Online equipment management evaluation method and system and electronic equipment

Abstract

The invention discloses an online equipment management evaluation method and system and electronic equipment, and belongs to the technical field of online equipment management evaluation, a sensing layer is established to obtain space image information, temperature and humidity information and personnel behavior information, and temperature and humidity data collected by a deployed temperature and humidity sensor are combined according to the space size and the number of workers, so that the online equipment management evaluation is realized. The method comprises the steps of establishing a heat balance model, determining target temperature and humidity, analyzing behaviors of workers, judging whether the workers leave an office or not, if yes, controlling an air conditioner to be turned off, recording on-duty and off-duty time of the workers, forming a historical data set, and calculating reference time by adopting a moving average algorithm. And according to the reference time, the air conditioner is controlled to be turned on by presetting time in advance, the relative position of a worker and an air outlet of the air conditioner and the manual adjusting behavior are analyzed, the preference of the worker for the air conditioner blowing mode is determined, and the air conditioner is controlled to be switched to the corresponding blowing mode.

Description

Technical Field

[0001] This invention relates to an online equipment management assessment, and more particularly to an online equipment management assessment method, system, and electronic device, belonging to the technical field of online equipment management assessment. Background Technology

[0002] Traditional air conditioning control relies heavily on manual adjustment or fixed-time start / stop, lacking awareness of human movement. Continuing to run the air conditioner after people leave leads to energy waste, or failing to start it before people enter results in an uncomfortable initial environment, making it difficult to balance energy consumption and comfort.

[0003] Poor spatial adaptability: Different office spaces have significant differences in structure and personnel density between small independent offices, medium-sized open office areas, and large conference rooms or exhibition halls. However, the existing control logic mostly adopts a uniform standard and does not formulate differentiated strategies for space size and layout characteristics. Large spaces require lower baseline temperatures due to their high heat capacity, while small spaces are easily affected by the heat generated by personnel and need to be dynamically adjusted. Traditional methods are difficult to balance these factors.

[0004] The precision of temperature and humidity control is insufficient. Current temperature and humidity regulation is mostly based on single sensor data or empirical values, without establishing a quantitative model that takes into account the space volume and the real-time number of people. It does not take into account the physical characteristics that the larger the space, the higher the air heat capacity, and the more energy is required to maintain a stable temperature, resulting in lag or deviation in regulation, which affects human comfort.

[0005] The lack of personalization and workstation adaptation means that employees have different preferences for air conditioning blowing modes, and the relative position of the workstation and the air vent directly affects comfort. Existing technologies lack the ability to learn from employees' behavioral habits and analyze workstation adaptability, making it difficult to achieve personalized control for each individual, and may even cause localized discomfort due to unreasonable air vent placement. Summary of the Invention

[0006] The main objective of this invention is to provide an online device management and evaluation method, system, and electronic device.

[0007] The objective of this invention can be achieved by adopting the following technical solution:

[0008] An online device management and evaluation method includes the following steps:

[0009] S11: Establish a perception layer to acquire spatial image information, temperature and humidity information, and personnel behavior information;

[0010] S12: Based on the size of the space and the number of staff, and combined with the temperature and humidity data collected by the deployed temperature and humidity sensors, establish a thermal balance model to determine the target temperature and humidity;

[0011] S13: Analyze staff behavior to determine whether staff have left the office; if staff have left, turn off the air conditioning.

[0012] S14: Record the staff's get off work hours to form a historical dataset, use a moving average algorithm to calculate the baseline time, and control the air conditioner to turn on in advance based on the baseline time;

[0013] S15: Analyze the relative position of staff and air conditioner vents and their manual adjustment behavior to determine staff preferences for air conditioner blowing modes and control the air conditioner to switch to the corresponding blowing mode.

[0014] Preferably, in S11, spatial images are captured by deployed cameras, the size of the office space is calculated by combining the reference objects in the images, and a target detection algorithm is used to perform human detection on the real-time frames of the cameras to obtain the number of staff.

