invention can also control the operation state of any air conditioner individually, as described below.

Referring to FIG. 7 , the computing device 30 can signally connect a first air conditioner 421 and a second air conditioner 422 . The installation positions of the first air conditioner 421 and the second air conditioner 422 can refer to FIG. 8 . An air conditioner 421 may be disposed in front of the space S, and the second air conditioner 422 may be disposed in the rear of the space S. FIG. 9 is one of the real-time images 200 captured by the computing device 30 from the camera device 20 ( FIG. 9 only shows a schematic diagram of a portrait and an air conditioner, other background scenes may be omitted), the computing device 30 according to a dividing line div The real-time image 200 is divided into a first sub-image 201 and a second sub-image 202, wherein the dividing line div can be defined horizontally at the center of the real-time image 200, but not limited to this, the first sub-image 201 The identification information (ie, at the front of the space S) is used to control the first air conditioner 421 , and the identification information of the second sub-image 202 (ie, at the front of the space S) is used to control the second air conditioner 422 accordingly. By analogy, when the real-time image 200 is divided into a plurality of sub-images, the computing device 30 can respectively control different air conditioners according to the identification information of the plurality of sub-images.

The computing device 30 respectively inputs the first sub-image 201 and the second sub-image 202 to the human body posture recognition model M(met) and the clothing recognition model M(col), and the recognition results are arranged in the following table: Attitude recognition results Personnel value Clothing recognition results Clothing quantity value first subimage standing 1 long coat 1 trousers 1 second sub-image standing 2 shorts 2 trousers 1 shorts 1

The computing device 30 obtains the metabolic reference value Xa_met corresponding to each posture recognition result Xa of the first sub-image 201 and the second sub-image 202 and the clothing reference value Xc_col corresponding to each clothing recognition result Xc by looking up a table, and according to respectively implement the calculation of the human comfort level program package P(comfort) to obtain a first comfort level evaluation value Y1 corresponding to the first sub-image 201 and a second comfort level evaluation value Y2 corresponding to the second sub-image 202 . In this way, the computing device 30 can implement zoned air conditioning control. For example, when the computing device 30 determines that the first comfort level evaluation value Y1 is greater than the upper limit TH_u but the second comfort level evaluation value Y2 is lower than the upper limit value TH_u The upper limit TH_u, that is, Y1>TH_u and Y2<TH_u, means that the person in front of the space S may feel warmer, but the person behind the space S still feels comfortable, and the computing device 30 can only activate the first air conditioner The device 421 discharges cold air, but the second air conditioning device 422 is not activated.

To sum up, the comfort evaluation value Y is one of the basis for controlling the operation state of the air-conditioning equipment 42 in the present invention, wherein the comfort evaluation value Y depends on the current situation of the people in the space S (such as the number of people, clothing and activities) It is related to the comfort that people feel at the moment; for example, when there are many people gathered in the indoor environment, more people wear more clothes or more people are active, the comfort evaluation value Y is higher, Otherwise lower. Therefore, as the computing device 30 determines that the comfort evaluation value Y is greater than the upper limit value TH_u or lower than the lower limit value TH_1, it can reflect the discomfort that the person feels too hot or too cold, and then activate or deactivate it in a timely manner The air-conditioning equipment 42 enables personnel to be in a more comfortable environment.

10: Environmental Sensing Unit 100: Environmental sensing data 11: Dry bulb thermometer 12: Black Ball Thermometer 13: Relative Humidity Meter 14: Air flow meter 20: Camera device 200: Live image 201: The first sub-image 202: Second sub-image 30: Computing device 31: Storage unit 40: Podium 41: Seat 42: Air conditioning equipment 421: First Air Conditioning Equipment 422: Second Air Conditioning Equipment 50: WiFi device 61, 62, 63, 64, 65: training image samples S: space tdb: dry bulb temperature signal tr: black ball temperature signal rh: relative humidity signal vr: Air velocity signal CS: control signal M(met): Human pose recognition model M(col): clothing recognition model P(comfort): Human Comfort Program Package X1: Human body posture information X2: Clothing Information Y: comfort evaluation value Y1: The first comfort evaluation value Y2: Second comfort evaluation value Table_met: Metabolic comparison table Table_col: clothing comparison table Xa: state identification result Xa_met: Metabolic reference value Xb: the value of the number of personnel Xc: Clothing recognition result Xc_col: clothing reference value Xd: clothing quantity value PMV: PMV Program Suite SET:SET program suite TH_u: upper limit value TH_l: lower limit Tset: temperature set point div: dividing line

