Temperature and humidity calibration method and device based on deep learning, equipment and storage medium

Through deep learning-based methods, the temperature and humidity calibration model is constructed and trained, and the accuracy problem caused by the components heating during measurement by traditional temperature and humidity collectors is solved, which improves calibration accuracy, enhances the accuracy of temperature and humidity detection, and improves the accuracy of temperature and humidity detection. Performance and user comfort of indoor environment control system.

CN118960813APending Publication Date: 2024-11-15GUANGDONG PHNIX ECO ENERGY SOLUTION
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202410872553.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-01
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

During measurement, the traditional temperature and humidity collector is too high due to the heating of internal components, which affects the accuracy of humidity measurement. Failure to fully consider factors such as heat dissipation and installation position, resulting in poor calibration accuracy and reducing the accuracy of temperature and humidity detection value. .

Method used

Using a deep learning-based method, the internal temperature and humidity array, uncorrected ambient temperature and humidity array and calibration temperature and humidity array of the temperature and humidity collector under various operating conditions is constructed and trained to adapt to the temperature and humidity calibration model. Different heating states and heat dissipation conditions improve calibration accuracy.

Benefits of technology

It significantly improves the accuracy of temperature and humidity prediction of temperature and humidity calibration model, enhances the accuracy of temperature and humidity detection values, improves the performance and user comfort of indoor environment control systems, and reduces unnecessary energy consumption.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118960813A_ABST
    Figure CN118960813A_ABST
Patent Text Reader

Abstract

The invention relates to the field of measurement data calibration, discloses a temperature and humidity calibration method, device and equipment based on deep learning and a storage medium, and is used for solving the technical problem of poor temperature and humidity calibration precision of a temperature and humidity collector. The temperature and humidity calibration method based on deep learning comprises the following steps: acquiring an internal temperature and humidity array, an uncorrected environment temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration correlation parameter corresponding to each working condition of a temperature and humidity collector; constructing an initial temperature and humidity calibration model according to the internal temperature and humidity array, the uncorrected environment temperature and humidity array and the at least one target calibration correlation parameter; training the initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the internal temperature and humidity array, the uncorrected environment temperature and humidity array and the at least one target calibration correlation parameter to obtain a target temperature and humidity calibration model; and performing temperature and humidity detection value calibration through the target temperature and humidity calibration model.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of measurement data calibration, and in particular to a temperature and humidity calibration method, device, equipment and storage medium based on deep learning. Background Art

[0002] Temperature and humidity collectors are an environmental monitoring device widely used in various scenarios. The accuracy of the temperature and humidity they measure is crucial to indoor environmental control. For example, indoor temperature control systems such as central air conditioners use the measured temperature and humidity as feedback signals to start cooling / heating. Inaccurate measurement data will result in additional energy consumption and poor user comfort.

[0003] However, the heat generated by the internal components of the temperature and humidity collector when powered on will affect the accuracy of the measured data. For example, the heat generated by the components will cause the measured temperature to be too high, and the humidity changes with temperature. If the measured temperature is too high, the measured humidity will also be inaccurate.

[0004] The traditional scheme uses different fitting methods to fit the temperature and humidity to achieve temperature and humidity calibration. It does not take into account the impact of different operating conditions of the components of the temperature and humidity collector on the heating conditions, nor does it consider the heat dissipation conditions and installation position of the temperature and humidity collector. It is impossible to establish an accurate functional relationship through mathematical deduction. The traditional temperature and humidity calibration is inaccurate, which reduces the accuracy of the temperature and humidity detection values. Summary of the invention

[0005] The present invention provides a temperature and humidity calibration method, device, equipment and storage medium based on deep learning, which is used for the traditional temperature and humidity calibration through functional relationship fitting, which cannot adapt to different internal heating states and the influence of the heat dissipation conditions and installation position of the temperature and humidity collector, and has poor calibration accuracy, thereby reducing the accuracy of the temperature and humidity detection values.

[0006] The first aspect of the present invention provides a temperature and humidity calibration method based on deep learning, including: obtaining the internal temperature and humidity array, uncorrected ambient temperature and humidity array, calibrated temperature and humidity array and at least one target calibration associated parameter corresponding to each working condition of the temperature and humidity collector; constructing an initial temperature and humidity calibration model according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and the at least one target calibration associated parameter; training the initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and the at least one target calibration associated parameter to obtain a target temperature and humidity calibration model; and calibrating the temperature and humidity detection values ​​through the target temperature and humidity calibration model.