[0015] Extract objects of known size from the image as reference objects, use the homography matrix to convert the image coordinates into physical coordinates, and calculate the area of ​​the office by multiplying its length and width.

[0016] Preferably, the specific process for establishing a thermal equilibrium model in a small space and determining the target temperature and humidity is as follows:

[0017] S21: Identify the office's length (L), width (W), and height (H), and calculate its volume.

[0018] V = L × W × H;

[0019] S22: Establishing a linear correction model based on the heat capacity formula:

[0020] -k ;

[0021] in, Based on the original comfort benchmark;

[0022] k is the correction factor;

[0023] log(V) represents the logarithmic weakening of the effect of extreme volumes;

[0024] S23: Determine the target temperature based on the number of staff:

[0025] = -0.3 = -k log(V)]-0.3×(N-1);

[0026] S24: Determine the target humidity:

[0027] = .

[0028] Preferably, the specific process for establishing a thermal balance model and determining the target temperature and humidity for medium-sized open-plan office areas and large spaces is as follows:

[0029] Combined volume V, personnel density D = N / V projection, where N is the number of people and the corrected temperature and humidity baseline.

[0030] Temperature regulation formula;

[0031] Basic comfort temperature In summer and winter (18-22℃), a volume correction factor α is introduced: ;

[0032] in:

[0033] α volume correction factor;

[0034] log(V) weakens the effect of extreme volumes;

[0035] β is the number of people correction factor.

[0036] Basic humidity =40-60%RH, combined with space ventilation characteristics;

[0037] = .

[0038] Preferably, the step of analyzing staff behavior through a camera to determine whether a staff member has left the office specifically involves: extracting the staff member's action sequence, movement trajectory, and action duration features;

[0039] Output the probability that a staff member leaves:

[0040] Human behavior analysis extracts key behavioral features through continuous frame analysis of cameras:

[0041] Pack your belongings, grab your bag, and head towards the door;

[0042] If the action lasts for more than 30 seconds and the direction of movement is stable, it is considered as leaving.

[0043] A brief departure is defined as a person getting up briefly, without a bag, and whose movement is indoors or outside without carrying the bag that is essential for commuting to and from get off work.

[0044] Output probability: ;

[0045] in, f 轨迹 , These are the normalized eigenvalues;

[0046] As weight;

[0047] b is the bias;

[0048] σ is the sigmoid function.

[0049] When P When the value is greater than 0.8, the air conditioner will be turned off.

[0050] Preferably, the step of recording employees' get off work hours via camera to form a historical dataset, calculating a baseline time using a moving average algorithm, and controlling the air conditioner to turn on in advance based on the baseline time specifically involves:

[0051] Commuting Habit Learning and Pre-start: Records employees' commuting times over 14 days using cameras to create a historical dataset. ;

[0052] The reference time treference is calculated using the moving average algorithm. ;

[0053] The preset time is 10 minutes. When t is reached... 基准 At -10 minutes, the air conditioner is turned on, and the baseline time is recalculated every 7 days.

[0054] Preferably, the step of analyzing the relative position of the staff member and the air conditioner vent through a camera and their manual adjustment behavior to determine the staff member's preference for the air conditioner's airflow mode, and controlling the air conditioner to switch to the corresponding airflow mode, specifically involves:

[0055] S31: When staff members are directly in front of the air outlet for more than 80% of the time and the distance is less than 2.5m, and no manual adjustment is made, it is judged as a direct airflow preference;

[0056] S32: When staff move frequently and the number of times is greater than 4 times / day or when they manually switch the swing mode, it is determined to be a swing preference;

[0057] S33: Based on the judgment result, control the air conditioner to switch to direct blowing mode or oscillation mode.

