Dry-burn identification model training and dry-burn detection method, system, device and medium

Abstract

The present disclosure provides a training method, system, device and medium of a dry burning identification model and a dry burning detection method, system, device and medium, the training method comprising: obtaining a plurality of groups of training sample data; wherein each group of training sample data comprises historical temperature field data, historical pot bottom temperature data and dry burning identification data corresponding to a cooking device; and training a preset network based on each group of training sample data to obtain a dry burning identification model. The present disclosure trains a dry burning identification model by obtaining the historical temperature field data, historical pot bottom temperature data and dry burning identification data of a plurality of groups of cooking devices, improves the accuracy and reliability of the temperature data, ensures the accuracy, reliability and timeliness of the dry burning identification model obtained by training in judging the dry burning state of the cooking device, ensures the safety in the cooking scene, and improves the user experience.

Description

Technical Field

[0001] This disclosure relates to the field of cooking equipment technology, and in particular to a method, system, device and medium for training a dry-burning recognition model and detecting dry-burning. Background Technology

[0002] Anti-dry-burning cooktops have always been popular with consumers. By installing a pressure-contact thermocouple temperature probe on the cooktop's burner, the probe activates when a pot is placed on it, sensing the pressure and measuring the temperature of the pot's bottom. In some dry-burning scenarios, it can provide an early warning and shut off the cooktop. However, because different pots have different temperature-sensing characteristics (i.e., reflectivity), the measured bottom temperature can be inaccurate. In addition, in some special scenarios, the temperature difference between the bottom of the pot and the internal temperature can be significant, and the temperature data is relatively singular, resulting in low reliability and a tendency for false alarms related to dry-burning.

[0003] Using infrared temperature sensors to measure temperature values ​​to prevent dry burning is a common method in the industry. Infrared temperature measurement has a fast response speed; it can measure the temperature of a target as long as it receives infrared thermal radiation, without needing to contact the target. It also allows for selection of conditions such as measurement distance and viewing angle. At the same time, by imaging the target using a thermal infrared sensitive CCD (charge-coupled device), it can reflect the temperature field of the target surface. By imaging and analyzing the cookware in the cooking area of ​​the stove using infrared thermal imaging, dry burning status can be identified and warnings can be issued. However, in dry burning scenarios such as when the cookware is covered with a lid, the infrared thermal radiation of the contents of the cookware is blocked by the lid, which may prevent timely judgment of whether dry burning has occurred. Summary of the Invention

[0004] The technical problem to be solved by this disclosure is to overcome the shortcomings of existing technologies that use pressure contact thermocouple temperature probes for dry burning prevention, such as low accuracy of measured temperature, low reliability of temperature data, and easy misjudgment of dry burning prevention, and the shortcomings of using infrared non-contact temperature sensors for dry burning prevention, such as inability to timely determine whether dry burning has occurred. The disclosure provides a method, system, device and medium for training a dry burning identification model and for dry burning detection.

[0005] This disclosure solves the above-mentioned technical problems through the following technical solution:

[0006] This disclosure provides a training method for a dry-burning recognition model, the training method comprising:

[0007] Acquire several sets of training sample data; wherein, each set of training sample data includes historical temperature field data, historical pot bottom temperature data, and dry burning indicator data corresponding to the cooking equipment;

[0008] The preset network is trained based on the training sample data of each group to construct the dry burning recognition model.

[0009] Preferably, the step of training the preset network based on each set of training sample data to construct the dry-burning recognition model includes:

[0010] Obtain historical image information corresponding to the cooking equipment;

[0011] Based on the historical image information, obtain the cooking equipment information corresponding to the cooking equipment;

[0012] The cooking equipment information includes lid information and contents information; the lid information indicates whether the cooking equipment is covered with a lid; the contents information indicates whether the cooking equipment contains contents.

[0013] Based on the cooking equipment information, a first historical weight corresponding to the historical temperature field data and a second historical weight corresponding to the historical pot bottom temperature data are determined.

[0014] The dry-burning identification model is trained based on the historical temperature field data with the first historical weight, the historical pot bottom temperature data with the second historical weight, and the dry-burning identification data.

[0015] Preferably, when the cooking device is covered with a lid, the first historical weight is less than the second historical weight;

[0016] When the cooking device contains contents, the first historical weight is greater than the second historical weight;

[0017] When the cooking device is not covered and / or the cooking device is empty, the first historical weight is equal to the second historical weight.

[0018] Preferably, the step of training the preset network based on each set of training sample data to construct the dry burning recognition model further includes:

[0019] Based on the historical temperature field data, a temperature matrix with multiple time window lengths corresponding to the cooking equipment is obtained;

[0020] The dry burning identification model is trained based on the temperature matrix, the historical pot bottom temperature data, and the dry burning identification data.

[0021] Preferably, the step of training the dry-burning identification model based on the temperature matrix, the historical pot bottom temperature data, and the dry-burning identification data includes:

[0022] Based on the temperature matrix, the temperature characteristics of the cooking device in each time window are calculated;

[0023] The dry burning identification model is trained based on the temperature characteristics, the historical pot bottom temperature data, and the dry burning identification data.