[0007] The second aspect of the present invention provides a temperature and humidity calibration device based on deep learning, including: an acquisition module, used to obtain the internal temperature and humidity array, uncorrected ambient temperature and humidity array, calibrated temperature and humidity array and at least one target calibration associated parameter corresponding to each working condition of the temperature and humidity collector; a construction module, used to construct an initial temperature and humidity calibration model according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and the at least one target calibration associated parameter; a training module, used to train the initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and the at least one target calibration associated parameter to obtain a target temperature and humidity calibration model; a calibration module, used to calibrate the temperature and humidity detection values ​​through the target temperature and humidity calibration model.

[0008] The third aspect of the present invention provides a temperature and humidity calibration device based on deep learning, comprising: a memory and at least one processor, wherein instructions are stored in the memory; the at least one processor calls the instructions in the memory so that the temperature and humidity calibration device based on deep learning performs the above-mentioned temperature and humidity calibration method based on deep learning.

[0009] A fourth aspect of the present invention provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is run on a computer, it enables the computer to execute the above-mentioned temperature and humidity calibration method based on deep learning.

[0010] In the technical solution provided in this embodiment, a temperature and humidity calibration model is constructed and trained through an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration associated parameter, fully considering the impact of heating of internal components of the temperature collector on the ambient temperature and humidity collection, and combining the corresponding target calibration associated parameters to improve the accuracy of temperature and humidity prediction by the target temperature and humidity calibration model. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 Schematic diagram of an embodiment of a temperature and humidity calibration method based on deep learning in an embodiment of the present invention;

[0012] Figure 2 Schematic diagram of another embodiment of a temperature and humidity calibration method based on deep learning in an embodiment of the present invention;

[0013] Figure 3 Schematic diagram of an embodiment of a temperature and humidity calibration device based on deep learning in an embodiment of the present invention;

[0014] Figure 4 Schematic diagram of another embodiment of a temperature and humidity calibration device based on deep learning in an embodiment of the present invention;

[0015] Figure 5 This is a schematic diagram of an embodiment of a temperature and humidity calibration device based on deep learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] The present invention provides a temperature and humidity calibration method, device, equipment and storage medium based on deep learning, which are used to solve the technical problem of poor temperature and humidity calibration accuracy of a temperature and humidity collector.

[0017] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0018] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 , an embodiment of a temperature and humidity calibration method based on deep learning in an embodiment of the present invention includes:

[0019] 101. Obtain an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array, and at least one target calibration-related parameter corresponding to each working condition of a temperature and humidity collector.

[0020] It is understandable that the execution subject of the present invention can be a temperature and humidity calibration device based on deep learning, and can also be a temperature and humidity collector control system, a temperature and humidity collector control terminal or a server, which is not limited here. The embodiment of the present invention is described by taking the temperature and humidity collector control system as the execution subject as an example.

[0021] The temperature and humidity collector in the temperature and humidity collector control system of this embodiment can be widely used in indoor environment control systems such as central air conditioners and heat pump air conditioners. If the temperature and humidity collected by the temperature and humidity collector are not corrected, this temperature difference will have an adverse effect on the performance of the indoor environment control system. For example, an indoor environment control system that uses the measured humidity as a feedback signal will perform dehumidification and humidification operations until the measured humidity reaches the humidity set point. If the actual humidity of the building area is lower than the measured humidity, then controlling the measured humidity to reach the set point will make the actual humidity of the building area lower than the set point, thereby reducing user comfort and increasing unnecessary energy consumption. The same is true for temperature.

[0022] In this embodiment, the temperature and humidity collector control system includes a temperature and humidity collector, which includes a shell. The shell includes one or more heating devices, such as a CPU, an electronic display and its backlight source, relay electronics and other components. The one or more heating devices make the temperature inside the shell exceed the actual temperature outside the shell, such as the actual indoor temperature.

[0023] The temperature and humidity collector includes a first temperature and humidity sensor and a second temperature and humidity sensor. The first temperature and humidity sensor is installed inside the shell to collect the internal temperature and humidity of the temperature and humidity collector and construct an internal temperature and humidity array, such as being installed on a PCB board that encapsulates the CPU; the second temperature and humidity sensor is installed on the shell, that is, arranged at the edge of the temperature and humidity collector, to collect the ambient temperature and humidity outside the shell and construct an uncorrected ambient temperature and humidity array.

[0024] In this embodiment, the calibrated temperature and humidity array is used to indicate a data set consisting of actual temperature and humidity. The temperature and humidity of the temperature and humidity collector under various working conditions can be detected by calibrated experimental equipment. For example, when the temperature and humidity collector is powered on, it is placed in an incubator to collect data under different constant temperature and humidity conditions to construct a calibrated temperature and humidity array.

[0025] In this embodiment, the target calibration associated parameters are used to indicate the influencing factors in the temperature and humidity calibration process, and may include one or more temperature calibration associated parameters, wherein the temperature calibration associated parameters are used to indicate the influencing factors on the temperature calibration process, and the humidity calibration associated parameters are used to indicate the influencing factors on the humidity calibration process.