[0058] Preferably, by combining airflow mode and personnel thermal comfort feedback through camera recognition, facial micro-expressions and postures, such as frowning or removing a coat, are used to determine discomfort and establish a workstation score:

[0059] S=1- (Personnel location, wind location) + ;

[0060] Wherein, D ( This refers to the mismatch between the distance / direction of personnel and the air vent;

[0061] ( ) represents the comfort feedback weight;

[0062] , This is a correction factor;

[0063] When S < 0.6S, the marking station needs to be adjusted.

[0064] An online equipment management and evaluation system, the perception layer includes a camera and a temperature and humidity sensor, the camera is used to collect spatial images and images of staff behavior, and the temperature and humidity sensor is used to collect temperature and humidity data;

[0065] The processing layer receives data collected by the perception layer, calculates the size of the office space, obtains the number of staff, establishes a thermal balance model to determine the target temperature and humidity, analyzes staff behavior to determine whether to leave, records staff arrival and departure times to calculate the base time, and analyzes staff preferences for air conditioning blowing modes.

[0066] The control layer is used to control the air conditioner's on / off state, temperature and humidity adjustment, and fan mode switching based on the analysis results from the processing layer.

[0067] An electronic device, including a processor and a storage medium.

[0068] Beneficial technical effects of the present invention:

[0069] This invention provides an online equipment management and evaluation method, system, and electronic device. It identifies the size (length, width, and height) of a space using a camera and calculates its volume. It then establishes a temperature and humidity benchmark correction model based on the principle of heat capacity. Differentiated control strategies are formulated for small spaces (10-30㎡), medium-sized office areas (30-100㎡), and large spaces (over 100㎡), solving the problems of stuffy small spaces and excessively cold large spaces caused by traditional uniform standards.

[0070] Based on the YOLOv8 object detection algorithm, the system counts the number of people in real time. Combined with LSTM neural network analysis of people's behavior, it enables dynamic adjustment of temperature and humidity and intelligent start / stop of air conditioning. The air conditioning automatically turns off when people leave to avoid energy waste; when the number of people increases, it adjusts by reducing the temperature by 0.2-0.3℃ for each additional person to ensure the comfort of the group.

[0071] By learning employees' get off work habits over 14 days using a moving average algorithm, the air conditioning is turned on 10 minutes in advance to avoid unnecessary waiting; it is automatically turned off when the probability of employees leaving is greater than 0.8, reducing idling energy consumption. According to actual tests, this strategy can reduce air conditioning energy consumption by 15-20%, and is especially suitable for medium-sized office areas and large conference rooms with high employee turnover.

[0072] Precise regulation reduces redundant energy consumption: The temperature and humidity regulation logic is optimized for different space volumes and personnel densities to avoid excessive cooling / heating. In large spaces, the humidity benchmark is corrected by combining ventilation characteristics to reduce the ineffective operation of dehumidification / humidification equipment. In small spaces, rapid response avoids the surge in energy consumption caused by excessive environmental fluctuations.

[0073] By analyzing the relative position of people to the air vents and their manual adjustment behavior through cameras, the system can identify whether the airflow is direct or oscillating, and automatically switch the air conditioning mode to meet individual needs.

[0074] A workstation rating model (S=1-γ・D+δ・F) is established by combining thermal comfort feedback from employees (discomfort behaviors such as frowning and taking off coats). When the rating is <0.6, the workstation needs to be adjusted to help optimize the spatial layout, reduce local discomfort caused by unreasonable air vent positions, and improve the overall office experience. Attached Figure Description

[0075] Figure 1 This is a flowchart of a preferred embodiment of an online device management and evaluation method, system, and electronic device according to the present invention. Detailed Implementation

[0076] To enable those skilled in the art to understand the technical solution of the present invention more clearly, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but the embodiments of the present invention are not limited thereto.

[0077] Establish a perception layer (i.e., a data acquisition layer);

[0078] Deploy temperature and humidity sensors and light sensors to collect indoor environmental data in real time, providing basic information for adjusting the air conditioning operation status.

