Air conditioner control method, device, medium and air conditioner

Abstract

The application discloses an air conditioner control method, device, medium and air conditioner, target face temperature data of a target object is acquired in real time, a target thermal sensation recognition model of the target object is matched based on target temperature characteristic parameters, the current target thermal sensation of the human body is recognized based on the target thermal sensation recognition model and the face temperature data, and finally the air conditioner is directly controlled according to the target thermal sensation; the application can adapt to the differences between individuals through dynamic matching of the model, and can directly capture the real thermal sensation of the human body, effectively overcoming the defects that the traditional control mode cannot perceive the real-time comfort feedback of the human body, and the output of the air conditioner is deviated from the actual thermal sensation of the human body, so that the accuracy and comfort of temperature regulation are significantly improved.

Description

Technical Field

[0001] This invention relates to the field of air conditioning technology, and in particular to an air conditioning control method, device, medium, and air conditioner. Background Technology

[0002] In existing technologies, air conditioning terminals are mainly controlled based on manual commands. However, since this method cannot obtain real-time thermal comfort feedback from the human body, there is a discrepancy between the air conditioning output and the actual hot or cold feeling of the human body, which in turn affects the accuracy and comfort of the control. Summary of the Invention

[0003] Therefore, it is necessary to provide air conditioning control methods, devices, media, and air conditioners to solve the problem that the output of existing air conditioning systems deviates from the actual hot and cold sensations of the human body, affecting the accuracy and comfort of control.

[0004] In a first aspect, embodiments of this application provide an air conditioning control method, the method comprising: Obtain the facial temperature data of the target object; Based on the target temperature feature parameters of the facial temperature data, a target thermal perception recognition model for the target object is determined. The facial temperature data is processed based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object. The operation of the air conditioner is controlled based on the target thermal sensation.

[0005] In some embodiments of this application, the target thermal sensation recognition model is obtained by filtering candidate thermal sensation recognition models based on the target temperature feature parameters of the facial temperature data. Thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The candidate thermal sensation recognition model includes a first critical temperature, which is used to distinguish between the comfortable state and the slightly cold state, or to distinguish between the comfortable state and the slightly hot state. The method for determining the first critical temperature includes: Acquire first training data and second training data; wherein, the first training data corresponds to the comfort state, and the second training data corresponds to the slightly cold state or the slightly hot state; Based on the first training data and the second training data, a target projection vector is determined; wherein, the target projection vector is used to minimize the difference between training data of the same class after projection and to maximize the difference between training data of different classes after projection. Based on the target projection vector, the first training data and the second training data are projected forward to obtain the target projection value; The target projection value is back-projected based on the target projection vector to obtain the first critical temperature.

[0006] In some embodiments of this application, determining the target projection vector based on the first training data and the second training data includes: The centroid is calculated on the first training data to obtain the first centroid value; The second training data is used to calculate the centroid to obtain the second centroid value; The first training data is subjected to discreteness calculation based on the first centroid value to obtain the first discreteness matrix; The second training data is discrete based on the second centroid value to obtain the second discreteness matrix; The first and second discrete values ​​are linearly summed to obtain the total discrete value matrix; The difference between the first centroid value and the second centroid value is multiplied by the inverse of the total discreteness matrix to obtain the target projection vector.

[0007] In some embodiments of this application, the step of forward projecting the first training data and the second training data based on the target projection vector to obtain the target projection value includes: The centroid is calculated on the first training data to obtain the first centroid value; The second training data is used to calculate the centroid to obtain the second centroid value; The target projection vector is projected onto the first centroid value to obtain the first projection value. The target projection vector is projected onto the second centroid value to obtain the second projection value; The target projection value is obtained by performing a weighted operation based on the number of the first training data, the number of the second training data, the first projection value, and the second projection value.

[0008] In some embodiments of this application, obtaining the facial temperature data of the target object includes: Acquire the first infrared image of the target object; Based on the first infrared image, a target temperature gradient image is determined; A connected component search is performed on the target temperature gradient image to obtain several zero-element regions; wherein, the zero-element regions represent connected regions formed by pixels with a temperature gradient of 0. Coordinate mapping is performed on the aforementioned zero-element regions to obtain several face mapping regions in the first infrared image; Calculate the first average temperature of the plurality of facial mapping regions, and determine the facial mapping regions whose first average temperature belongs to the human body temperature range as the target temperature region. The temperature within the target temperature region is extracted to obtain the facial temperature data of the target object.

[0009] In some embodiments of this application, determining the target thermal perception recognition model of the target object based on the target temperature feature parameters of the facial temperature data includes: Obtain candidate thermal sensation recognition models, and the different candidate temperature feature parameters corresponding to different candidate thermal sensation recognition models; The distance between the target temperature feature parameter of the facial temperature data and the different candidate temperature feature parameters is calculated to obtain the parameter distance; The candidate thermal sensing recognition model corresponding to the smallest parameter distance is determined as the target thermal sensing recognition model.

[0010] In some embodiments of this application, the target temperature feature parameter and the candidate temperature feature parameter include a first category parameter, a second category parameter and a third category parameter. The first category parameter represents the mean between the maximum and minimum values ​​in the facial temperature data, the second category parameter represents the volatility of the facial temperature data, and the third category parameter represents the difference between the maximum and minimum values ​​in the facial temperature data. The step of calculating the distance between the target temperature feature parameter of the facial temperature data and the different candidate temperature feature parameters to obtain the parameter distance includes: The distance between the first category parameter of the target temperature feature parameter and the first category parameter of the candidate temperature feature parameter is calculated to obtain a first distance; The distance between the second category parameter of the target temperature feature parameter and the second category parameter of the candidate temperature feature parameter is calculated to obtain the second distance; The distance between the third category parameter of the target temperature feature parameter and the third category parameter of the candidate temperature feature parameter is calculated to obtain the third distance; The first distance, the second distance, and the third distance are weighted and summed to obtain the parameter distance.

[0011] In some embodiments of this application, thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The target thermal sensation recognition model includes a second critical temperature and a third critical temperature. The second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state. The step of processing the facial temperature data based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object includes: The average facial temperature is calculated by averaging the facial temperature data. If the average facial temperature is less than the second critical temperature, then the cool state is determined as the target thermal sensation of the target object. If the average facial temperature is greater than or equal to the second critical temperature and less than or equal to the third critical temperature, then the comfortable state is determined as the target thermal sensation of the target object. If the average facial temperature is greater than the third critical temperature, then the preheated state is determined as the target thermal sensation of the target object.