[0026] Among them, the temperature calibration associated parameters may include the working state parameters of each heating device, the number of heating components, the shell material parameters of the temperature and humidity collector, heat dissipation associated parameters, the relative distance of the sensor, whether it is suitable for wall installation, etc. The corresponding temperature calibration associated parameters can be selected as the input variables of the model according to actual conditions. This embodiment does not impose any specific restrictions on the number of temperature calibration associated parameters.

[0027] As mentioned above, the working status parameters of each heating device are used to indicate the operating status of each heating device, such as whether the electronic display is activated, the brightness level of the backlight source, how many heating devices are activated, etc. For example, the backlight source is configured to illuminate at multiple different brightness levels. Due to the different operating states of the heating devices, the power consumption is not constant and the heat generated is also different.

[0028] As mentioned above, the shell material parameters can be represented by numerical values ​​such as thermal conductivity and heat capacity, and the heat dissipation related parameters can be represented by numerical values ​​such as the number and size of heat sinks and the number of heat dissipation holes.

[0029] As mentioned above, the relative distance of sensors is the relative position of the sensors between the first temperature and humidity sensor and the second temperature and humidity sensor. Different temperature and humidity sensors have different design layouts, and the first temperature and humidity sensor is usually installed near the heating device. The greater the relative distance of the sensors, the greater the influence of the heating of the internal heating device on the temperature and humidity detection value of the second temperature and humidity sensor, and the smaller the relative distance of the sensors, the opposite is true.

[0030] It should be further explained that the target calibration associated parameters may also include one or more humidity calibration associated parameters, wherein the humidity calibration associated parameters may include real-time wind speed, whether it is close to a ventilated area, etc. In this embodiment, the influencing factors of the temperature calibration process and the humidity calibration process are digitized and further normalized to obtain at least one temperature calibration associated parameter and at least one humidity calibration associated parameter as model input variables, and the uncorrected temperature and humidity data collected by the temperature and humidity collector at the same time are bound to the internal temperature and humidity, and the calibrated temperature and humidity data collected by professional instruments, and the constructed training data set is used to construct and train the temperature and humidity calibration model to fit various factors of the temperature and humidity monitoring process, thereby improving the accuracy of the target temperature and humidity calibration model in predicting temperature and humidity.

[0031] 102. Construct an initial temperature and humidity calibration model according to an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter.

[0032] Specifically, the number of network layers of the initial temperature and humidity calibration model and the number of nodes corresponding to each network layer are determined according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and at least one target calibration associated parameter; the weight matrix between each network layer is initialized, and the ReLU function is determined as the activation function of the initial temperature and humidity calibration model to obtain the initial temperature and humidity calibration model.

[0033] The initial temperature and humidity calibration model of this embodiment is a deep learning model, which includes an input layer, a hidden layer and an output layer. The number of neuron nodes in the input layer and the output layer can be determined based on the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and at least one target calibration associated parameter, and the number of neuron nodes in the hidden layer can be determined based on actual conditions.

[0034] The model parameters of the above-mentioned initial temperature and humidity calibration model include the weight matrix between the front and rear two network layer structures, the activation function, the learning rate, the loss threshold and the iteration number threshold. The model parameters of the initial temperature and humidity calibration model are initialized to obtain the initial temperature and humidity calibration model.

[0035] 103. Train an initial temperature and humidity calibration model according to a calibrated temperature and humidity array, an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter to obtain a target temperature and humidity calibration model.

[0036] In a feasible implementation, an internal temperature array, an uncorrected ambient temperature array and at least one target calibration-related parameter are input into an initial temperature calibration network to obtain a temperature prediction result; the temperature calibration network is back-propagated according to the calibrated temperature array and the temperature prediction result in the calibrated temperature and humidity array to obtain a target temperature calibration network; prediction is performed through the target temperature calibration network to obtain a corrected target ambient temperature array; the target ambient temperature array, the internal humidity array and the uncorrected ambient humidity array are input into an initial humidity calibration network to obtain a humidity prediction result; the humidity calibration network is back-propagated according to the calibrated humidity array and the humidity prediction result to obtain a target humidity calibration network; the target temperature calibration network and the target humidity calibration network are determined as a target temperature and humidity calibration model.

[0037] In this embodiment, the initial humidity calibration network is trained and parameter updated through the target ambient temperature array, the internal humidity array, the uncorrected ambient humidity array, and the calibrated humidity array, wherein the target ambient temperature array is used as the humidity calibration associated parameter in the humidity calibration process, and each ambient temperature corresponds to a different saturated humidity. Model training is performed through the calibrated ambient temperature to improve the accuracy of humidity calibration.