[0024] Preferably, the step of training the dry-burning recognition model based on the temperature features, the historical pot bottom temperature data, and the dry-burning identification data includes:

[0025] The temperature features, the historical pot bottom temperature data, and the dry burning identification data are input into a machine learning supervised model to train the dry burning identification model.

[0026] Preferably, the step of obtaining several sets of training sample data includes:

[0027] Obtain historical temperature field data of the cooktop surface corresponding to the cooktop surface where the cooking equipment is located;

[0028] Based on the historical cooktop temperature field data, the historical temperature field data corresponding to the cooking equipment is obtained according to the area corresponding to the cooking equipment.

[0029] And / or,

[0030] The cooking equipment information also includes material information;

[0031] The step of obtaining several sets of training sample data also includes:

[0032] Based on the material information, the temperature emissivity corresponding to the cooking device is obtained;

[0033] Based on the temperature emissivity, the historical pot bottom temperature data corresponding to the cooking device is obtained.

[0034] This disclosure also provides a dry-burning detection method, which is implemented based on the training method of the dry-burning recognition model described above, and the dry-burning detection method includes:

[0035] Obtain the actual temperature field data and actual pot bottom temperature data corresponding to the cooking equipment;

[0036] The actual temperature field data and the actual pot bottom temperature data are input into the dry burning recognition model to output the dry burning state corresponding to the cooking device.

[0037] This disclosure also provides a training system for a dry-burning recognition model, the training system comprising:

[0038] The training sample data acquisition module is used to acquire several sets of training sample data; wherein, each set of training sample data includes historical temperature field data, historical pot bottom temperature data and dry burning indicator data corresponding to the cooking equipment;

[0039] The dry-burning recognition model training module is used to train a preset network based on each set of training sample data to construct the dry-burning recognition model.

[0040] Preferably, the dry-burning recognition model training module includes a historical image information acquisition unit, a cooking equipment information acquisition unit, a weight determination unit, and a dry-burning recognition model training unit;

[0041] The historical image information acquisition unit is used to acquire historical image information corresponding to the cooking equipment;

[0042] The cooking equipment information acquisition unit is used to acquire cooking equipment information corresponding to the cooking equipment based on the historical image information;

[0043] The cooking equipment information includes lid information and contents information; the lid information indicates whether the cooking equipment is covered with a lid; the contents information indicates whether the cooking equipment contains contents.

[0044] The weight determination unit is used to determine the first historical weight corresponding to the historical temperature field data and the second historical weight corresponding to the historical pot bottom temperature data based on the cooking equipment information.

[0045] The dry-burning identification model training unit is used to train the dry-burning identification model based on the historical temperature field data with the first historical weight, the historical pot bottom temperature data with the second historical weight, and the dry-burning identification data.

[0046] Preferably, when the cooking device is covered with a lid, the first historical weight is less than the second historical weight;

[0047] When the cooking device contains contents, the first historical weight is greater than the second historical weight;

[0048] When the cooking device is not covered and / or the cooking device is empty, the first historical weight is equal to the second historical weight.

[0049] Preferably, the dry burning identification model training module further includes a temperature matrix acquisition unit;

[0050] The temperature matrix acquisition unit is used to acquire a temperature matrix with multiple time window lengths corresponding to the cooking device based on the historical temperature field data.

[0051] The dry-burning identification model training unit is also used to train the dry-burning identification model based on the temperature matrix, the historical pot bottom temperature data, and the dry-burning identification data.

[0052] Preferably, the dry-burning identification model training unit includes a temperature feature calculation subunit and a dry-burning identification model training subunit;

[0053] The temperature feature calculation subunit is used to calculate the temperature features of the cooking device in each time window based on the temperature matrix.

[0054] The dry-burning identification model training subunit is used to train the dry-burning identification model based on the temperature features, the historical pot bottom temperature data, and the dry-burning identification data.

[0055] Preferably, the dry-burning identification model training subunit is further configured to input the temperature features, the historical pot bottom temperature data, and the dry-burning identification data into a machine learning supervised model to train the dry-burning identification model.

[0056] Preferably, the training sample data acquisition module includes a historical stove surface temperature field data acquisition unit and a historical temperature field data acquisition unit;

[0057] The historical stove surface temperature field data acquisition unit is used to acquire the historical stove surface temperature field data corresponding to the stove surface where the cooking equipment is located.

[0058] The historical temperature field data acquisition unit is used to acquire the historical temperature field data corresponding to the cooking equipment based on the historical stove surface temperature field data and the area corresponding to the cooking equipment.

[0059] And / or,

[0060] The cooking equipment information also includes material information;

[0061] The training sample data acquisition module also includes a temperature emissivity acquisition unit and a historical pot bottom temperature data acquisition unit;

[0062] The temperature emissivity acquisition unit is used to acquire the temperature emissivity corresponding to the cooking device based on the material information;

[0063] The historical pot bottom temperature data acquisition unit is used to acquire the historical pot bottom temperature data corresponding to the cooking device based on the temperature emissivity.