[0012] In some embodiments of this application, thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The target thermal sensation recognition model includes a second critical temperature and a third critical temperature. The second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state. After controlling the operation of the air conditioner based on the target thermal sensation, the method further includes: When the parameter adjustment instruction of the target object is obtained, the parameter adjustment instruction is mapped to obtain the equivalent thermal sensation vote corresponding to the parameter adjustment instruction; The current facial temperature data of the target object and the equivalent thermal sensation vote corresponding to the parameter adjustment command are determined as the current training data, and the second critical temperature or the third critical temperature is updated based on the current training data.

[0013] Secondly, embodiments of this application also provide an air conditioning control device, the air conditioning control device comprising: The data acquisition module is used to acquire the facial temperature data of the target object; The model determination module is used to determine the target thermal perception recognition model of the target object based on the target temperature feature parameters of the facial temperature data. A thermal sensation recognition module is used to recognize and process the facial temperature data based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object; The operation control module is used to control the operation of the air conditioner based on the target thermal sensation.

[0014] Thirdly, this application also provides an air conditioner, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps in the above-described air conditioning control method.

[0015] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps in the aforementioned air conditioning control method.

[0016] Fifthly, embodiments of this application also provide a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in the various optional implementations described in embodiments of this application.

[0017] This invention provides an air conditioning control method, device, medium, and air conditioner. It acquires real-time facial temperature data of a target object and matches a target thermal sensation recognition model specific to that object based on its target temperature characteristic parameters. Then, based on this target thermal sensation recognition model and facial temperature data, it identifies the current target thermal sensation of the human body and ultimately controls the air conditioner's operation directly based on this target thermal sensation. This invention can adapt to individual differences through a dynamic matching model and can directly capture the human body's actual thermal sensation, effectively overcoming the shortcomings of traditional control methods that suffer from "deviation between air conditioning output and the actual hot / cold sensation of the human body" due to the inability to perceive real-time comfort feedback. This significantly improves the accuracy and comfort of temperature regulation. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] in: Figure 1 A schematic flowchart illustrating the air conditioning control method provided in an embodiment of this application; Figure 2 A flowchart illustrating the process of obtaining facial temperature data of a target object; Figure 3 This is a flowchart illustrating the method for determining the first critical temperature. Figure 4 A flowchart illustrating the target thermal sensing recognition model for identifying target objects; Figure 5 A flowchart illustrating the process of determining the target thermal sensation of a target object; Figure 6 This is a schematic diagram of the air conditioning control device. Figure 7 This is a structural block diagram of an air conditioner. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0022] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0023] This invention provides an air conditioning control method, apparatus, medium, and air conditioner. In some embodiments of this application, the provided air conditioning control method can be applied to an air conditioner. Specifically, the air conditioner can be applied to different scenarios, including but not limited to industrial air conditioners or household air conditioners. In some embodiments of this application, the air conditioner can be a single unit, such as a cabinet air conditioner or a wall-mounted air conditioner; in some embodiments of this application, the air conditioner can also be a central air conditioning system composed of multiple air conditioner units, such as a multi-split air conditioner, an air-cooled heat pump system, or an air conditioning system with heat recovery function.

[0024] Please see Figure 1 , Figure 1 This is a flowchart illustrating an air conditioning control method provided in an embodiment of this application. Although the logical sequence is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown in the figures. Specifically, the specific flow of this air conditioning control method is as follows: S101, Obtain the facial temperature data of the target object.

[0025] The target object represents the human body for which face extraction is required, including but not limited to children, adults, and the elderly. Facial temperature data represents the temperature value of the facial region of the target object.

[0026] Optionally, the face of the target object can be scanned by a temperature acquisition device to obtain facial temperature data; the temperature acquisition device includes, but is not limited to, contact temperature sensors and non-contact infrared thermometers.

[0027] In some embodiments of this application, such as Figure 2 As shown, S101, which involves acquiring the facial temperature data of the target object, includes steps S1011-S1016, as detailed below: S1011, acquire the first infrared image of the target object.

[0028] Optionally, the surface temperature distribution characteristics of the target object represented by the first infrared image can be obtained through the following m-row n-column temperature matrix. Indicate:

[0029] In the above formula, m is the total number of rows of pixels in the infrared image, and n is the total number of columns of pixels in the infrared image; The value of the pixel in the i-th row and j-th column. , .

[0030] Optionally, an infrared imaging device can be used to capture images of the target's face and upper body. This infrared imaging device includes, but is not limited to, miniature infrared thermal imaging modules, low-cost infrared temperature measurement cameras, and portable infrared imagers. It can also be integrated with infrared acquisition modules in terminal products such as smart access control systems, infrared temperature measurement terminals, and smart wearable devices; the options are not limited here.

[0031] S1012, Based on the first infrared image, determine the target temperature gradient image.

[0032] In some embodiments of this application, S1012, determining the target temperature gradient image based on the first infrared image, specifically includes the following steps: performing temperature standard deviation sliding calculation on the first infrared image based on a temperature gradient window to obtain a first temperature gradient image; segmenting the first temperature gradient image based on a segmentation threshold to obtain a first segmented region and a second segmented region in the first temperature gradient image; performing denoising processing on the first segmented region, and / or enhancing processing on the second segmented region to obtain the target temperature gradient image.

[0033] The temperature gradient window represents a window used to define the pixel range for each temperature standard deviation sliding calculation. The first temperature gradient image represents an initial image of the temperature change intensity at the corresponding calculation location. The temperature gradient of pixels within the first segmentation region is less than the temperature gradient of pixels within the second segmentation region.

[0034] Optionally, the temperature gradient window adopts a square pixel window design with a resolution of k×k pixels. The value of k can be flexibly set according to the actual resolution of the first infrared image and the requirements for temperature feature extraction. Within each window unit... Within a given window, the temperature standard deviation can be calculated using the following formula to characterize the degree of temperature variation within that range:

[0035] In the above formula, SD is the standard deviation of temperature within the temperature gradient window. This represents the average temperature of all pixels within the temperature gradient window. This represents the total number of pixels within the temperature gradient window. .

[0036] When performing sliding calculations for temperature standard deviation, using As a single computational unit, starting from the top-left pixel position of the first infrared image, the temperature standard deviation of the entire region on the first infrared image is calculated in a row-by-row, column-by-column, non-overlapping manner to obtain the first temperature gradient image. The pixel distribution of this first temperature gradient image corresponds one-to-one with the result of the sliding calculation, which can be represented by the following temperature gradient matrix. Indicate:

[0037] In the above formula, Let be the temperature gradient of the pixel in the i-th row and j-th column. , .