[0038] Specifically, the temperature calibration network is back-propagated according to the calibrated temperature array and the temperature prediction results in the calibrated temperature and humidity array to obtain a target temperature calibration network, including: obtaining a temperature prediction loss value according to the calibrated temperature array, the temperature prediction results and a preset first loss function; if the temperature prediction loss value is greater than a preset first loss threshold, the initial temperature calibration network is iteratively updated through the first loss function and a preset learning rate until the first loss function converges to obtain a target temperature and humidity calibration model.

[0039] Specifically, the humidity calibration network is back-propagated according to the calibrated humidity array and the humidity prediction results to obtain the target humidity calibration network, including: obtaining the humidity prediction loss value according to the calibrated humidity array, the humidity prediction results and the preset second loss function; if the humidity prediction loss value is greater than the preset second loss threshold, the initial humidity calibration network is iteratively updated through the second loss function and the preset learning rate until the second loss function converges to obtain the target temperature and humidity calibration model.

[0040] 104. Calibrate the temperature and humidity detection values ​​through the target temperature and humidity calibration model.

[0041] In a feasible implementation, internal temperature and humidity detection values, ambient temperature and humidity detection values, and at least one target calibration-related parameter are obtained; the internal temperature detection value, ambient temperature detection value, and at least one target calibration-related parameter are input into the temperature calibration network of the target temperature and humidity calibration model to obtain the calibrated target ambient temperature; the target ambient temperature, internal humidity detection value, and ambient humidity detection value are input into the humidity calibration network of the target temperature and humidity calibration model to obtain the calibrated target ambient humidity. This embodiment performs temperature calibration on the internal temperature detection value, ambient temperature detection value, and at least one target calibration-related parameter through a deep learning network, adapts to the temperature correction requirements of the temperature and humidity collector under various working conditions, improves the accuracy of ambient temperature acquisition, and uses the calibrated target ambient temperature as the input variable for humidity calibration, and combines the internal humidity detection value and the ambient humidity detection value to improve the accuracy of humidity calibration network calibration, thereby improving the accuracy of ambient humidity acquisition.

[0042] In the technical solution provided in this embodiment, a temperature and humidity calibration model is constructed and trained through an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration associated parameter, fully considering the impact of heating of internal components of the temperature collector on the ambient temperature and humidity collection, and combining the corresponding target calibration associated parameters to improve the accuracy of temperature and humidity prediction by the target temperature and humidity calibration model.

[0043] See also Figure 2 Another embodiment of the temperature and humidity calibration method based on deep learning in the embodiment of the present invention includes:

[0044] 201. Obtain an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array, and at least one target calibration-related parameter corresponding to each working condition of a temperature and humidity collector.

[0045] 202. Construct an initial temperature and humidity calibration model according to an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter.

[0046] The initial temperature and humidity calibration model of this embodiment includes an initial temperature calibration network and an initial humidity calibration network, the internal temperature and humidity array includes an internal temperature array and an internal humidity array, the uncorrected ambient temperature and humidity array includes an uncorrected ambient temperature array and an uncorrected ambient humidity array, and the at least two target calibration associated parameters include at least one temperature calibration associated parameter and at least one humidity calibration associated parameter.

[0047] Specifically, the number of network layers of the initial temperature calibration network and the number of nodes corresponding to each network layer are determined according to the internal temperature array, the uncorrected ambient temperature array and at least one temperature calibration associated parameter. For example, the input variables include the internal detection temperature, the uncalibrated ambient detection temperature, and two target calibration associated parameters. The two target calibration associated parameters are whether the temperature fits the wall and the shell material parameters. The input neuron nodes of the initial temperature calibration network can be set to 4, and the output layer of the initial temperature calibration network includes a node to output the predicted temperature of the initial temperature calibration network, that is, the calibrated target ambient temperature.

[0048] Specifically, the number of network layers of the initial humidity calibration network and the number of nodes corresponding to each network layer are determined according to the target ambient temperature array, the internal humidity array, the uncorrected ambient humidity array and at least one humidity calibration associated parameter. For example, the input variables include the predicted temperature of the initial temperature calibration network, the internal detected humidity, the uncorrected ambient detected humidity and a humidity calibration associated parameter. The humidity calibration associated parameter is such as the current wind speed. The input neuron nodes of the initial humidity calibration network can be set to 4. The output layer of the initial humidity calibration network includes a node to output the predicted humidity of the initial humidity calibration network, that is, the calibrated target ambient humidity.

[0049] 203. Train an initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the internal temperature and humidity array, the uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter to obtain a target temperature and humidity calibration model.

[0050] In a feasible implementation, the internal temperature array, the uncorrected ambient temperature array and at least one target calibration-related parameter are input into the initial temperature calibration network to obtain a temperature prediction result; the internal humidity array, the uncorrected ambient humidity array, the temperature prediction result and at least one humidity calibration-related parameter are input into the initial humidity calibration network to obtain a humidity prediction result; the initial temperature and humidity calibration model is updated according to the calibrated temperature and humidity array, the temperature prediction result and the humidity prediction result to obtain a target temperature and humidity calibration model.

