Sensor data error correction method and system based on big data

Abstract

The invention provides a sensor data processing method and system based on big data, and belongs to the technical field of data processing, in the method, abnormal data in an original data set uploaded by a target sensor cluster can be automatically identified and repaired under the condition of determining that the target sensor cluster meets the security condition, and the security of the target sensor cluster is improved. According to the method, a Merkel tree is established, a target data set subjected to data restoration is obtained, the integrity and accuracy of the data are ensured, and meanwhile, newly-added sensor data in each target data set can be stored to a proper target leaf node in the Merkel tree according to the updating moment of the sensor data in each target data set; according to the method, the storage overhead is effectively reduced while orderly storage of the newly-added sensor data is realized, and by performing consensus on the newly-added sensor data in the block chain, illegal control and illegal data tampering of the sensor can be effectively avoided, so that the security of data storage is ensured.

Description

Technical Field

[0001] The present disclosure relates to the field of sensor technology, and in particular, to a sensor data error correction method and system based on big data. Background Art

[0002] With the development of sensor technology, the application of sensors is becoming more and more extensive. Temperature sensor is a sensor that senses temperature and converts it into a usable output signal. Temperature sensor is the core part of temperature measuring instrument and there are many varieties.

[0003] When the temperature sensor collects temperature data, it will be affected by many factors, which may cause errors in the temperature data. Therefore, it is necessary to correct the errors of the temperature data collected by the temperature sensor. At present, the data correction of the temperature sensor has the problems of low efficiency and difficulty in ensuring accuracy. Summary of the invention

[0004] The purpose of the present disclosure is to provide a sensor data error correction method and system based on big data, which can realize effective and accurate correction of temperature sensor data.

[0005] In order to achieve the above-mentioned objectives, in a first aspect, the present disclosure provides a sensor data error correction method based on big data, comprising: obtaining sensor data to be corrected, temperature sensor information corresponding to the sensor data to be corrected, and measurement object information corresponding to the sensor data to be corrected; determining an error type according to the temperature sensor information and the measurement object information, wherein the error type is used to characterize the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information; determining an error correction parameter according to the error type, the sensor data to be corrected, and a pre-trained first prediction model; and correcting the sensor data to be corrected according to the error correction parameter to obtain corrected sensor data.

[0006] Optionally, the temperature sensor information includes: protection tube information, temperature sensing element information and measurement information, and the measurement information is used to characterize the measurement method of the temperature sensor; the measurement object information includes: measurement object morphology, measurement object state and measurement object size; the error type is determined according to the temperature sensor information and the measurement object information, including: if the measurement information characterizes that the measurement method of the temperature sensor is insertion measurement, the error influence weight of the temperature sensor information is determined according to the insertion depth in the measurement information and the protection tube information, and the error influence weight of the measurement object information is determined according to the measurement object morphology and the measurement object state; if the measurement information characterizes that the measurement method of the temperature sensor is contact measurement, the error influence weight of the temperature sensor information is determined according to the protection tube information and the temperature sensing element information, and the error influence weight of the measurement object information is determined according to the measurement object morphology and the measurement object size; the error type is determined according to the error influence weight of the temperature sensor information and the error influence weight of the measurement object information.

[0007] Optionally, determining the error type based on the error influence weight of the temperature sensor information and the error influence weight of the measured object information includes: if the error influence weight of the temperature sensor information is equal to the error influence weight of the measured object information, determining the error type as a first error type; if the error influence weight of the temperature sensor information is greater than the error influence weight of the measured object information, determining the error type as a second error type; if the error influence weight of the temperature sensor information is less than the error influence weight of the measured object information, determining the error type as a third error type.

[0008] Optionally, the sensor data error correction method based on big data also includes: determining an error correction evaluation result according to the error type, the error correction parameter, the corrected sensor data and a pre-trained second prediction model, wherein the error correction evaluation result is used to characterize the error correction effect of the error correction parameter; and feeding back the corrected sensor data and the error correction evaluation result.