[0038] Optionally, the global standard deviation of the temperature gradient matrix is ​​used as the segmentation threshold τ. All pixels in the image with a temperature gradient less than or equal to τ are grouped into the same set to form the first segmentation region. The pixels in this region have a smaller temperature gradient, indicating that the temperature change at the corresponding location is gradual. All pixels in the image with a temperature gradient greater than τ are grouped into another set to form the second segmentation region. The pixels in this region have a larger temperature gradient, indicating that there is a significant temperature step at the corresponding location.

[0039] In some embodiments of this application, the first segmented region is subjected to noise reduction processing, which specifically includes the following steps: setting the temperature gradient of each pixel in the first segmented region to a first value.

[0040] In some embodiments of this application, the second segmented region includes a first pixel. The enhancement processing of the second segmented region specifically includes the following steps: subtracting a segmentation threshold from the first temperature gradient of the first pixel to obtain a first calculated value; performing a maximum operation on the first calculated values ​​corresponding to different pixels in the second segmented region to obtain a second calculated value; dividing the first calculated value by the second calculated value to obtain a third calculated value; adding the third calculated value to the second calculated value to obtain a fourth calculated value; and multiplying the fourth calculated value by the first temperature gradient to obtain a second temperature gradient.

[0041] The second temperature gradient is the first temperature gradient after the enhancement treatment.

[0042] Optionally, the first value A value of 0 quickly eliminates weak temperature gradient noise within the region. Of course, the first value... You can also set other smaller preset values.

[0043] Alternatively, the formula for the enhancement process can be expressed as:

[0044] In the above formula, This represents the second temperature gradient; The second value, the second value It can be set to 1, or other values.

[0045] In the above embodiment, by performing gradient unification denoising on the first segmented region, the interference of temperature gradient noise inside the face is eliminated; at the same time, nonlinear enhancement operation is performed on the second segmented region, which greatly amplifies the temperature boundary step features between the face and the background, making the contour distinction more significant.

[0046] Optionally, after completing the initial region segmentation, denoising, and enhancement processing to obtain the target temperature gradient image, the obtained target temperature gradient image can be used as a new first temperature gradient image to restart the region optimization process. The complete steps of "segmenting the first temperature gradient image based on the segmentation threshold to obtain the first segmented region and the second segmented region in the first temperature gradient image; denoising the first segmented region; and / or enhancing the second segmented region to obtain the target temperature gradient image" can be repeated to further enhance the overall optimization effect, making the target temperature gradient image more accurate in representing the temperature boundary step features and more distinct in the regional feature contrast.

[0047] S1013, perform connected component search in the target temperature gradient image to obtain several zero-element regions.

[0048] The zero-element region represents the connected region formed by pixels with a temperature gradient of 0.

[0049] Optionally, the identification and extraction of zero-element regions can be achieved based on connected component search algorithms such as depth-first search and breadth-first search. Taking the depth-first search algorithm as an example, its specific process includes: First, performing a full traversal of the pixel matrix of the target temperature gradient image row by row and column by column, sequentially reading the coordinates of each pixel and its corresponding temperature gradient value, and taking the first pixel with a temperature gradient value of 0 and no label as the initial seed point; then, starting from this initial seed point, performing a depth traversal of the surrounding pixels according to the four-neighbor connectivity rule (up, down, left, right) or the eight-neighbor connectivity rule (up, down, left, right, and diagonal), checking whether the gradient value of each neighboring pixel is 0 and no label, and if the condition is met, including the neighboring pixel in the current connected region. The process continues, starting from the current pixel and traversing its neighborhood. During this traversal, all pixels included in the connected region are uniformly marked to avoid repeated traversal and region confusion, until all neighborhoods of the current seed point have no matching pixels, thus completing the search and extraction of a zero-element region. Then, the process returns to the pixel traversal process, finding the next unmarked pixel with a temperature gradient value of 0 as the new seed point, and repeating the neighborhood traversal, pixel marking, and region aggregation steps. This process is repeated until all pixels in the target temperature gradient image are traversed and detected, and finally, all pixels with a temperature gradient value of 0 in the image are aggregated into several independent, non-overlapping zero-element regions based on connectivity.

[0050] S1014, coordinate mapping is performed on several zero-element regions to obtain several face mapping regions in the first infrared image.

[0051] In some embodiments of this application, the target temperature gradient image is obtained by processing the first infrared image with a temperature gradient window. The length of the temperature gradient window is a third value, the width of the temperature gradient window is a fourth value, and several zero-element regions include a first zero-element region, the first zero-element region includes a second pixel, and several face mapping regions include a first face mapping region corresponding to the first zero-element region, the first face mapping region includes a third pixel.

[0052] S1014 performs coordinate mapping on several zero-element regions to obtain several face mapping regions in the first infrared image, specifically including the following steps: halving the third value to obtain the fifth value; halving the fourth value to obtain the sixth value; adding the horizontal coordinate of the second pixel to the fifth value, and adding the vertical coordinate of the second pixel to the sixth value to obtain the coordinates of the third pixel.

[0053] Optionally, the third value = the fourth value = k, and the coordinate mapping process can be represented by the following formula:

[0054]

[0055] In the above formula, Let be the original coordinates of any second pixel within the first zero-element region. y is the x-coordinate of the second pixel, and y is the y-coordinate of the second pixel; Let be the target coordinates of the third pixel within the first face mapping region. It is the fifth or sixth operand value.

[0056] The above embodiments, through a coordinate mapping method based on the temperature gradient window size, accurately map the zero-element region in the temperature gradient image back to the original infrared image, achieving precise alignment between gradient features and original image pixels, and ensuring the accuracy of face mapping region positioning.

[0057] S1015, calculate the first average temperature of several face mapping regions, and determine the face mapping regions whose first average temperature belongs to the human body temperature range as the target temperature region.

[0058] Optionally, the human body temperature range can be set according to the actual application scenario of infrared thermometry. In a normal scenario, it can be set to 35℃-42℃. Only the facial mapping area with the first average temperature falling within the range of 35℃ to 42℃ is retained as the target temperature area.

[0059] S1016, extract the temperature within the target temperature region to obtain the facial temperature data of the target object.

[0060] The embodiments S1011-S1016 described above not only utilize the step characteristics of the temperature gradient to quickly lock the facial contour, but also eliminate non-human interference through temperature verification, which greatly improves the accuracy of face extraction in low-resolution infrared scenes.

[0061] S102, Based on the target temperature feature parameters of facial temperature data, determine the target thermal sensation recognition model of the target object.

[0062] Among them, the target temperature characteristic parameters represent the statistical characteristics obtained after quantitative analysis of facial temperature data, including but not limited to the average temperature of the whole face and the regional temperature difference.