[0064] This disclosure also provides a dry-burning detection system, which is implemented based on the training system of the dry-burning recognition model described above, and the dry-burning detection system includes:

[0065] The actual temperature data acquisition module is used to acquire the actual temperature field data and actual pot bottom temperature data corresponding to the cooking equipment;

[0066] The dry-burning status output module is used to input the actual temperature field data and the actual pot bottom temperature data into the dry-burning recognition model to output the dry-burning status corresponding to the cooking device.

[0067] This disclosure also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and for running on the processor, wherein the processor executes the computer program to implement the training method for the dry burning recognition model as described above; or to implement the dry burning detection method as described above.

[0068] This disclosure also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the training method for the dry burning recognition model as described above; or implements the dry burning detection method as described above.

[0069] Based on common knowledge in the field, the preferred conditions described can be combined arbitrarily to obtain various preferred embodiments of this disclosure.

[0070] The positive and progressive effects of this disclosure are as follows:

[0071] This disclosure trains a dry-burning recognition model by acquiring historical temperature field data, historical pot bottom temperature data, and dry-burning indicator data from multiple sets of cooking equipment. The actual temperature field data and actual pot bottom temperature data of the cooking equipment are then input into this dry-burning recognition model to output the corresponding dry-burning state of the cooking equipment. This overcomes the problem of low accuracy of measured temperature data due to different temperature measurement characteristics of different cooking equipment, as well as the low reliability of using a single temperature data point for dry-burning state judgment. It improves the accuracy and reliability of temperature data, ensuring the accuracy, reliability, and timeliness of dry-burning state judgment of cooking equipment in different usage scenarios, guaranteeing safety in cooking scenarios, and enhancing the user experience. Attached Figure Description

[0072] Figure 1 This is a flowchart of the training method for the dry burning recognition model in Embodiment 1 of this disclosure.

[0073] Figure 2 This is a first flowchart of the training method for the dry burning recognition model in Embodiment 2 of this disclosure.

[0074] Figure 3 This is a second flowchart of the training method for the dry burning recognition model in Embodiment 2 of this disclosure.

[0075] Figure 4This is a flowchart of the dry burning detection method of Embodiment 3 of this disclosure.

[0076] Figure 5 This is a schematic diagram of the training system for the dry burning recognition model of Embodiment 4 of this disclosure.

[0077] Figure 6 This is a schematic diagram of the training system for the dry burning recognition model of Embodiment 5 of this disclosure.

[0078] Figure 7 This is a schematic diagram of the dry burning detection system of Embodiment 6 of this disclosure.

[0079] Figure 8 This is a schematic diagram of the structure of the electronic device according to Embodiment 7 of this disclosure. Detailed Implementation

[0080] The present disclosure is further illustrated below by way of embodiments, but the present disclosure is not limited to the scope of the embodiments described herein.

[0081] Example 1

[0082] This embodiment provides a training method for a dry-burning recognition model, such as... Figure 1 As shown, the training method includes:

[0083] S101. Obtain several sets of training sample data; wherein, each set of training sample data includes historical temperature field data, historical pot bottom temperature data and dry burning label data corresponding to the cooking equipment.

[0084] Specifically, an infrared thermal imaging module that can cover the countertop of gas stoves (including natural gas stoves, manufactured gas stoves, liquefied petroleum gas stoves, etc.) at a specific sampling frequency (the area where the stove used by the user is located) can be used to collect historical temperature field data. The infrared thermal imaging module can output countertop temperature field data images and temperature field data, where the temperature field data can be a temperature matrix.

[0085] A pressure-contact thermocouple temperature sensor is installed in the center of the stove. When cooking equipment (such as pots and pans) is placed on it, the pressure sensor activates and outputs the temperature data of the bottom of the pot.

[0086] The dry-burning label data is the tag value indicating whether the cooking equipment is in a dry-burning state.

[0087] S102. Train the preset network based on each set of training sample data to build a dry burning recognition model.

[0088] Specifically, a dry-burning identification model is trained by using historical temperature field data and historical pot bottom temperature data as inputs and dry-burning identification data as outputs.

[0089] In this embodiment, a dry-burning identification model is trained by acquiring historical temperature field data, historical pot bottom temperature data, and dry-burning identification data of multiple sets of cooking equipment. This overcomes the problem of low accuracy of measured temperature data due to different temperature measurement characteristics of different cooking equipment, as well as the problem of low reliability of using a single temperature data to judge the dry-burning state. It improves the accuracy and reliability of temperature data and ensures the accuracy, reliability, and timeliness of the dry-burning identification model trained on the cooking equipment to judge the dry-burning state.

[0090] Example 2

[0091] This embodiment provides a training method for a dry burning recognition model, which is a further improvement on Embodiment 1.

[0092] In a feasible solution, such as Figure 2 As shown, step S102 includes:

[0093] S1021. Obtain historical image information corresponding to the cooking equipment.

[0094] S1022. Based on historical image information, obtain the cooking equipment information corresponding to the cooking equipment.

[0095] The cooking equipment information includes lid information and contents information; the lid information indicates whether the cooking equipment is covered with a lid; the contents information indicates whether the cooking equipment contains contents.