[0051] Specifically, the initial temperature and humidity calibration model is back-propagated according to the calibrated temperature and humidity array, the temperature prediction results and the humidity prediction results to obtain the target temperature and humidity calibration model, including: obtaining the temperature prediction loss value and the humidity prediction loss value according to the calibrated temperature and humidity array, the temperature prediction results, the humidity prediction results and the preset loss function; if the temperature prediction loss value and / or the humidity prediction loss value is greater than the preset loss threshold, the initial temperature and humidity calibration model is iteratively updated through the loss function and the preset learning rate until the temperature prediction loss value and the humidity prediction loss value are less than or equal to the loss threshold, so as to obtain the target temperature and humidity calibration model.

[0052] 204. Obtain internal temperature and humidity detection values, ambient temperature and humidity detection values, at least one temperature calibration-related parameter, and at least one humidity calibration-related parameter.

[0053] Specifically, the pre-trained target temperature and humidity calibration model is deployed to the temperature and humidity collector, installed at the corresponding position in the building area, and the corresponding temperature calibration-related parameters and humidity calibration-related parameters in the building area are configured, such as input shell material parameters, heat dissipation-related parameters, sensor relative distance, whether it is installed close to the wall, etc., so as to subsequently calibrate the real-time temperature detection value and humidity detection value, and collect other real-time calibration-related parameters, such as taking internal temperature and humidity detection values, ambient temperature and humidity detection values, real-time wind speed, and working status parameters of each heating device as input variables.

[0054] 205. Input the internal temperature detection value, the ambient temperature detection value, and at least one temperature calibration associated parameter into a temperature calibration network of a target temperature and humidity calibration model to obtain a calibrated target ambient temperature.

[0055] The internal temperature detection value, the ambient temperature detection value and at least one temperature calibration associated parameter are input into the first input layer of the temperature calibration network; the operation is performed according to the first weight matrix between the first input layer and the first hidden layer, and the operation is input into the first activation layer of the temperature calibration network; the operation is performed according to the second weight matrix between the first activation layer and the second hidden layer, and the operation is input into the first output layer of the temperature calibration network; and the target ambient temperature is output through the first output layer.

[0056] Optionally, a temperature calibration parameter matrix is ​​constructed by combining the internal temperature detection value, the ambient temperature detection value, the working state parameters of each heating device and the shell material parameters, and the matrix is ​​input into the first input layer. To prevent the network learning from falling into a local area, a first bias matrix is ​​added, and the temperature calibration parameter matrix, the first weight matrix and the first bias matrix are subjected to matrix operations and input into the first activation layer of the temperature calibration network. The second weight matrix and the second bias matrix are operated between the first activation layer and the second hidden layer, and the target ambient temperature is output through the first output layer.

[0057] 206. Input the target ambient temperature, the internal humidity detection value, the ambient humidity detection value, and at least one humidity calibration associated parameter into a humidity calibration network of a target temperature and humidity calibration model to obtain a calibrated target ambient humidity.

[0058] The internal humidity detection value, the ambient humidity detection value, at least one humidity calibration associated parameter, and the target ambient temperature output by the first output layer are input into the second input layer of the humidity calibration network; operations are performed according to the third weight matrix between the second input layer and the third hidden layer, and the results are input into the second activation layer of the humidity calibration network; operations are performed according to the fourth weight matrix between the second activation layer and the fourth hidden layer, and the results are input into the second output layer of the humidity calibration network; and the target ambient humidity is output through the second output layer.

[0059] Optionally, the internal temperature detection value, the ambient temperature detection value, the real-time wind speed and the target ambient humidity are used to construct a humidity calibration parameter matrix and input it into the first input layer. In order to prevent the network learning from falling into a local area, a third bias matrix is ​​added, and the humidity calibration parameter matrix, the third weight matrix and the third bias matrix are subjected to matrix operations and input into the second activation layer of the temperature calibration network. The third weight matrix and the fourth bias matrix are operated between the second activation layer and the fourth hidden layer, and the target ambient humidity is output through the first output layer.

[0060] In the technical solution provided in this embodiment, a temperature and humidity calibration model is constructed and trained through an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration associated parameter, fully considering the impact of heating of internal components of the temperature collector on the ambient temperature and humidity collection, and combining the corresponding target calibration associated parameters to improve the accuracy of temperature and humidity prediction by the target temperature and humidity calibration model.