[0063] In some embodiments of this application, the target thermal sensation recognition model is obtained by filtering candidate thermal sensation recognition models based on target temperature feature parameters of facial temperature data. Thermal sensation includes a cool state, a comfortable state, and a warm state. The candidate thermal sensation recognition model includes a first critical temperature, which is used to distinguish between a comfortable state and a cool state, or between a comfortable state and a warm state. It is understood that the logic for determining the first critical temperature for distinguishing between a comfortable state and a cool state is basically the same as the logic for determining the first critical temperature for distinguishing between a comfortable state and a warm state. For ease of explanation, the following description mainly uses the distinction between a comfortable state and a cool state as an example.

[0064] like Figure 3 As shown, the method for determining the first critical temperature includes steps S102a-S102d: S102a, Obtain the first training data and the second training data.

[0065] The first training data corresponds to a comfortable state, and the second training data corresponds to a slightly cold or slightly hot state.

[0066] For example, multiple sets of data are collected in a stable environment for a candidate user A beforehand. When candidate user A explicitly reports that they are in a comfortable state, facial temperature data of candidate user A is continuously collected and compiled into the first training data. When candidate user A explicitly reports that they are in a slightly cold state, facial temperature data of candidate user A is collected simultaneously and compiled into the second training data.

[0067] S102b, based on the first training data and the second training data, determines the target projection vector.

[0068] The target projection vector is used to minimize the difference between training data of the same class after projection and to maximize the difference between training data of different classes after projection.

[0069] In some embodiments of this application, S102b, determining the target projection vector based on the first training data and the second training data, specifically includes the following steps: Calculating the centroid of the first training data to obtain a first centroid value; Calculating the centroid of the second training data to obtain a second centroid value; Calculating the dispersion of the first training data based on the first centroid value to obtain a first dispersion matrix; Calculating the dispersion of the second training data based on the second centroid value to obtain a second dispersion matrix; Linearly summing the first and second dispersion values ​​to obtain a total dispersion matrix; Multiplying the difference between the first and second centroid values ​​by the inverse of the total dispersion matrix to obtain the target projection vector.

[0070] Optionally, the first training data is labeled as The second training data is labeled as For the first training data respectively With the second training data The centroid is calculated to obtain the corresponding centroid value, which is expressed as:

[0071] In the above formula, The centroid values ​​corresponding to the training data include the first centroid value and the second centroid value; This represents the total number of samples included in the corresponding training data; The training data currently being used for computation; This refers to a single sample data point in the training data.

[0072] Next, we calculate the deviation of all samples within each class of training data from their corresponding centroid values, and construct the intra-class scatter matrix, which is represented as follows:

[0073] In the above formula, The discreteness matrix corresponding to the training data includes a first discreteness matrix and a second discreteness matrix; This is the transpose operation of a matrix.

[0074] Next, the two discreteness matrices are linearly summed to obtain the total discreteness description matrix, which is represented as:

[0075] In the above formula, This is the matrix describing the total discreteness. This is the first discreteness matrix; This is the second discreteness matrix.

[0076] Next, based on constrained optimization theory, a projection direction is selected that minimizes the difference after projection of training data of the same class and maximizes the difference after projection of training data of different classes. The target projection vector is then calculated and expressed as:

[0077] In the above formula, The target projection vector; This is the first centroid value; This is the second centroid value.

[0078] The above embodiments can obtain a target projection vector that maximizes the differentiation between different types of data and minimizes the differences within the same type.

[0079] S102c, Based on the target projection vector, perform forward projection on the first training data and the second training data to obtain the target projection value.

[0080] In some embodiments of this application, S102c, which involves forward projection of the first training data and the second training data based on the target projection vector to obtain a target projection value, specifically includes the following steps: Calculating the centroid of the first training data to obtain a first centroid value; Calculating the centroid of the second training data to obtain a second centroid value; Performing a projection operation between the target projection vector and the first centroid value to obtain a first projection value; Performing a projection operation between the target projection vector and the second centroid value to obtain a second projection value; Performing a weighted calculation based on the number of first training data points, the number of second training data points, the first projection value, and the second projection value to obtain the target projection value.

[0081] Optionally, the first centroid value can be calculated in the same manner as above. and the second centroid value Let's redefine the linear projection function to perform the projection calculation. The function expression is:

[0082] In the above formula, For the projection result, This is the transpose of the target projection vector. Input data for projection calculations.

[0083] Next, the first center of gravity value and the second centroid value Substituting these values ​​into the linear projection function above, the projection calculation is completed, and the corresponding first projection value on the projection axis is obtained. With the second projection value .

[0084] Next, considering that in actual data collection scenarios, the number of training data samples corresponding to different states is often imbalanced, with the number of samples corresponding to the comfortable state often exceeding that of the less comfortable state, directly using a simple mean would reduce the accuracy of the classification boundary. Therefore, a sample size weighting mechanism is introduced to correct the projection value by combining the sample sizes of the two classes of training data, resulting in a target projection value suitable for imbalanced samples, expressed as:

[0085] In the above formula, For the target projection value; The number of training data points; This represents the number of training data points.

[0086] In the above embodiment, a linear projection function is first constructed, and the centroids of the two types of training data are substituted into the function to calculate their respective projection values. Then, considering the imbalance in the number of actual data samples (comfort samples are usually more numerous), the final projection value is calculated using a sample weighting method, thereby obtaining a target projection value that is more in line with the real collection scenario.

[0087] S102d, the target projection value is back-projected based on the target projection vector to obtain the first critical temperature.

[0088] Optionally, the formula for calculating back projection is:

[0089] In the above formula, This is the first critical temperature.

[0090] Understandably, in practical applications, the first critical temperature has two distinct uses. When the first critical temperature... The first critical temperature used to distinguish between a comfortable state and a slightly cold state. It can be used as the second critical temperature When the first critical temperature The first critical temperature used to distinguish between a comfortable state and a hot state. It can be used as the third critical temperature The third critical temperature The value is always greater than the second critical temperature. The numerical value. Furthermore, due to differences in physical characteristics, temperature perception abilities, and thermal comfort habits among different users, the corresponding second critical temperature varies for different users. With the third critical temperature The values ​​are not exactly the same.

[0091] The above S102a-S102d, by acquiring training data, determining the target projection vector, calculating the target projection value and back-projecting, can accurately determine the first critical temperature for classifying thermal sensation states, which not only improves the accuracy and stability of thermal sensation recognition, but also adapts to non-equilibrium sample scenarios.

[0092] In some embodiments of this application, such as Figure 4 As shown, S102 determines the target thermal perception recognition model of the target object based on the target temperature feature parameters of facial temperature data, including steps S102A-S102C, as follows: S102A: Obtain candidate thermal sensation recognition models and different candidate temperature feature parameters corresponding to different candidate thermal sensation recognition models.