[0079] The deployment of temperature and humidity sensors and light sensors is based on;

[0080] For small offices (10-30㎡): Deploy 1-2 combined sensors in a single area, integrating temperature, humidity, and light functions, and install them on the wall in the center of the area, 1.5-1.8m above the ground, at the same height as human perception.

[0081] For medium-sized open-plan office areas (30-100㎡): sensors are deployed at a density of 5-8㎡ per sensor, using a grid pattern. In our application, a 60㎡ office area was divided into four zones, with one sensor installed in each zone on the ceiling or pillars, focusing on covering densely populated workstation areas and areas near air conditioning vents.

[0082] Large spaces (meeting rooms, exhibition halls, 100㎡ and above): The sensors are arranged in different areas according to the spatial structure. The meeting rooms are arranged in three rows: front, middle and back, with 2-3 sensors in each row. At the same time, one sensor is added near the air conditioner return air vent to help judge the overall environmental balance.

[0083] The temperature and humidity sensor is selected from industrial-grade devices (SensirionSHT3x series) with an accuracy of ±0.5℃ (temperature) and ±2%RH (humidity), supporting a wide range of -40℃ to 85℃ to adapt to fluctuations in the office environment.

[0084] The light sensor uses a photoresistor or a digital sensor (BH1750), with a measurement range of 0-65535 lux. It can identify environments with a mixture of natural light and artificial light, and the data update frequency is set to 10 seconds / time.

[0085] It uses low-power human infrared sensors and cameras to identify the presence and behavior of people and trigger the air conditioning control logic.

[0086] For the deployment of human infrared sensors and cameras;

[0087] Small office (10-30㎡):

[0088] One or two human infrared sensors can be deployed in suitable locations within the area, mounted on the wall at a height that allows for effective monitoring of indoor human activity, with a focus on monitoring the presence of people in the office area.

[0089] The cameras are placed in indoor corners to avoid directly facing the work areas of office workers, while covering the main activity areas. At the same time, privacy protection algorithms are used to process the collected images to protect the privacy of the staff.

[0090] Medium-sized open-plan office space (30-100㎡):

[0091] Human infrared sensors are deployed at a reasonable density in a grid pattern. In a 60㎡ office area, four zones can be divided, with one sensor installed in each zone. These sensors are mounted on the ceiling or pillars, focusing on covering densely populated workstation areas to ensure comprehensive monitoring of the presence of people within the zone.

[0092] A suitable number of cameras are installed in key locations in the office area, such as entrances and main passageways. Wide-angle lenses are used to expand the monitoring range, while privacy protection algorithms are used to process sensitive information in the images.

[0093] Large spaces (conference rooms, exhibition halls, over 100 square meters):

[0094] Human infrared sensors are deployed in different areas based on the spatial structure. For example, in a conference room, they can be arranged in three rows: front, middle, and back, with 2-3 sensors installed in each row. An additional sensor can be installed near the air conditioning return vent to help determine the overall distribution of people and provide more accurate data for air conditioning control.

[0095] Multiple cameras are installed in suitable locations within a large space to achieve coverage of the entire space. The installation locations should avoid causing direct visual oppression to people. Privacy protection algorithms are used to ensure that people's privacy is not leaked, and only necessary information such as the presence and general behavior of people is extracted.

[0096] Human body infrared sensors can be wall-mounted or ceiling-mounted. When wall-mounted, they should be fixed at an appropriate height on the wall; when ceiling-mounted, they should be embedded in the ceiling or fixed with brackets to ensure that the sensor's detection direction covers the target monitoring area.

[0097] The camera can be mounted on the ceiling, wall, or other locations using a bracket. Adjust the shooting angle to clearly monitor the people in the target area. At the same time, pay attention to the wiring layout to ensure normal power supply and data transmission for the equipment.

[0098] The system obtains air conditioner operating parameters (compressor frequency, refrigerant pressure) and control status via RS485 and WiFi communication protocols.

[0099] After the perception layer is established, the transmission layer is then established.