[0096] Specifically, the countertop temperature field data image output by the infrared thermal imaging module is analyzed to identify the cooking equipment information and output historical temperature field data of the area containing the cooking equipment. Image information corresponding to the cooking equipment can be obtained from the countertop temperature field data image.

[0097] Historical countertop temperature field data images can be acquired using an infrared thermal imaging module. A dataset can be collected, including the location, size, material, whether a lid is on, and whether there are contents on the cooking equipment placed on the stove's burner. Then, a feature extraction network and a classification network are designed and trained to create a countertop temperature field analysis model. By acquiring current countertop temperature field data images and inputting them into the model, the model can output information about the cooking equipment, including its location, size, material, whether a lid is on, and whether there are contents. Additionally, a camera can be used to identify the cooking equipment's information.

[0098] S1023. Based on the cooking equipment information, determine the first historical weight corresponding to the historical temperature field data and the second historical weight corresponding to the historical pot bottom temperature data.

[0099] S1024. Based on historical temperature field data with the first historical weight, historical pot bottom temperature data with the second historical weight, and dry burning identification data, a dry burning identification model is trained.

[0100] In this solution, the initial weights of historical temperature field data and historical pot bottom temperature data are adjusted based on the information of the pot lid and contents of the cooking equipment. This improves the accuracy and reliability of the temperature data and ensures the accuracy, reliability, and timeliness of the dry burning recognition model trained on the cooking equipment in judging its dry burning status.

[0101] In a feasible solution, when the cooking device is covered with a lid, the first historical weight is less than the second historical weight.

[0102] When the cooking device contains contents, the first historical weight is greater than the second historical weight;

[0103] The contents include food, oil, water, etc. from the cooking equipment;

[0104] When the cooking device is not covered and / or the cooking device is empty, the first historical weight is equal to the second historical weight.

[0105] Specifically, when the cooking device is covered with a lid, the first historical weight corresponding to the historical temperature field data is less than the second historical weight corresponding to the historical pot bottom temperature; when the cooking device contains contents, the first historical weight corresponding to the historical temperature field data is less than the second historical weight corresponding to the historical pot bottom temperature; in other cases, the first historical weight corresponding to the historical temperature field data is equal to the second historical weight corresponding to the historical pot bottom temperature.

[0106] In this solution, the first historical weight corresponding to the historical temperature field data and the second historical weight corresponding to the historical pot bottom temperature are adjusted according to whether the cooking equipment is covered with a lid and whether it contains contents. This improves the accuracy, reliability and rationality of the temperature data, and ensures the accuracy, reliability and timeliness of the dry burning recognition model trained on the cooking equipment in judging the dry burning state.

[0107] In a feasible solution, such as Figure 3 As shown, step S102 further includes:

[0108] S1025. Based on historical temperature field data, obtain the temperature matrix corresponding to the cooking equipment with multiple time window lengths.

[0109] Specifically, after obtaining historical temperature field data, it is filled into multiple time windows to form a multidimensional temperature matrix with multiple time window lengths, including 1 second, 2 seconds, 5 seconds, 10 seconds, 15 seconds, 30 seconds, and 60 seconds.

[0110] S1026. Based on the temperature matrix, historical pot bottom temperature data, and dry burning identification data, a dry burning identification model is trained.

[0111] In this solution, a temperature matrix with multiple time window lengths is obtained based on historical temperature field data to train a dry-burning identification model, ensuring the accuracy and reliability of the dry-burning identification model in judging the dry-burning state of cooking equipment.

[0112] In one feasible embodiment, step S1026 includes:

[0113] S10261. Based on the temperature matrix, calculate the temperature characteristics of the cooking equipment in each time window;

[0114] Among them, the calculation methods for temperature characteristics include variance, mean, mean difference, extreme values, and range;

[0115] S10262. Based on temperature characteristics, historical pot bottom temperature data, and dry burning identification data, a dry burning identification model is trained.

[0116] In this scheme, the dry-burning identification data consists of a label value indicating whether the cooking equipment is in a dry-burning state within a time window. Based on the temperature matrix, multiple methods are used to calculate the temperature characteristics of the cooking equipment to train a dry-burning identification model, ensuring the accuracy and reliability of the dry-burning identification model in judging the dry-burning state of the cooking equipment.

[0117] In one feasible embodiment, step S10262 includes:

[0118] S102621. Input the temperature characteristics, historical pot bottom temperature data and dry burning identification data into the machine learning supervised model to train a dry burning identification model.

[0119] The supervised machine learning model includes support vector machines, KNN (K nearest neighbor classification algorithm), decision trees, Naive Bayes, regression and neural networks, etc. The supervised machine learning model is trained to converge in the direction of improving prediction accuracy in order to obtain the dry burning recognition model.

[0120] In this solution, a dry-burning recognition model is trained through machine learning supervised model training, which ensures the accuracy, reliability and timeliness of the dry-burning recognition model in judging the dry-burning state of cooking equipment.