[0093] Optionally, the candidate thermal sensing recognition model can be obtained through the steps S102a-S102d described above, which will not be repeated here.

[0094] For example, the candidate thermal perception recognition model generated for child users corresponds to a set of candidate temperature feature parameters specific to children; the candidate thermal perception recognition model generated for elderly users corresponds to a set of candidate temperature feature parameters specific to the elderly; and the candidate thermal perception recognition model generated for adult male users corresponds to a set of candidate temperature feature parameters specific to adult males. Of course, it is also possible to generate specific candidate temperature feature parameters for each candidate user, which is not limited here.

[0095] S102B calculates the distance between the target temperature feature parameter and different candidate temperature feature parameters of the facial temperature data to obtain the parameter distance.

[0096] Alternatively, the Manhattan distance method can be used to calculate the absolute difference between the target temperature feature parameter and the candidate temperature feature parameter in different dimensions, and then summing all the differences to obtain the final parameter distance. Alternatively, the cosine similarity method can be used to construct multi-dimensional feature vectors from the two types of parameters, calculate the cosine similarity of the vectors, and then subtract the similarity from 1 to obtain the parameter distance.

[0097] In some embodiments of this application, the target temperature feature parameters and candidate temperature feature parameters include a first category parameter, a second category parameter, and a third category parameter. The first category parameter represents the mean between the maximum and minimum values ​​in the facial temperature data, the second category parameter represents the volatility of the facial temperature data, and the third category parameter represents the difference between the maximum and minimum values ​​in the facial temperature data.

[0098] S102B calculates the distance between the target temperature feature parameters and different candidate temperature feature parameters of the facial temperature data to obtain the parameter distance. Specifically, this includes the following steps: calculating the distance between the first category of the target temperature feature parameters and the first category of the candidate temperature feature parameters to obtain a first distance; calculating the distance between the second category of the target temperature feature parameters and the second category of the candidate temperature feature parameters to obtain a second distance; calculating the distance between the third category of the target temperature feature parameters and the third category of the candidate temperature feature parameters to obtain a third distance; and finally, performing a weighted sum of the first, second, and third distances to obtain the parameter distance.

[0099] Optionally, the first category parameters, the second category parameters, and the third category parameters can be combined into a three-dimensional temperature feature vector, specifically expressed as follows:

[0100]

[0101]

[0102]

[0103] In the above formula, For the first category parameter; This represents the maximum value in the facial temperature data; This represents the minimum value in the facial temperature data; The volatility of facial temperature data; This represents the total number of facial temperature data. This is the i-th temperature data; This is a third-category parameter.

[0104] Optionally, the algorithm used to calculate the distance between feature parameters includes, but is not limited to, weighted Euclidean distance algorithm, Manhattan distance algorithm, Chebyshev distance algorithm and cosine similarity distance algorithm, and is not limited here.

[0105] S102C determines the candidate thermal sensing recognition model corresponding to the smallest parameter distance as the target thermal sensing recognition model.

[0106] The embodiments S102A-S102C described above can quickly match a personalized thermal perception recognition model for the current user, effectively adapt to the temperature perception differences of different groups of people, and improve the accuracy of thermal perception judgment.

[0107] S103, based on the target thermal sensation recognition model, the facial temperature data is identified and processed to obtain the target thermal sensation of the target object.

[0108] Among them, target thermal sensation represents the current thermal comfort state of the target object, which can be a slightly cold state, a comfortable state, or a slightly hot state.

[0109] Optionally, the target thermal perception recognition model has a built-in trained and optimized temperature judgment standard, which can quickly compare and classify facial temperature data. By judging whether the facial temperature data falls within the temperature range corresponding to the model, it can automatically determine whether the target object is currently in a state of being too cold, comfortable, or too hot.

[0110] In some embodiments of this application, thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The target thermal sensation recognition model includes a second critical temperature and a third critical temperature. The second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state.

[0111] like Figure 5 As shown, step S103, based on the target thermal sensing recognition model, processes facial temperature data to obtain the target thermal sensation of the target object. This includes steps S1031-S1034, as follows: S1031, calculate the average value of the facial temperature data to obtain the average facial temperature.

[0112] S1032, if the average facial temperature is less than the second critical temperature, then the cool state is determined as the target thermal sensation of the target object.

[0113] S1033, if the average facial temperature is greater than or equal to the second critical temperature and less than or equal to the third critical temperature, then the comfort state is determined as the target thermal sensation of the target object.

[0114] S1034, if the average temperature of the face is greater than the third critical temperature, then the pre-heated state is determined as the target thermal sensation of the target object.

[0115] For example, suppose the second critical temperature It is 35.8℃, the third critical temperature. The temperature was 36.5℃. The calculated average facial temperature was 35.5℃, which is lower than... The system determines that the target's current thermal sensation is on the cooler side. If the average facial temperature is 36.2℃, then... and Between these values, the target subject is judged to be in a comfortable state. If the average facial temperature is 36.8℃, which is higher than... The target object is determined to be in a slightly overheated state.

[0116] S104, controls the operation of the air conditioner based on target thermal sensing.

[0117] Optionally, the specific control logic is as follows:

[0118] In the above formula, In the temperature compensation mode, when the real-time facial temperature is below the boundary temperature between comfortable and slightly cold, the system determines that the person is currently feeling too cold and immediately initiates the temperature compensation operation, raising the air conditioning set temperature. And / or, lower the air conditioning fan speed. It can quickly improve the feeling of coldness among people. In cooling compensation mode, if the system determines that the person feels too hot, it will immediately activate the cooling compensation function and lower the air conditioner's set temperature. And / or, raise the air conditioning fan speed. It can quickly improve the feeling of heat among people. The command to maintain temperature and windshield indicates that the user is currently feeling comfortable and there is no need to adjust the air conditioning operating parameters; simply maintain the current operating state. This represents the magnitude of a single temperature adjustment, and can be set to 0.5℃. This represents the single windshield adjustment range, and can be set to 1 windshield increment.

[0119] In the above embodiments, by acquiring the facial temperature data of the target object in real time, and matching a target thermal sensation recognition model specific to the target object based on its target temperature feature parameters, the current target thermal sensation of the human body is identified based on the target thermal sensation recognition model and facial temperature data, and finally the operation of the air conditioner is directly controlled based on the target thermal sensation. This embodiment can adapt to the differences between individuals through dynamic matching models, and can directly capture the real thermal sensation of the human body, effectively overcoming the defect of "deviation between air conditioner output and actual human body temperature sensation" caused by the inability to perceive real-time comfort feedback of the human body in traditional control methods, thereby significantly improving the accuracy and comfort of temperature regulation.