[0100] For small offices (10-30㎡), WiFi 6 wireless networks can be primarily used. Deploy WiFi 6 routers in suitable locations within the office to ensure signal coverage throughout the entire area, meeting the data transmission needs of devices such as temperature and humidity sensors, light sensors, infrared human body sensors, and cameras. Due to the relatively small number of devices, a single high-performance WiFi 6 router can provide a stable, high-speed connection.

[0101] For medium-sized open-plan office areas (30-100㎡), it is necessary to combine WiFi 6 and appropriate 5G supplementation. Multiple WiFi 6 routers can be deployed in reasonable locations in the office area, using Mesh networking technology to achieve seamless coverage and avoid signal blind spots. For sensors and cameras in some remote locations or where the signal is easily interfered with, 5G modules can be equipped to access the 5G network and ensure the stability and real-time performance of data transmission.

[0102] For large spaces (conference rooms, exhibition halls, over 100㎡), WiFi 6 and 5G networks should be used in combination. Multiple WiFi 6 routers should be evenly deployed throughout the space to build a high-density Mesh network. At the same time, 5G signal enhancement devices should be set up in key areas to ensure that all sensing layer devices can access the high-speed wireless network. For areas with high personnel flow and concentrated device use in the conference room, network capacity and stability need to be optimized.

[0103] After the transmission layer is built, the core analysis technology layer of this invention is entered. The design of this invention is to first use a camera to determine the size of the office and the number of staff. After the determination is completed, it is divided into small independent offices. Then, the temperature and humidity are monitored and adjusted to suit the office size. The behavior of the staff is analyzed to determine whether they need to leave the office to pack things or just leave temporarily. If they need to leave to pack things, the air conditioner is automatically turned off. Secondly, it is also necessary to determine the staff's work and commuting time habits to turn on the air conditioner 10 minutes in advance, and to determine whether they are used to direct airflow or oscillating airflow.

[0104] Therefore, the specific technical solution established is as follows:

[0105] Space size determination: Based on panoramic images captured by cameras, the space size is calculated through perspective transformation and feature matching. Objects of known size in the image (such as a standard desk 1.2m long and a door 0.9m wide) are extracted as reference objects. The image coordinates are converted into physical coordinates using the homography matrix. The area of ​​the office is calculated by multiplying the length and width, and it is determined whether it is a small space of 10-30㎡.

[0106] The calculation formula uses S= (Unit: m) 2 );

[0107] Where L pixel and W pixel are the spatial length / width pixel values ​​in the image, and k is the pixel-to-physical size conversion coefficient (calibrated by a reference object).

[0108] The personnel count uses the YOLOv8 target detection algorithm to perform human detection on real-time camera frames and output the number of people N (accuracy ≥ 95%, supports 1-10 person scenarios).

[0109] In the regulation of air conditioning temperature, a heat balance model is established based on the control dimensions collected by the camera, the collection of temperature and humidity data, and the analysis of the number of people.

[0110] Small spaces (10㎡) have limited air volume, and the heat generated by people can cause the temperature to rise rapidly, requiring a higher baseline temperature;

[0111] In large spaces (30㎡), the amount of air is large, and the heat generated by people is difficult to change the overall temperature, so a lower base temperature is required;

[0112] Using cameras and deep learning, the length (L), width (W), and height (H) of the office are first identified, and the volume is calculated as: V = L × W × H;

[0113] A model relating reference temperature to spatial volume;

[0114] Based on the heat capacity formula Q=c m ΔT (where c is the specific heat capacity of air and m is the mass of air), derivation: the larger the volume of space (the more air), the more heat is required to maintain the same temperature change;

[0115] Therefore, the larger the volume, the lower the reference temperature should be, and a linear correction model should be established:

[0116] -k ;

[0117] in, The original comfort standard is 25°C, which can be slightly adjusted according to the season.

[0118] k is a correction factor (empirical value is 0.5 to 1.2, needs to be calibrated on site, k is smaller in small spaces);

[0119] log(V) is used to weaken the effect of extreme volumes (to avoid excessively large volumes that could lead to temperature baseline anomalies).