FIG. 1 is a block schematic diagram of an embodiment of an air conditioning control system based on personnel comfort according to the present invention. Figure 2: A schematic diagram of a usage scenario of an embodiment of the present invention. FIG. 3 is a schematic diagram of the computing device in the present invention inputting real-time images to the human body posture recognition model and the clothing recognition model and obtaining the recognition information. FIG. 4A is a schematic diagram (1) of an example of a training image sample of the human body posture recognition model and the clothing recognition model in the present invention. FIG. 4B is a schematic diagram (2) of an example of a training image sample of the human body posture recognition model and the clothing recognition model in the present invention. FIG. 4C is a schematic diagram (3) of an example of training image samples of the human body posture recognition model and the clothing recognition model in the present invention. FIG. 4D is a schematic diagram (4) of an example of training image samples of the human body posture recognition model and the clothing recognition model in the present invention. FIG. 4E is a schematic diagram (5) of an example of a training image sample of the human body posture recognition model and the clothing recognition model in the present invention. FIG. 5 is a schematic diagram of the real-time image received by the computing device from the camera device in the present invention. FIG. 6 is a schematic diagram of the implementation of the table look-up method and the human body comfort program kit implemented by the computing device in the present invention. FIG. 7 is a schematic block diagram of another embodiment of the air conditioning control system based on the comfort of the personnel according to the present invention. FIG. 8 is a schematic diagram of a usage scenario of another embodiment of the present invention. FIG. 9 is a schematic diagram of the real-time image received by the computing device from the camera device in the present invention.

10: Environmental Sensing Unit

100: Environmental sensing data

11: Dry bulb thermometer

12: Black Ball Thermometer

13: Relative Humidity Meter

14: Air flow meter

20: Camera device

200: Live image

30: Computing device

31: Storage unit

42: Air conditioning equipment

tdb: dry bulb temperature signal

tr: black ball temperature signal

rh: relative humidity signal

vr: Air velocity signal

CS: control signal

M(met): Human pose recognition model

M(col): clothing recognition model

P(comfort): Human Comfort Program Package

Table_met: Metabolic comparison table

Table_col: clothing comparison table

TH_u: upper limit value

TH_l: lower limit

Tset: temperature set point

Claims (9)

An air-conditioning control system based on personnel comfort, which is supplied to at least one air-conditioning device in a space, the air-conditioning control system includes: an environment sensing unit that detects the environment of the space and outputs an environment sensing data correspondingly; a a camera device for capturing and outputting a real-time image of the space, the real-time image includes at least one person; and a computing device for signally connecting the environment sensing unit, the camera device and the at least one air conditioner, the computing device stores a human body A posture recognition model, a clothing recognition model, and a human body comfort program kit; the computing device inputs the real-time image to the human posture recognition model and the clothing recognition model, and implements the human body according to the recognition information and the environment sensing data The calculation of the comfort degree program kit is to obtain a comfort degree evaluation value, and the operation device controls the operation state of the at least one air conditioner according to the comfort degree evaluation value; wherein, the operation device includes a storage unit to store the human body posture recognition model, Program data of the clothing recognition model and the human comfort program kit. The air-conditioning control system based on the comfort of persons according to claim 1, wherein the environment sensing unit comprises at least one of a dry bulb thermometer, a black bulb thermometer, a relative hygrometer and an air velocity meter; correspondingly, The environmental sensing data includes at least one of a dry bulb temperature signal, a black bulb temperature signal, a relative humidity signal and an air velocity signal. The air-conditioning control system based on the comfort of the person as claimed in claim 1, wherein the storage unit stores a metabolism comparison table and a clothing comparison table; the computing device inputs the real-time image into the human body posture recognition model to obtain a human body posture information, the human body posture information includes at least one posture recognition result and at least one person number value; the computing device obtains at least one metabolic reference value corresponding to the at least one posture recognition result in the metabolic comparison table by looking up a table; The computing device inputs the real-time image into the clothing recognition model to obtain clothing information, and the clothing information includes at least one clothing recognition result and at least one clothing quantity value; the computing device obtains and compares the clothing comparison table with the clothing quantity by looking up the table. at least one clothing reference value corresponding to the at least one clothing identification result; the computing device according to the environment sensing data, the at least one personnel quantity value, the at least one metabolic reference value, the at least one clothing quantity value and the at least one clothing reference value value implements the calculation of the human comfort program suite to obtain the comfort evaluation value. The air conditioning control system based on occupant comfort according to claim 3, wherein the storage unit stores an upper limit value, and the upper limit value is a numerical value; when the computing device determines that the comfort evaluation value is greater than the upper limit value, the control command output by the computing device to the at least one air conditioner is an activation command. The air conditioning control system based on occupant comfort according to claim 4, wherein the storage unit stores a lower limit value, and the lower limit value is a numerical value; when the computing device determines that the comfort evaluation value is smaller than the lower limit value , the control command output by the computing device to the at least one air conditioner is a shutdown command. The air-conditioning control system based on the comfort level of persons according to claim 3, wherein the storage unit stores a temperature setting value; when the computing device determines that the comfort evaluation value is greater than the temperature setting value, the computing device outputs the output to the temperature setting value. The control command of at least one air conditioner is a start command. The air-conditioning control system based on the comfort level of persons according to claim 1, wherein the at least one air-conditioning apparatus is a plurality of air-conditioning apparatuses, and the computing device synchronously controls the operation states of the plurality of air-conditioning apparatuses. The air-conditioning control system based on the comfort level of persons according to claim 1, wherein the at least one air-conditioning apparatus is a plurality of air-conditioning apparatuses, and the computing device controls the operation state of any one of the air-conditioning apparatuses respectively. The air-conditioning control system based on the comfort of persons as claimed in claim 8, wherein the real-time image is divided into a plurality of sub-images, and the computing device correspondingly controls the operation states of different air-conditioning equipment according to the identification information of the plurality of sub-images.

Images