[0121] In one feasible embodiment, step S101 includes:

[0122] S1011. Obtain historical temperature field data of the cooktop surface corresponding to the cooktop surface where the cooking equipment is located;

[0123] S1012. Based on historical stove surface temperature field data, obtain the historical temperature field data corresponding to the cooking equipment according to the area corresponding to the cooking equipment.

[0124] In this solution, the historical temperature field data of the cooktop is obtained through an infrared thermal imaging module, and the historical temperature field data of the cooking equipment is further obtained, ensuring the accuracy and reliability of the historical temperature field data.

[0125] In a feasible solution, the cooking equipment information also includes material information;

[0126] Step S101 also includes:

[0127] S1013. Based on material information, obtain the temperature emissivity corresponding to the cooking equipment;

[0128] S1014. Based on the temperature emissivity, obtain the historical bottom temperature data of the cooking equipment.

[0129] Specifically, after obtaining the material information of the cooking utensils output by the countertop temperature field analysis model, it is sent to the pressure contact thermocouple temperature probe on the stove through a communication link. According to the material information of the cooking equipment, the temperature probe is adjusted to the temperature emissivity corresponding to the material information to ensure the accuracy of historical pot bottom temperature data collection. When the temperature probe has pressure sensing contact, it starts to work and outputs historical pot bottom temperature data through the communication link.

[0130] In this solution, historical image information of the cooking equipment is obtained based on the infrared thermal imaging module, and the material information of the cooking equipment is further obtained. Based on the material information, the temperature emissivity of the pressure contact thermocouple temperature sensing probe is measured, thereby improving the accuracy and reliability of historical pot bottom temperature data.

[0131] The working principle of the training method of the dry burning recognition model in this embodiment is explained below with specific examples:

[0132] Historical temperature field data of the cooktop surface where the cooking equipment is located is acquired using an infrared thermal imaging module. Based on this historical temperature field data, historical temperature field data of the cooking equipment is obtained. Historical image information of the cooking equipment is acquired using the infrared thermal imaging module to identify whether the equipment is covered with a lid, contains any contents, and its material information. The emissivity of the pressure-contact thermocouple temperature sensor is adjusted based on the material information to obtain historical pot bottom temperature data. The historical temperature field data and historical pot bottom temperature data are weighted according to the information such as whether the equipment is covered with a lid and whether it contains any contents, resulting in historical temperature field data with first historical weight and historical pot bottom temperature data with second historical weight. A temperature matrix with multiple time window lengths is obtained from the historical temperature field data, and temperature features are calculated from this matrix. Dry-burning label data of the cooking equipment is obtained. Based on the temperature features of the cooking equipment, historical pot bottom temperature data, and dry-burning label data, a machine learning supervised model is trained to obtain a dry-burning recognition model.

[0133] In this embodiment, historical temperature field data, historical pot bottom temperature data, and dry-burning identification data of multiple cooking devices are acquired. The historical pot bottom temperature data is weighted according to material information, and the historical temperature field data and historical pot bottom temperature data are weighted according to pot lid information and contents information. A temperature matrix with multiple time window lengths is obtained based on the historical temperature field data. Temperature features are then used to train a machine learning supervised model to obtain a dry-burning identification model. This overcomes the problem of low accuracy of measured temperature data due to different temperature measurement characteristics of different cooking devices, as well as the problem of low reliability of using a single temperature data to judge the dry-burning state. It improves the accuracy and reliability of temperature data, and ensures the accuracy, reliability, and timeliness of the dry-burning identification model trained to judge the dry-burning state of cooking devices.

[0134] Example 3

[0135] This embodiment provides a dry-burning detection method, which is implemented based on the training method of the dry-burning recognition model described in Embodiment 1 or 2, such as... Figure 4 As shown, the dry-burning test method includes:

[0136] S201. Obtain the actual temperature field data and actual pot bottom temperature data corresponding to the cooking equipment;

[0137] S202. Input the actual temperature field data and the actual pot bottom temperature data into the dry burning recognition model to output the dry burning state corresponding to the cooking equipment.

[0138] Specifically, the actual temperature field data is obtained through an infrared thermal imaging module, and the actual pot bottom temperature data is obtained through a pressure-contact thermocouple temperature sensing probe. These data are then input into a trained dry-burning recognition model to output the corresponding dry-burning state of the cooking equipment, i.e., the final dry-burning detection and recognition result (yes or no).

[0139] Furthermore, upon detecting a dry-burning scenario, the system can trigger the stove to shut off and issue a notification to the user, allowing them to take further action. Applying this dry-burning detection method to stoves can enhance their level of intelligence.

[0140] In this embodiment, by inputting the acquired actual temperature field data and actual pot bottom temperature data into the trained dry-burning recognition model, the dry-burning status of the cooking equipment is output. This overcomes the problem of low accuracy of the measured temperature data due to the different temperature measurement characteristics of different cooking equipment, as well as the problem of low reliability of using a single temperature data to judge the dry-burning status. It improves the accuracy and reliability of the temperature data, ensures the accuracy, reliability and timeliness of judging the dry-burning status of the cooking equipment in different usage scenarios, ensures safety in cooking scenarios, and improves the user experience.