[0120] Final temperature and humidity control formula:

[0121] General temperature formula (integrating "space volume + number of people"): = -0.3 = -k log(V)]-0.3×(N-1);

[0122] Humidity formula (retaining the original logic, supplementing extreme scenarios related to volume):

[0123] = ;

[0124] When the sensor reading deviates from the target by ±0.5℃ (small space) When the humidity is ±3%RH (small space) or ±5%RH (large space), the adjustment is triggered (because the temperature and humidity fluctuate faster in small spaces, a more sensitive response is required).

[0125] Personnel behavior analysis (determination of absence type) extracts key behavioral features through continuous frame analysis of cameras:

[0126] Gathering items (frequent hand contact with desktop / drawer), carrying bag, and heading towards the door (movement trajectory towards exit).

[0127] Actions lasting more than 30 seconds and with a stable direction of movement are considered leaving; brief standing up without a bag, and the trajectory being indoors or outside without carrying the bag that is usually carried to and from get off work are considered brief leaving.

[0128] An LSTM neural network-based behavior classifier outputs probabilities:

[0129] ;

[0130] in, f 轨迹 , For normalized eigenvalues, σ is the weight, b is the bias, and σ is the sigmoid function.

[0131] When P((away)>0.8, the air conditioner is triggered to turn off.

[0132] Commuting Habit Learning and Pre-start: Records employees' commuting times over 14 days using cameras to create a historical dataset.

[0133] The base time t is calculated using the moving average algorithm. 基准 ;

[0134] When P (away) > 0.8, the air conditioner is triggered to turn off.

[0135] Adaptive updates: The baseline time is recalculated every 7 days to adapt to changes in habits (such as weekend adjustments).

[0136] Hair blowing mode compatibility (direct blowing / oscillating);

[0137] Record the correlation between the employee's workstation coordinates (x, y) and the direction of the air conditioner vent over a 30-day period:

[0138] Personnel spend more than 80% of their time directly in front of the air outlet (at a distance of less than 2 meters) and do not manually adjust the air outlet;

[0139] Frequent personnel movement (>3 times / day) or manual switching of swing mode.

[0140] If a direct airflow preference is detected, the air conditioner will fix the direction of the air outlet; otherwise, it will automatically switch to oscillation mode.

[0141] Using spatial images captured by a camera, the spatial type and size are determined through semantic segmentation and 3D reconstruction techniques.

[0142] Using the DeepLabV3+ model, the system identifies structures such as walls, ceilings, doors, and windows in images, delineates spatial boundaries, and, combined with monocular visual geometry, calculates the length (L), width (W), and height (H) of the space using known reference objects (standard office desk and chair dimensions, door and window specifications). It then derives the volume V = L × W × H and determines the type based on the volume range.

[0143] Medium-sized open-plan office space: 30㎡ (Default floor height H=3m);

[0144] Large spaces (meeting rooms / exhibition halls) .

[0145] The larger the space, the higher the air heat capacity. Therefore, the temperature and humidity baseline needs to be adjusted based on volume (V) and personnel density (D=N / V projection, where N is the number of people).

[0146] Temperature regulation formula

[0147] Basic comfort temperature (Summer), 18-22℃ (Winter), introduce a volume correction factor α (negatively correlated with space volume, with a more significant impact on heat production by people in small spaces): ;

[0148] Wherein: α is the volume correction coefficient (empirical value of 0.3 for medium-sized spaces and 0.5 for large spaces, which needs to be calibrated on site), and log(V) weakens the influence of extreme volumes;

[0149] β-number of people correction factor (uniformly set at 0.2; for every 1 increase in the number of people, the temperature decreases by 0.2℃ to avoid excessive cooling in large spaces).

[0150] Basic humidity =40-60%RH, taking into account the ventilation characteristics of the space (large spaces have good ventilation, but the humidity tends to be low).