[0120] In some embodiments of this application, thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The target thermal sensation recognition model includes a second critical temperature and a third critical temperature. The second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state.

[0121] After the target thermal sensing control air conditioner operates in step S104, the following steps are also included: When the parameter adjustment command of the target object is obtained, the parameter adjustment command is mapped to obtain the equivalent thermal sensing vote corresponding to the parameter adjustment command. The current facial temperature data of the target object and the equivalent thermal sensing vote corresponding to the parameter adjustment command are determined as the current training data, and the second critical temperature or the third critical temperature is updated based on the current training data.

[0122] Optionally, parameter adjustment commands issued by the target object are obtained. These commands include two categories: air conditioning temperature adjustment commands and air conditioning fan speed adjustment commands. Temperature adjustment commands include raising and lowering the set temperature, while fan speed adjustment commands include raising and lowering the fan speed. Following the preset mapping rules in Table 1, the parameter adjustment commands are standardized and mapped to convert the user's actions into equivalent thermal perception votes that can be used for model training, thus clarifying the user's current actual preference for hot or cold sensations.

[0123]

[0124] Table 1. Mapping Rules between Parameter Adjustment Commands and Equivalent Thermal Sensation Voting Next, the current facial temperature data of the target object is collected synchronously. This real-time facial temperature data is then bound to the equivalent thermal sensation vote obtained by mapping the parameter adjustment command, forming a new set of current training data. This current training data is then added to the model's training sample library. Following the training logic from S102a to S102d, the second and third critical temperatures in the thermal sensation recognition model are recalculated and corrected, completing the real-time update of the critical temperatures. The updated second and third critical temperatures replace the original values ​​in the model, ensuring that subsequent thermal sensation recognition and air conditioning control processes are executed based on the new critical temperatures, continuously improving the matching degree between body sensation recognition and air conditioning adjustment.

[0125] Optionally, the system can also be set to update periodically, updating all candidate thermal sensation recognition models in batches every fixed interval (e.g., 3 months) to adapt to the differences in user sensation caused by seasonal and physical changes, so that air conditioning control always meets the user's real-time comfort needs.

[0126] To facilitate better implementation of the air conditioning control method of this application, this application also provides an air conditioning control device based on the above-described air conditioning control method. The meanings of the terms used are the same as in the above-described air conditioning control method, and specific implementation details can be found in the descriptions of the method embodiments.

[0127] Please see Figure 6 , Figure 6 This is a schematic diagram of the structure of the air conditioning control device provided in the embodiments of this application, which may specifically include: The data acquisition module 601 is used to acquire the facial temperature data of the target object; The model determination module 602 is used to determine the target thermal perception recognition model of the target object based on the target temperature feature parameters of the facial temperature data. The thermal sensation recognition module 603 is used to recognize and process facial temperature data based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object; The operation control module 604 is used to control the operation of the air conditioner based on the target thermal sensation.

[0128] In the above embodiment, the data acquisition module 601 is used to acquire the facial temperature data of the target object in real time, the model determination module 602 is used to match the target thermal sensation recognition model of the target object based on its target temperature feature parameters, the thermal sensation recognition module 603 is used to identify the current target thermal sensation of the human body based on the target thermal sensation recognition model and facial temperature data, and the operation control module 604 is used to directly control the operation of the air conditioner based on the target thermal sensation. This embodiment can adapt to the differences between individuals through dynamic matching models and can directly capture the real thermal sensation of the human body, effectively overcoming the defect of "deviation between air conditioner output and actual cold and hot sensation of the human body" caused by the inability to perceive the real-time comfort feedback of the human body in traditional control methods, thereby significantly improving the accuracy and comfort of temperature regulation.

[0129] In some embodiments of this application, the target thermal sensation recognition model is obtained by the thermal sensation recognition module 603 from the candidate thermal sensation recognition models based on the target temperature feature parameters of the face temperature data. Thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The candidate thermal sensation recognition model includes a first critical temperature, which is used to distinguish between a comfortable state and a slightly cold state, or to distinguish between a comfortable state and a slightly hot state. The methods for determining the first critical temperature include: Acquire first training data and second training data; wherein, the first training data corresponds to a comfortable state, and the second training data corresponds to a slightly cold or slightly hot state; Based on the first training data and the second training data, a target projection vector is determined; wherein, the target projection vector is used to minimize the difference between training data of the same class after projection and to maximize the difference between training data of different classes after projection. The first training data and the second training data are forward-projected based on the target projection vector to obtain the target projection value; The target projection value is back-projected based on the target projection vector to obtain the first critical temperature.

[0130] In some embodiments of this application, the thermal sensing recognition module 603 determines the target projection vector based on first training data and second training data, including: The centroid is calculated on the first training data to obtain the first centroid value; The centroid is calculated on the second training data to obtain the second centroid value; The first training data is discrete based on the first centroid value to obtain the first discreteness matrix; The second training data is discrete based on the second centroid value to obtain the second discreteness matrix; The first and second discrete values ​​are linearly summed to obtain the total discrete value matrix; The target projection vector is obtained by multiplying the difference between the first and second centroid values ​​by the inverse of the total discreteness matrix.

[0131] In some embodiments of this application, the thermal sensing recognition module 603 performs forward projection of the first training data and the second training data based on the target projection vector to obtain the target projection value, including: The centroid is calculated on the first training data to obtain the first centroid value; The centroid is calculated on the second training data to obtain the second centroid value; The first projection value is obtained by projecting the target projection vector onto the first centroid value. The second projection value is obtained by projecting the target projection vector onto the second centroid value. The target projection value is obtained by weighting the number of first training data, the number of second training data, the first projection value, and the second projection value.

[0132] In some embodiments of this application, the data acquisition module 601 acquires facial temperature data of the target object, including: Acquire the first infrared image of the target object; Based on the first infrared image, determine the target temperature gradient image; A connected component search is performed on the target temperature gradient image to obtain several zero-element regions; where the zero-element regions represent the connected regions formed by pixels with a temperature gradient of 0. Coordinate mapping is performed on several zero-element regions to obtain several face mapping regions in the first infrared image; Calculate the first average temperature of several facial mapping regions, and determine the facial mapping regions whose first average temperature belongs to the human body temperature range as the target temperature region. Extract the temperature within the target temperature region to obtain the facial temperature data of the target object.