[0061] The above describes the temperature and humidity calibration method based on deep learning in the embodiment of the present invention. The following describes the temperature and humidity calibration device based on deep learning in the embodiment of the present invention. Figure 3 , an embodiment of a temperature and humidity calibration device based on deep learning in an embodiment of the present invention includes:

[0062] An acquisition module 301 is used to acquire an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration associated parameter corresponding to each working condition of the temperature and humidity collector;

[0063] A construction module 302 is used to construct an initial temperature and humidity calibration model according to an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter;

[0064] A training module 303 is used to train an initial temperature and humidity calibration model according to a calibrated temperature and humidity array, an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter to obtain a target temperature and humidity calibration model;

[0065] The calibration module 304 is used to calibrate the temperature and humidity detection values ​​using a target temperature and humidity calibration model.

[0066] In the technical solution provided in this embodiment, a temperature and humidity calibration model is constructed and trained through an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration associated parameter, fully considering the impact of heating of internal components of the temperature collector on the ambient temperature and humidity collection, and combining the corresponding target calibration associated parameters to improve the accuracy of temperature and humidity prediction by the target temperature and humidity calibration model.

[0067] See also Figure 4 , another embodiment of the temperature and humidity calibration device based on deep learning in the embodiment of the present invention includes:

[0068] An acquisition module 301 is used to acquire an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration associated parameter corresponding to each working condition of the temperature and humidity collector;

[0069] A construction module 302 is used to construct an initial temperature and humidity calibration model according to an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter;

[0070] A training module 303 is used to train an initial temperature and humidity calibration model according to a calibrated temperature and humidity array, an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, and at least one target calibration associated parameter to obtain a target temperature and humidity calibration model;

[0071] The calibration module 304 is used to calibrate the temperature and humidity detection values ​​using a target temperature and humidity calibration model.

[0072] Optionally, the calibration module 304 includes:

[0073] An acquisition unit 3041 is used to acquire an internal temperature and humidity detection value, an ambient temperature and humidity detection value, at least one temperature calibration-related parameter, and at least one humidity calibration-related parameter;

[0074] A temperature calibration unit 3042 is used to input the internal temperature detection value, the ambient temperature detection value and at least one temperature calibration associated parameter into a temperature calibration network of a target temperature and humidity calibration model to obtain a calibrated target ambient temperature;

[0075] The humidity calibration unit 3043 is used to input the target ambient temperature, the internal humidity detection value, the ambient humidity detection value and at least one humidity calibration associated parameter into the humidity calibration network of the target temperature and humidity calibration model to obtain the calibrated target ambient humidity.

[0076] Optionally, the temperature calibration unit 3042 is specifically used to: input the internal temperature detection value, the ambient temperature detection value and at least one temperature calibration associated parameter into a first input layer of the temperature calibration network;

[0077] Performing operations according to a first weight matrix between a first input layer and a first hidden layer and inputting into a first activation layer of a temperature calibration network;

[0078] Performing an operation based on a second weight matrix between the first activation layer and the second hidden layer and inputting the result into a first output layer of the temperature calibration network;

[0079] The target ambient temperature is output through the first output layer.

[0080] Optionally, the humidity calibration unit 3043 is specifically used to: input the internal humidity detection value, the ambient humidity detection value, at least one humidity calibration associated parameter, and the target ambient temperature output by the first output layer into the second input layer of the humidity calibration network;

[0081] Performing operations according to a third weight matrix between the second input layer and the third hidden layer and inputting into a second activation layer of the humidity calibration network;

[0082] Performing operations according to a fourth weight matrix between the second activation layer and the fourth hidden layer, and inputting into a second output layer of the humidity calibration network;

[0083] The target ambient humidity is output through the second output layer.

[0084] Optionally, the building block 302 includes:

[0085] The temperature prediction unit 3021 is used to input the internal temperature array, the uncorrected ambient temperature array and at least one target calibration associated parameter into the initial temperature calibration network to obtain a temperature prediction result;

[0086] A humidity prediction unit 3022 is used to input the internal humidity array, the uncorrected ambient humidity array, the temperature prediction result and at least one humidity calibration associated parameter into an initial humidity calibration network to obtain a humidity prediction result;

[0087] The model updating unit 3023 is used to update the parameters of the initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the temperature prediction result and the humidity prediction result to obtain the target temperature and humidity calibration model.

[0088] Optionally, the model updating unit 3023 is specifically used to: obtain a temperature prediction loss value and a humidity prediction loss value according to the calibrated temperature and humidity array, the temperature prediction result, the humidity prediction result and a preset loss function;

[0089] If the temperature prediction loss value and / or the humidity prediction loss value is greater than the preset loss threshold, the initial temperature and humidity calibration model is iteratively updated through the loss function and the preset learning rate until the temperature prediction loss value and the humidity prediction loss value are less than or equal to the loss threshold, and the target temperature and humidity calibration model is obtained.