[0151] = ;

[0152] When the temperature and humidity sensor (deployment density: 50㎡ / unit for medium-sized areas, 100㎡ / unit for large spaces) deviates from the target by ±1℃ (medium-sized) / ±1.5℃ (large-sized) or ±5%RH (medium-sized) / ±8%RH (large-sized) the air conditioning will be adjusted accordingly.

[0153] Data is collected continuously for 14 days by recording employees' arrival and departure times via cameras.

[0154] The baseline time (tbaseline) is calculated using a moving average plus outlier filtering algorithm. ;

[0155] Pre-start time: = -10 minutes, recalculate adaptation habit changes weekly.

[0156] The camera analysis reveals the relative position of people to the air conditioning vents and their manual adjustment behavior.

[0157] Direct airflow preference: People are directly in front of the air outlet (distance < 2.5m) 70% of the time, and no manual adjustment is made;

[0158] Swing preference: Frequent personnel movement (>4 times / day) or manual switching of swing mode.

[0159] Output mode label M∈{direct blowing, oscillating}, automatically adapts to air conditioning mode.

[0160] Workstation adaptability analysis: Combining airflow mode and employee thermal comfort feedback (detected through camera, facial micro-expressions, and postures such as frowning or removing coats to determine discomfort), a workstation rating was established: S=1- (Personnel location, wind location) + ;

[0161] Wherein, D ( The distance / direction mismatch between the person and the air vent (in direct airflow mode, the greater the distance and the more off-center the direction, the larger D is);

[0162] ( The F-value is the weight for comfort feedback (the more uncomfortable behaviors, the higher the F-value, and the lower the score).

[0163] , This is a correction factor (empirical values ​​are 0.6 and 0.4).

[0164] When S < 0.6S, the marked workstation needs to be adjusted (prompt the administrator to optimize the workstation layout and move the air vent).

[0165] The above description is merely a further embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope disclosed in the present invention, based on the technical solution and concept of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. An online equipment management and evaluation method, characterized in that: Includes the following steps: S11: Establish a perception layer to acquire spatial image information, temperature and humidity information, and personnel behavior information; S12: Based on the size of the space and the number of staff, and combined with the temperature and humidity data collected by the deployed temperature and humidity sensors, establish a thermal balance model to determine the target temperature and humidity; S13: Analyze staff behavior to determine whether staff have left the office; if staff have left, turn off the air conditioning. S14: Record the staff's get off work hours to form a historical dataset, use a moving average algorithm to calculate the baseline time, and control the air conditioner to turn on in advance based on the baseline time; S15: Analyze the relative position of staff and air conditioner vents and their manual adjustment behavior to determine staff preferences for air conditioner blowing modes and control the air conditioner to switch to the corresponding blowing mode.

2. The online equipment management and evaluation method according to claim 1, characterized in that: In S11, spatial images are captured by deployed cameras, and the size of the office space is calculated by combining the reference objects in the images. A target detection algorithm is used to perform human detection on the real-time frames of the cameras to obtain the number of staff. Extract objects of known size from the image as reference objects, use the homography matrix to convert the image coordinates into physical coordinates, and calculate the area of ​​the office by multiplying its length and width.