[0133] In some embodiments of this application, the model determination module 602 determines a target thermal perception recognition model for a target object based on target temperature feature parameters of facial temperature data, including: Obtain candidate thermal sensation recognition models, and the different candidate temperature feature parameters corresponding to different candidate thermal sensation recognition models; The distance between the target temperature feature parameter and different candidate temperature feature parameters of the facial temperature data is calculated to obtain the parameter distance; The candidate thermal sensing recognition model corresponding to the smallest parameter distance is determined as the target thermal sensing recognition model.

[0134] In some embodiments of this application, the target temperature feature parameters and candidate temperature feature parameters include a first category parameter, a second category parameter and a third category parameter. The first category parameter represents the mean between the maximum and minimum values ​​in the facial temperature data, the second category parameter represents the volatility of the facial temperature data, and the third category parameter represents the difference between the maximum and minimum values ​​in the facial temperature data. The model determination module 602 calculates the distance between the target temperature feature parameters of the facial temperature data and different candidate temperature feature parameters to obtain the parameter distance, including: The distance between the first category parameter of the target temperature feature parameter and the first category parameter of the candidate temperature feature parameter is calculated to obtain the first distance; The distance between the second category parameter of the target temperature feature parameter and the second category parameter of the candidate temperature feature parameter is calculated to obtain the second distance; The distance between the third category parameter of the target temperature feature parameter and the third category parameter of the candidate temperature feature parameter is calculated to obtain the third distance; The first, second, and third distances are weighted and summed to obtain the parameter distance.

[0135] In some embodiments of this application, thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state, and the target thermal sensation recognition model includes a second critical temperature and a third critical temperature; wherein, the second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state. The thermal sensation recognition module 603 processes facial temperature data based on a target thermal sensation recognition model to obtain the target thermal sensation of the target object, including: The average facial temperature was calculated by averaging the facial temperature data. If the average facial temperature is less than the second critical temperature, then the cool state is determined as the target thermal sensation of the target object. If the average facial temperature is greater than or equal to the second critical temperature and less than or equal to the third critical temperature, then the comfortable state is defined as the target thermal sensation of the target object. If the average facial temperature is greater than the third critical temperature, then the pre-heated state is determined as the target thermal sensation of the target object.

[0136] In some embodiments of this application, thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state, and the target thermal sensation recognition model includes a second critical temperature and a third critical temperature; wherein, the second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state. After the operation control module 604 controls the operation of the air conditioner based on the target thermal sensing, it also includes: When the parameter adjustment command of the target object is obtained, the parameter adjustment command is mapped to obtain the equivalent thermal sensation vote corresponding to the parameter adjustment command; The current facial temperature data of the target object and the equivalent thermal sensation vote corresponding to the parameter adjustment command are determined as the current training data, and the second or third critical temperature is updated based on the current training data.

[0137] In addition, this application also provides an air conditioner, such as Figure 7 As shown, it illustrates the structural diagram of the air conditioner involved in this application, specifically: The air conditioner may include components such as a processor 701 with one or more processing cores, a memory 702 with one or more computer-readable storage media, a power supply 703, and an input unit 704. Those skilled in the art will understand that... Figure 7 The air conditioner structure shown does not constitute a limitation on the air conditioner and may include more or fewer components than shown, or combine certain components, or have different component arrangements. Wherein: The processor 701 is the control center of the air conditioner. It connects to various parts of the air conditioner via various interfaces and lines. By running or executing software programs and / or modules stored in the memory 702, and by calling data stored in the memory 702, it performs various functions and processes data, thereby providing overall monitoring of the air conditioner. Optionally, the processor 701 may include one or more processing cores; preferably, the processor 701 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may not be integrated into the processor 701.

[0138] The memory 702 can be used to store software programs and modules. The processor 701 executes various functional applications and data processing by running the software programs and modules stored in the memory 702. The memory 702 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function, etc.; the data storage area may store data created based on the use of the air conditioner, etc. In addition, the memory 702 may include high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 702 may also include a memory controller to provide the processor 701 with access to the memory 702.

[0139] The air conditioner also includes a power supply 703 that supplies power to the various components. Preferably, the power supply 703 can be logically connected to the processor 701 through a power management system, thereby enabling functions such as charging, discharging, and power consumption management through the power management system. The power supply 703 may also include one or more DC or AC power supplies, recharging systems, power equipment debugging circuits, power converters or inverters, power status indicators, and other arbitrary components.

[0140] The air conditioner may also include an input unit 704, which can be used to receive input digital or character information, and generate keyboard, mouse, joystick, optical or trackball signal inputs related to object settings and function control.

[0141] Although not shown, the air conditioner may also include a display unit, etc., which will not be described in detail here. Specifically, in this embodiment, the processor 701 in the air conditioner will load the executable files corresponding to the processes of one or more application programs into the memory 702 according to the following instructions, and the processor 701 will run the application programs stored in the memory 702 to realize the steps in any of the air conditioning control methods provided in this application embodiment: acquiring the facial temperature data of the target object; determining the target thermal sensation recognition model of the target object based on the target temperature feature parameters of the facial temperature data; performing recognition processing on the facial temperature data based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object; and controlling the operation of the air conditioner based on the target thermal sensation.

[0142] The above embodiment acquires the facial temperature data of the target object in real time, matches a target thermal sensation recognition model for the target object based on its target temperature feature parameters, and then identifies the current target thermal sensation of the human body based on the target thermal sensation recognition model and facial temperature data. Finally, the air conditioner is directly controlled based on the target thermal sensation. This embodiment can adapt to the differences between individuals through dynamic matching models and can directly capture the real thermal sensation of the human body. It effectively overcomes the defect of traditional control methods that cause "deviation between air conditioner output and actual human body temperature" due to the inability to perceive real-time comfort feedback of the human body, thereby significantly improving the accuracy and comfort of temperature regulation.

[0143] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0144] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be performed by instructions, or by instructions controlling related hardware. These instructions can be stored in a computer-readable storage medium and loaded and executed by a processor.

[0145] Therefore, this application provides a computer-readable storage medium storing a computer program that can be loaded by a processor to execute the steps of any of the air conditioning control methods provided in this application.

[0146] For details on the implementation of each of the above operations, please refer to the previous examples, which will not be repeated here.

[0147] The computer-readable storage medium may include: read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0148] Since the instructions stored in the computer-readable storage medium can execute the steps of any of the air conditioning control methods provided in this application, the beneficial effects that any of the air conditioning control methods provided in this application can achieve can be realized, as detailed in the preceding embodiments, and will not be repeated here.

[0149] The above provides a detailed description of an air conditioning control method, apparatus, air conditioner, and computer-readable storage medium provided in this application. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, those skilled in the art will recognize that there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. An air conditioning control method, characterized in that, The method includes: Obtain the facial temperature data of the target object; Based on the target temperature feature parameters of the facial temperature data, a target thermal perception recognition model for the target object is determined. The facial temperature data is processed based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object. The operation of the air conditioner is controlled based on the target thermal sensation.