[0141] Example 4

[0142] This embodiment provides a training system for a dry-burning recognition model, such as... Figure 5 As shown, the training system includes:

[0143] Training sample data acquisition module 1 is used to acquire several sets of training sample data; wherein, each set of training sample data includes historical temperature field data, historical pot bottom temperature data and dry burning identification data corresponding to the cooking equipment;

[0144] The dry-burning recognition model training module 2 is used to train the preset network based on each set of training sample data to build a dry-burning recognition model.

[0145] The working principle of this embodiment is the same as that of the training method of the dry burning recognition model corresponding to Embodiment 1, and will not be discussed further here.

[0146] In this embodiment, a dry-burning identification model is trained by acquiring historical temperature field data, historical pot bottom temperature data, and dry-burning identification data of multiple sets of cooking equipment. This overcomes the problem of low accuracy of measured temperature data due to different temperature measurement characteristics of different cooking equipment, as well as the problem of low reliability of using a single temperature data to judge the dry-burning state. It improves the accuracy and reliability of temperature data and ensures the accuracy, reliability, and timeliness of the dry-burning identification model trained on the cooking equipment to judge the dry-burning state.

[0147] Example 5

[0148] This embodiment provides a training system for a dry-burning recognition model, which is a further improvement on embodiment 4, such as... Figure 6 As shown.

[0149] In one feasible solution, the dry-burning recognition model training module 2 includes a historical image information acquisition unit 21, a cooking equipment information acquisition unit 22, a weight determination unit 23, and a dry-burning recognition model training unit 24.

[0150] The historical image information acquisition unit 21 is used to acquire historical image information corresponding to the cooking equipment;

[0151] The cooking equipment information acquisition unit 22 is used to acquire cooking equipment information corresponding to the cooking equipment based on historical image information;

[0152] The cooking equipment information includes lid information and contents information; the lid information indicates whether the cooking equipment is covered with a lid; the contents information indicates whether the cooking equipment contains contents.

[0153] The weight determination unit 23 is used to determine the first historical weight corresponding to the historical temperature field data and the second historical weight corresponding to the historical pot bottom temperature data based on the cooking equipment information;

[0154] The dry-burning identification model training unit 24 is used to train the dry-burning identification model based on historical temperature field data with the first historical weight, historical pot bottom temperature data with the second historical weight, and dry-burning identification data.

[0155] In a feasible solution, when the cooking device is covered with a lid, the first historical weight is less than the second historical weight.

[0156] When the cooking device contains contents, the first historical weight is greater than the second historical weight;

[0157] When the cooking device is not covered and / or the cooking device is empty, the first historical weight is equal to the second historical weight.

[0158] In one feasible solution, the dry-burning identification model training module also includes a temperature matrix acquisition unit 25;

[0159] The temperature matrix acquisition unit 25 is used to acquire a temperature matrix with multiple time window lengths corresponding to the cooking equipment based on historical temperature field data.

[0160] The dry-burning identification model training unit 24 is also used to train a dry-burning identification model based on the temperature matrix, historical pot bottom temperature data and dry-burning identification data.

[0161] In one feasible solution, the dry burning identification model training unit 24 includes a temperature feature calculation subunit 241 and a dry burning identification model training subunit 242.

[0162] The temperature feature calculation subunit 241 is used to calculate the temperature features of the cooking device in each time window based on the temperature matrix.

[0163] The dry-burning identification model training subunit 242 is used to train a dry-burning identification model based on temperature features, historical pot bottom temperature data, and dry-burning identification data.

[0164] In one feasible scheme, the dry burning recognition model training subunit 242 is also used to input temperature features, historical pot bottom temperature data and dry burning identification data into a machine learning supervised model to train a dry burning recognition model.

[0165] In one feasible solution, the training sample data acquisition module 1 includes a historical stove surface temperature field data acquisition unit 11 and a historical temperature field data acquisition unit 12;

[0166] The historical cooktop temperature field data acquisition unit 11 is used to acquire the historical cooktop temperature field data corresponding to the cooktop where the cooking equipment is located.

[0167] The historical temperature field data acquisition unit 12 is used to acquire the historical temperature field data corresponding to the cooking equipment based on the historical stove surface temperature field data and the area corresponding to the cooking equipment.

[0168] In a feasible solution, the cooking equipment information also includes material information;

[0169] The training sample data acquisition module also includes a temperature emissivity acquisition unit 13 and a historical pot bottom temperature data acquisition unit 14;

[0170] The temperature emissivity acquisition unit 13 is used to acquire the temperature emissivity of the cooking device based on the material information;

[0171] The historical pot bottom temperature data acquisition unit 14 is used to acquire historical pot bottom temperature data corresponding to the cooking equipment based on the temperature emissivity.

[0172] The working principle of this embodiment is the same as that of the training method of the dry burning recognition model corresponding to Embodiment 2, and will not be discussed further here.