3. The online equipment management and evaluation method according to claim 1, characterized in that: The specific process of establishing a thermal equilibrium model in a small space and determining the target temperature and humidity is as follows: S21: Identify the office's length (L), width (W), and height (H), and calculate its volume. V = L × W × H; S22: Establishing a linear correction model based on the heat capacity formula: -k ； in, Based on the original comfort benchmark; k is the correction factor; log(V) represents the logarithmic weakening of the effect of extreme volumes; S23: Determine the target temperature based on the number of staff: = -0.3 = -k log(V)]-0.3×(N-1); S24: Determine the target humidity: = 。 4. The online equipment management and evaluation method according to claim 1, characterized in that: The specific process for establishing a thermal balance model and determining target temperature and humidity for medium-sized open-plan office areas and large spaces is as follows: Combined volume V, personnel density D = N / V projection, where N is the number of people and the corrected temperature and humidity baseline. Temperature regulation formula; Basic comfort temperature In summer and winter (18-22℃), a volume correction factor α is introduced: ; in: α volume correction factor; log(V) weakens the effect of extreme volumes; β is the number of people correction factor. Basic humidity =40-60%RH, combined with space ventilation characteristics; = 。 5. The online equipment management and evaluation method according to claim 2, characterized in that: The step of analyzing staff behavior through cameras to determine whether staff have left the office specifically involves: extracting the staff's action sequence, movement trajectory, and action duration features; Output the probability that a staff member leaves: Human behavior analysis extracts key behavioral features through continuous frame analysis of cameras: Pack your belongings, grab your bag, and head towards the door; If the action lasts for more than 30 seconds and the direction of movement is stable, it is considered as leaving. A brief departure is defined as a person getting up briefly, without a bag, and whose movement is indoors or outside without carrying the bag that is essential for commuting to and from get off work. Output probability: ; in, f 轨迹 , These are the normalized eigenvalues; As weight; b is the bias; σ is the sigmoid function. When P When the value is greater than 0.8, the air conditioner will be turned off.

6. The online equipment management and evaluation method according to claim 1, characterized in that: The process involves recording employees' arrival and departure times via cameras to create a historical dataset, calculating a baseline time using a moving average algorithm, and then controlling the air conditioning to turn on in advance based on this baseline time. Specifically: Commuting Habit Learning and Pre-start: Records employees' commuting times over 14 days using cameras to create a historical dataset. ; The reference time treference is calculated using the moving average algorithm. ; The preset time is 10 minutes. When t is reached... 基准 At -10 minutes, the air conditioner is turned on, and the baseline time is recalculated every 7 days.

7. The online equipment management and evaluation method according to claim 1, characterized in that: The method involves analyzing the relative position of staff members to the air conditioning vents and their manual adjustment behavior using a camera to determine the staff members' preferred air conditioning blowing mode and controlling the air conditioning to switch to the corresponding blowing mode. Specifically: S31: When staff members are directly in front of the air outlet for more than 80% of the time and the distance is less than 2.5m, and no manual adjustment is made, it is judged as a direct airflow preference; S32: When staff move frequently and the number of times is greater than 4 times / day or when they manually switch the swing mode, it is determined to be a swing preference; S33: Based on the judgment result, control the air conditioner to switch to direct blowing mode or oscillation mode.

8. The online equipment management and evaluation method according to claim 7, characterized in that: Combining airflow patterns and employee thermal comfort feedback via camera recognition, micro-expressions and postures such as frowning and removing coats are used to determine discomfort and establish a workstation rating system. S=1- (Personnel location, wind location) + ; Wherein, D ( This refers to the mismatch between the distance / direction of personnel and the air vent; ( ) represents the comfort feedback weight; , This is a correction factor; When S < 0.6S, the marking station needs to be adjusted.

9. An online equipment management and evaluation system, based on the online equipment management and evaluation method described in claims 1-8, characterized in that: The perception layer includes a camera and a temperature and humidity sensor. The camera is used to acquire spatial images and images of staff behavior, and the temperature and humidity sensor is used to acquire temperature and humidity data. The processing layer receives data collected by the perception layer, calculates the size of the office space, obtains the number of staff, establishes a thermal balance model to determine the target temperature and humidity, analyzes staff behavior to determine whether to leave, records staff arrival and departure times to calculate the base time, and analyzes staff preferences for air conditioning blowing modes. The control layer is used to control the air conditioner's on / off state, temperature and humidity adjustment, and fan mode switching based on the analysis results from the processing layer.

10. An electronic device, characterized in that: It includes a processor and a storage medium, wherein a computer program is stored on the storage medium, and the processor executes the computer program to implement the online device management and evaluation method as described in any one of claims 1-8.

Images