2. The air conditioning control method according to claim 1, characterized in that, The target thermal sensation recognition model is obtained by filtering candidate thermal sensation recognition models based on the target temperature feature parameters of the facial temperature data. Thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The candidate thermal sensation recognition model includes a first critical temperature, which is used to distinguish between the comfortable state and the slightly cold state, or to distinguish between the comfortable state and the slightly hot state. The method for determining the first critical temperature includes: Acquire first training data and second training data; wherein, the first training data corresponds to the comfort state, and the second training data corresponds to the slightly cold state or the slightly hot state; Based on the first training data and the second training data, a target projection vector is determined; wherein, the target projection vector is used to minimize the difference between training data of the same class after projection and to maximize the difference between training data of different classes after projection. Based on the target projection vector, the first training data and the second training data are projected forward to obtain the target projection value; The target projection value is back-projected based on the target projection vector to obtain the first critical temperature.

3. The air conditioning control method according to claim 2, characterized in that, Determining the target projection vector based on the first training data and the second training data includes: The centroid is calculated on the first training data to obtain the first centroid value; The second training data is used to calculate the centroid to obtain the second centroid value; The first training data is subjected to discreteness calculation based on the first centroid value to obtain the first discreteness matrix; The second training data is discrete based on the second centroid value to obtain the second discreteness matrix; The first and second discrete values ​​are linearly summed to obtain the total discrete value matrix; The difference between the first centroid value and the second centroid value is multiplied by the inverse of the total discreteness matrix to obtain the target projection vector.

4. The air conditioning control method according to claim 2, characterized in that, The step of forward projecting the first training data and the second training data based on the target projection vector to obtain the target projection value includes: The centroid is calculated on the first training data to obtain the first centroid value; The second training data is used to calculate the centroid to obtain the second centroid value; The target projection vector is projected onto the first centroid value to obtain the first projection value. The target projection vector is projected onto the second centroid value to obtain the second projection value; The target projection value is obtained by performing a weighted operation based on the number of the first training data, the number of the second training data, the first projection value, and the second projection value.

5. The air conditioning control method according to claim 1, characterized in that, The acquisition of the facial temperature data of the target object includes: Acquire the first infrared image of the target object; Based on the first infrared image, a target temperature gradient image is determined; A connected component search is performed on the target temperature gradient image to obtain several zero-element regions; wherein, the zero-element regions represent connected regions formed by pixels with a temperature gradient of 0. Coordinate mapping is performed on the aforementioned zero-element regions to obtain several face mapping regions in the first infrared image; Calculate the first average temperature of the plurality of facial mapping regions, and determine the facial mapping regions whose first average temperature belongs to the human body temperature range as the target temperature region. The temperature within the target temperature region is extracted to obtain the facial temperature data of the target object.

6. The air conditioning control method according to claim 1, characterized in that, The determination of the target thermal perception recognition model for the target object based on the target temperature feature parameters of the facial temperature data includes: Obtain candidate thermal sensation recognition models, and the different candidate temperature feature parameters corresponding to different candidate thermal sensation recognition models; The distance between the target temperature feature parameter of the facial temperature data and the different candidate temperature feature parameters is calculated to obtain the parameter distance; The candidate thermal sensing recognition model corresponding to the smallest parameter distance is determined as the target thermal sensing recognition model.

7. The air conditioning control method according to claim 6, characterized in that, The target temperature feature parameters and the candidate temperature feature parameters include a first category parameter, a second category parameter, and a third category parameter. The first category parameter represents the mean between the maximum and minimum values ​​in the facial temperature data, the second category parameter represents the volatility of the facial temperature data, and the third category parameter represents the difference between the maximum and minimum values ​​in the facial temperature data. The step of calculating the distance between the target temperature feature parameter of the facial temperature data and the different candidate temperature feature parameters to obtain the parameter distance includes: The distance between the first category parameter of the target temperature feature parameter and the first category parameter of the candidate temperature feature parameter is calculated to obtain a first distance; The distance between the second category parameter of the target temperature feature parameter and the second category parameter of the candidate temperature feature parameter is calculated to obtain the second distance; The distance between the third category parameter of the target temperature feature parameter and the third category parameter of the candidate temperature feature parameter is calculated to obtain the third distance; The first distance, the second distance, and the third distance are weighted and summed to obtain the parameter distance.

8. The air conditioning control method according to claim 1, characterized in that, Thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The target thermal sensation recognition model includes a second critical temperature and a third critical temperature. The second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state. The step of processing the facial temperature data based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object includes: The average facial temperature is calculated by averaging the facial temperature data. If the average facial temperature is less than the second critical temperature, then the cool state is determined as the target thermal sensation of the target object. If the average facial temperature is greater than or equal to the second critical temperature and less than or equal to the third critical temperature, then the comfortable state is determined as the target thermal sensation of the target object. If the average facial temperature is greater than the third critical temperature, then the preheated state is determined as the target thermal sensation of the target object.

9. The air conditioning control method according to claim 1, characterized in that, Thermal sensation includes a slightly cold state, a comfortable state, and a slightly hot state. The target thermal sensation recognition model includes a second critical temperature and a third critical temperature. The second critical temperature is used to distinguish between the comfortable state and the slightly cold state, and the third critical temperature is used to distinguish between the comfortable state and the slightly hot state. After controlling the operation of the air conditioner based on the target thermal sensation, the method further includes: When the parameter adjustment instruction of the target object is obtained, the parameter adjustment instruction is mapped to obtain the equivalent thermal sensation vote corresponding to the parameter adjustment instruction; The current facial temperature data of the target object and the equivalent thermal sensation vote corresponding to the parameter adjustment command are determined as the current training data, and the second critical temperature or the third critical temperature is updated based on the current training data.

10. An air conditioning control device, characterized in that, The air conditioning control device includes: The data acquisition module is used to acquire the facial temperature data of the target object; The model determination module is used to determine the target thermal perception recognition model of the target object based on the target temperature feature parameters of the facial temperature data. A thermal sensation recognition module is used to recognize and process the facial temperature data based on the target thermal sensation recognition model to obtain the target thermal sensation of the target object; The operation control module is used to control the operation of the air conditioner based on the target thermal sensation.

11. A computer-readable storage medium, characterized in that, The device stores a computer program that, when executed by a processor, causes the processor to perform the steps of the method as described in any one of claims 1 to 9.

12. An air conditioner, characterized in that, It includes a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method as described in any one of claims 1 to 9.

Images