[0173] In this embodiment, historical temperature field data, historical pot bottom temperature data, and dry-burning identification data of multiple cooking devices are acquired. The historical pot bottom temperature data is weighted according to material information, and the historical temperature field data and historical pot bottom temperature data are weighted according to pot lid information and contents information. A temperature matrix with multiple time window lengths is obtained based on the historical temperature field data. Temperature features are then used to train a machine learning supervised model to obtain a dry-burning identification model. This overcomes the problem of low accuracy of measured temperature data due to different temperature measurement characteristics of different cooking devices, as well as the problem of low reliability of using a single temperature data to judge the dry-burning state. It improves the accuracy and reliability of temperature data, and ensures the accuracy, reliability, and timeliness of the dry-burning identification model trained to judge the dry-burning state of cooking devices.

[0174] Example 6

[0175] This embodiment provides a dry-burning detection system, which is implemented based on the training system of the dry-burning recognition model described in Embodiment 4 or 5, such as... Figure 7 As shown, the dry-burn detection system includes:

[0176] The actual temperature data acquisition module 3 is used to acquire the actual temperature field data and actual pot bottom temperature data corresponding to the cooking equipment.

[0177] The dry-burning status output module 4 is used to input the actual temperature field data and the actual pot bottom temperature data into the dry-burning recognition model to output the dry-burning status corresponding to the cooking device.

[0178] The working principle of this embodiment is the same as that of the dry burning detection method corresponding to Embodiment 3, and will not be discussed further here.

[0179] In this embodiment, by inputting the acquired actual temperature field data and actual pot bottom temperature data into the trained dry-burning recognition model, the dry-burning status of the cooking equipment is output. This overcomes the problem of low accuracy of the measured temperature data due to the different temperature measurement characteristics of different cooking equipment, as well as the problem of low reliability of using a single temperature data to judge the dry-burning status. It improves the accuracy and reliability of the temperature data, ensures the accuracy, reliability and timeliness of judging the dry-burning status of the cooking equipment in different usage scenarios, ensures safety in cooking scenarios, and improves the user experience.

[0180] Example 7

[0181] like Figure 8 The diagram shown is a structural schematic of an electronic device provided in Embodiment 7 of this disclosure. It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the training method for the dry-burning recognition model or the dry-burning detection method as described in the above embodiments. Figure 8 The electronic device 90 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0182] The electronic device 90 may be in the form of a general-purpose computing device, such as a server device. The components of the electronic device 90 may include, but are not limited to: at least one processor 91, at least one memory 92, and a bus 93 connecting different system components (including memory 92 and processor 91).

[0183] Bus 93 includes a data bus, an address bus, and a control bus.

[0184] The memory 92 may include volatile memory, such as random access memory (RAM) 921 and / or cache memory 922, and may further include read-only memory (ROM) 923.

[0185] The memory 92 may also include a program / utility 925 having a set (at least one) of program modules 924, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0186] The processor 91 executes various functional applications and data processing by running computer programs stored in the memory 92, such as the dry burning identification model training method or dry burning detection method in the above embodiments of this disclosure.

[0187] Electronic device 90 can also communicate with one or more external devices 94 (e.g., keyboard, pointing device, etc.). This communication can be performed via input / output (I / O) interface 95. Furthermore, the model-generated device 90 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public network, such as the Internet) via network adapter 96. As shown, network adapter 96 communicates with other modules of the model-generated device 90 via bus 93. It should be understood that, although not shown in the figure, other hardware and / or software modules can be used in conjunction with the model-generated device 90, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID (disk array) systems, tape drives, and data backup storage systems.

[0188] It should be noted that although several units / modules or sub-units / modules of the electronic device have been mentioned in the detailed description above, this division is merely exemplary and not mandatory. In fact, according to embodiments of this disclosure, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided and embodied by multiple units / modules.

[0189] Example 8

[0190] This embodiment provides a computer-readable storage medium storing a computer program thereon. When the program is executed by a processor, it implements the training method or dry burning detection method of the dry burning recognition model as described in the above embodiment.

[0191] The readable storage medium may be more specifically adopted, including but not limited to: portable disk, hard disk, random access memory, read-only memory, erasable programmable read-only memory, optical storage device, magnetic storage device, or any suitable combination thereof.

[0192] In possible implementations, this disclosure can also be implemented as a program product comprising program code that, when the program product is run on a terminal device, causes the terminal device to execute a training method or a dry-burning detection method that implements the dry-burning recognition model as described in the above embodiments.

[0193] The program code for executing this disclosure can be written in any combination of one or more programming languages. The program code can be executed entirely on a user device, partially on a user device, as a standalone software package, partially on a user device and partially on a remote device, or entirely on a remote device.

[0194] While specific embodiments of this disclosure have been described above, those skilled in the art should understand that these are merely illustrative examples, and the scope of protection of this disclosure is defined by the appended claims. Those skilled in the art can make various changes or modifications to these embodiments without departing from the principles and essence of this disclosure, but all such changes and modifications fall within the scope of protection of this disclosure.