[0090] Optionally, the construction module 302 is specifically used to: determine the number of network layers of the initial temperature and humidity calibration model and the number of nodes corresponding to each network layer according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and at least one target calibration associated parameter;

[0091] The weight matrix between each network layer is initialized, and the ReLU function is determined as the activation function of the initial temperature and humidity calibration model to obtain the initial temperature and humidity calibration model.

[0092] In the technical solution provided in this embodiment, a temperature and humidity calibration model is constructed and trained through an internal temperature and humidity array, an uncorrected ambient temperature and humidity array, a calibrated temperature and humidity array and at least one target calibration associated parameter, fully considering the impact of heating of internal components of the temperature collector on the ambient temperature and humidity collection, and combining the corresponding target calibration associated parameters to improve the accuracy of temperature and humidity prediction by the target temperature and humidity calibration model.

[0093] above Figure 3 and Figure 4 The temperature and humidity calibration device based on deep learning in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The temperature and humidity calibration device based on deep learning in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0094] See also Figure 5 As shown, the temperature and humidity calibration device based on deep learning includes a processor 500 and a memory 501, wherein the memory 501 stores machine executable instructions that can be executed by the processor 500, and the processor 500 executes the machine executable instructions to implement the above-mentioned temperature and humidity calibration method based on deep learning.

[0095] Further, Figure 5 The deep learning-based temperature and humidity calibration device shown also includes a bus 502 and a communication interface 503 , and the processor 500 , the communication interface 503 and the memory 501 are connected via the bus 502 .

[0096] The memory 501 may include a high-speed random access memory (RAM), and may also include a non-volatile memory (non-volatile memory), for example, at least one disk storage. The communication connection between the system network element and at least one other network element is realized through at least one communication interface 503 (which may be wired or wireless), and the Internet, wide area network, local area network, metropolitan area network, etc. may be used. The bus 502 may be an ISA bus, a PCI bus, or an EISA bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5Only one bidirectional arrow is used in the diagram, but this does not mean that there is only one bus or only one type of bus.

[0097] The processor 500 may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method can be completed by the hardware integrated logic circuit or software instructions in the processor 500. The above processor 500 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present disclosure can be implemented or executed. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. The steps of the method disclosed in conjunction with the embodiments of the present disclosure can be directly embodied as a hardware decoding processor to be executed, or a combination of hardware and software modules in the decoding processor can be executed. The software module can be located in a mature storage medium in the field such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in the memory 501 , and the processor 500 reads the information in the memory 501 and completes the method steps of the above-mentioned embodiment in combination with its hardware.

[0098] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of a temperature and humidity calibration method based on deep learning.

[0099] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0100] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0101] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A temperature and humidity calibration method based on deep learning, characterized in that: The temperature and humidity calibration method based on deep learning includes: Obtaining the internal temperature and humidity array, the uncorrected ambient temperature and humidity array, the calibrated temperature and humidity array and at least one target calibration-related parameter corresponding to each working condition of the temperature and humidity collector; constructing an initial temperature and humidity calibration model according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array, and the at least one target calibration associated parameter; Training the initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the internal temperature and humidity array, the uncorrected ambient temperature and humidity array, and the at least one target calibration associated parameter to obtain a target temperature and humidity calibration model; The temperature and humidity detection values ​​are calibrated using the target temperature and humidity calibration model.

2. The temperature and humidity calibration method based on deep learning according to claim 1, characterized in that: The step of calibrating the temperature and humidity detection values ​​by using the target temperature and humidity calibration model includes: Acquire internal temperature and humidity detection values, ambient temperature and humidity detection values, at least one temperature calibration-related parameter, and at least one humidity calibration-related parameter; Inputting the internal temperature detection value, the ambient temperature detection value and at least one temperature calibration associated parameter into the temperature calibration network of the target temperature and humidity calibration model to obtain a calibrated target ambient temperature; The target ambient temperature, the internal humidity detection value, the ambient humidity detection value and at least one humidity calibration associated parameter are input into a humidity calibration network of the target temperature and humidity calibration model to obtain a calibrated target ambient humidity.

3. The temperature and humidity calibration method based on deep learning according to claim 2, characterized in that: The step of inputting the internal temperature detection value, the ambient temperature detection value and at least one temperature calibration associated parameter into the temperature calibration network of the target temperature and humidity calibration model to obtain the calibrated target ambient temperature comprises: Inputting an internal temperature detection value, an ambient temperature detection value and at least one temperature calibration associated parameter into a first input layer of the temperature calibration network; Performing operations based on a first weight matrix between the first input layer and the first hidden layer and inputting into a first activation layer of the temperature calibration network; Performing an operation based on a second weight matrix between the first activation layer and the second hidden layer and inputting the result into a first output layer of the temperature calibration network; The target ambient temperature is outputted through the first output layer.