Claims

1. A training method for a dry-burning recognition model, characterized in that, The training method comprises: obtaining a plurality of sets of training sample data; wherein each set of training sample data comprises historical temperature field data, historical pot bottom temperature data and dry burning identification data corresponding to a cooking device; training a preset network based on each set of training sample data to construct the dry burning identification model; The step of training the preset network based on each set of training sample data to construct the dry burning identification model comprises: obtaining historical image information corresponding to the cooking device; based on the historical image information, obtaining cooking device information corresponding to the cooking device; wherein the cooking device information comprises lid information and content information; the lid information is used to represent whether the cooking device is covered with a pot cover; the content information is used to represent whether the cooking device has content; based on the cooking device information, determining a first historical weight corresponding to the historical temperature field data and a second historical weight corresponding to the historical pot bottom temperature data; based on the historical temperature field data of the first historical weight, the historical pot bottom temperature data of the second historical weight and the dry burning identification data, the dry burning identification model is trained; when the cooking device is covered with a pot cover, the first historical weight is less than the second historical weight; when the cooking device has content, the first historical weight is greater than the second historical weight; when the cooking device is not covered with a pot cover and / or the cooking device has no content, the first historical weight is equal to the second historical weight. 2.The training method of the dry boiling identification model of claim 1, wherein, The step of training the preset network based on each set of training sample data to construct the dry burning identification model further comprises: based on the historical temperature field data, obtaining a temperature matrix of multiple time window lengths corresponding to the cooking device; based on the temperature matrix, the historical pot bottom temperature data and the dry burning identification data, the dry burning identification model is trained. 3.The method of claim 2, wherein, The step of training the preset network based on each set of training sample data to construct the dry burning identification model further comprises: based on the temperature matrix, the temperature features corresponding to the cooking device in each time window are calculated; based on the temperature features, the historical pot bottom temperature data and the dry burning identification data, the dry burning identification model is trained. 4.The method of claim 3, wherein, The step of training the preset network based on each set of training sample data to construct the dry burning identification model further comprises: inputting the temperature features, the historical pot bottom temperature data and the dry burning identification data into a machine learning supervised model to train the dry burning identification model. 5.The method of claim 1-4, wherein, The step of obtaining a plurality of sets of training sample data comprises: obtaining historical cooktop temperature field data corresponding to the cooktop on which the cooking device is located; based on the historical cooktop temperature field data, obtaining the historical temperature field data corresponding to the cooking device according to the region corresponding to the cooking device; and / or, The cooking device information further comprises material information; The step of obtaining a plurality of sets of training sample data further comprises: based on the material information, obtaining the temperature measurement emissivity corresponding to the cooking device; Based on the temperature measurement emissivity, the historical pot bottom temperature data corresponding to the cooking equipment is acquired.

6. A dry boil detection method characterized by, The dry burning detection method is implemented based on the training method of the dry burning identification model in any one of claims 1-5, and the dry burning detection method comprises: Acquiring actual temperature field data and actual pot bottom temperature data corresponding to the cooking equipment; The actual temperature field data and the actual pot bottom temperature data are input into the dry burning identification model to output the dry burning state corresponding to the cooking equipment. 7.A system for training a dry boil recognition model, the system comprising: The training system comprises: A training sample data acquisition module is configured to acquire a plurality of sets of training sample data; wherein each set of training sample data comprises historical temperature field data, historical pot bottom temperature data and dry burning identification data corresponding to a cooking equipment; A dry burning identification model training module is configured to train a preset network based on each set of training sample data to construct the dry burning identification model; The dry burning identification model training module comprises a historical image information acquisition unit, a cooking equipment information acquisition unit, a weight determination unit and a dry burning identification model training unit; The historical image information acquisition unit is configured to acquire historical image information corresponding to the cooking equipment; The cooking equipment information acquisition unit is configured to acquire cooking equipment information corresponding to the cooking equipment based on the historical image information; The cooking equipment information comprises lid information and content information; the lid information is used to represent whether the cooking equipment is covered with a lid; and the content information is used to represent whether the cooking equipment has content; The weight determination unit is configured to determine a first historical weight corresponding to the historical temperature field data and a second historical weight corresponding to the historical pot bottom temperature data based on the cooking equipment information; The dry burning identification model training unit is configured to train the dry burning identification model based on the historical temperature field data with the first historical weight, the historical pot bottom temperature data with the second historical weight and the dry burning identification data; When the cooking equipment is covered with a lid, the first historical weight is less than the second historical weight; When the cooking equipment has content, the first historical weight is greater than the second historical weight; When the cooking equipment is not covered with a lid and / or the cooking equipment has no content, the first historical weight is equal to the second historical weight.

8. A dry boil detection system characterized by, The dry burning detection system is implemented based on the training system of the dry burning identification model in claim 7, and the dry burning detection system comprises: An actual temperature data acquisition module is configured to acquire actual temperature field data and actual pot bottom temperature data corresponding to the cooking equipment; A dry burning state output module is configured to input the actual temperature field data and the actual pot bottom temperature data into the dry burning identification model to output the dry burning state corresponding to the cooking equipment.

9. An electronic device comprising a memory, a processor, and a computer program stored on the memory for running on the processor, characterized in that, The processor executes the computer program to implement the training method of the dry burning identification model in any one of claims 1-5; or, implement the dry burning detection method in claim 6.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the training method of the dry burning identification model according to any one of claims 1-5; or, implement the dry burning detection method according to claim 6.

Images