4. The temperature and humidity calibration method based on deep learning according to claim 3, characterized in that: The second input layer of the humidity calibration network is connected to the first output layer, and the target ambient temperature, the internal humidity detection value, the ambient humidity detection value and at least one humidity calibration associated parameter are input into the humidity calibration network of the target temperature and humidity calibration model to obtain the calibrated target ambient humidity, including: Inputting the internal humidity detection value, the ambient humidity detection value, at least one humidity calibration associated parameter, and the target ambient temperature output by the first output layer into the second input layer of the humidity calibration network; Performing operations according to a third weight matrix between the second input layer and the third hidden layer and inputting into a second activation layer of the humidity calibration network; Performing operations according to a fourth weight matrix between the second activation layer and the fourth hidden layer, and inputting into a second output layer of the humidity calibration network; The target ambient humidity is outputted through the second output layer.

5. The temperature and humidity calibration method based on deep learning according to claim 1, characterized in that: The initial temperature and humidity calibration model includes an initial temperature calibration network and an initial humidity calibration network, the internal temperature and humidity array includes an internal temperature array and an internal humidity array, the uncorrected ambient temperature and humidity array includes an uncorrected ambient temperature array and an uncorrected ambient humidity array, and the at least two target calibration associated parameters include at least one temperature calibration associated parameter and at least one humidity calibration associated parameter. The initial temperature and humidity calibration model is trained according to the calibrated temperature and humidity array, the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and the at least one temperature calibration associated parameter to obtain a target temperature and humidity calibration model, including: Inputting the internal temperature array, the uncorrected ambient temperature array and the at least one target calibration associated parameter into the initial temperature calibration network to obtain a temperature prediction result; Inputting the internal humidity array, the uncorrected ambient humidity array, the temperature prediction result and the at least one humidity calibration associated parameter into the initial humidity calibration network to obtain a humidity prediction result; The parameters of the initial temperature and humidity calibration model are updated according to the calibrated temperature and humidity array, the temperature prediction result and the humidity prediction result to obtain a target temperature and humidity calibration model.

6. The temperature and humidity calibration method based on deep learning according to claim 5, characterized in that: The back propagation of the initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the temperature prediction result and the humidity prediction result to obtain a target temperature and humidity calibration model includes: Obtaining a temperature prediction loss value and a humidity prediction loss value according to the calibrated temperature and humidity array, the temperature prediction result, the humidity prediction result and a preset loss function; If the temperature prediction loss value and / or the humidity prediction loss value is greater than a preset loss threshold, the initial temperature and humidity calibration model is iteratively updated through the loss function and the preset learning rate until the temperature prediction loss value and the humidity prediction loss value are less than or equal to the loss threshold, thereby obtaining the target temperature and humidity calibration model.

7. The temperature and humidity calibration method based on deep learning according to claim 1, characterized in that: The constructing of an initial temperature and humidity calibration model according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and the at least one target calibration associated parameter comprises: Determine the number of network layers of the initial temperature and humidity calibration model and the number of nodes corresponding to each network layer according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array and the at least one target calibration associated parameter; The weight matrix between each network layer is initialized, and the ReLU function is determined as the activation function of the initial temperature and humidity calibration model to obtain the initial temperature and humidity calibration model.

8. A temperature and humidity calibration device based on deep learning, characterized in that: The temperature and humidity calibration device based on deep learning includes: An acquisition module, used to acquire the internal temperature and humidity array, the uncorrected ambient temperature and humidity array, the calibrated temperature and humidity array and at least one target calibration associated parameter corresponding to each working condition of the temperature and humidity collector; A construction module, configured to construct an initial temperature and humidity calibration model according to the internal temperature and humidity array, the uncorrected ambient temperature and humidity array, and the at least one target calibration associated parameter; A training module, configured to train the initial temperature and humidity calibration model according to the calibrated temperature and humidity array, the internal temperature and humidity array, the uncorrected ambient temperature and humidity array, and the at least one target calibration associated parameter to obtain a target temperature and humidity calibration model; The calibration module is used to calibrate the temperature and humidity detection values ​​through the target temperature and humidity calibration model.

9. A temperature and humidity calibration device based on deep learning, characterized in that: The temperature and humidity calibration device based on deep learning includes: a memory and at least one processor, wherein instructions are stored in the memory; The at least one processor calls the instructions in the memory so that the deep learning-based temperature and humidity calibration device performs the deep learning-based temperature and humidity calibration method as described in any one of claims 1-7.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instruction is read and executed, the temperature and humidity calibration method based on deep learning as described in any one of claims 1 to 7 is executed.

Citation Information

Cited By

  • Cereal quality analysis method based on prediction-correction double neural network fusion

    CN120473030A

  • Gas detector calibration system

    CN120609981A

  • Gas detector calibration system

    CN120609981B