[0009] Optionally, the pre-trained first prediction model is used to output the error correction parameter based on the input error type and the sensor data to be corrected, and the sensor data error correction method based on big data also includes: obtaining a first training data set, the first training data set including: first sample sensor data, error correction parameters corresponding to the first sample sensor data, and the error type corresponding to the first sample sensor data; training the first prediction model to be trained according to the first training data set to obtain the pre-trained first prediction model.

[0010] Optionally, the pre-trained second prediction model is used to output the error correction evaluation result based on the input error type, the error correction parameters and the corrected sensor data, and the sensor data error correction method based on big data also includes: obtaining a second training data set, the second training data set including: second sample sensor data, error type, target error parameters and error correction evaluation labels of original sample sensor data corresponding to the second sample sensor data, wherein the original sample sensor data is sensor data selected from the first sample sensor data; training the second prediction model to be trained according to the second training data set to obtain the pre-trained second prediction model.

[0011] Optionally, obtaining the second training data set includes: determining the original sample sensor data from the first sample sensor data; inputting the original sample sensor data and the error type corresponding to the original sample sensor data into the pre-trained first prediction model to obtain the predicted error correction parameters output by the pre-trained first prediction model; determining the target error correction parameters and the first error correction evaluation label according to the error correction parameters corresponding to the original sample sensor data and the predicted error correction parameters; determining the second sample sensor data according to the target error correction parameters and the original sample sensor data; determining the error correction evaluation label according to the first error correction evaluation label; generating the second training data set according to the second sample sensor data, the error type of the original sample sensor data corresponding to the second sample sensor data, the target error correction parameters and the error correction evaluation label.

[0012] Optionally, determining the error correction evaluation label based on the first error correction evaluation label includes: feeding back the target error correction parameter and the original sample sensor data; receiving a second error correction evaluation label input by the user based on the fed-back target error correction parameter, the original sample sensor data and a preset error correction parameter range; and determining the error correction evaluation label based on the first error correction evaluation label and the second error correction evaluation label.

[0013] Optionally, the error type includes a first error type, a second error type and a third error type, the first error type characterizing that the error influence weight of the temperature sensor information is equal to the error influence weight of the measurement object information, the second error type characterizing that the error influence weight of the temperature sensor information is greater than the error influence weight of the measurement object information, and the third error type characterizing that the error influence weight of the temperature sensor information is less than the error influence weight of the measurement object information, and determining the error correction parameter according to the error type, the sensor data to be corrected and the pre-trained first prediction model includes: if the error type is the first error type, inputting the sensor data to be corrected and the error type into the pre-trained first prediction model to obtain the pre-trained first prediction model output; if the error type is the second error type, determining the first sensor data to be corrected from the sensor data to be corrected, inputting the first sensor data to be corrected and the error type into the pre-trained first prediction model, and obtaining the error correction parameters output by the pre-trained first prediction model, wherein the acquisition time of the first sensor data to be corrected covers the target acquisition cycle; if the error type is the third error type, determining the second sensor data to be corrected from the sensor data to be corrected, inputting the second sensor data to be corrected and the error type into the pre-trained first prediction model, and obtaining the error correction parameters output by the pre-trained first prediction model, wherein the acquisition time of the second sensor data to be corrected covers multiple acquisition cycles.

[0014] In a second aspect, an embodiment of the present disclosure provides a sensor data error correction system based on big data, comprising: a temperature sensor, which is used to collect temperature data of a measurement object; and a terminal, which is connected to the temperature sensor; wherein the temperature sensor is used to transmit the temperature data of the measurement object to the terminal as the sensor data to be corrected, and the terminal is used to execute the sensor data error correction method based on big data as described in the first aspect.

[0015] Through the above technical solution, the error type is first determined using the temperature sensor information and the measurement information object, and then the error correction parameters are determined using the error type, the sensor data to be corrected, and the pre-trained first prediction model. Then, the sensor data to be corrected is corrected according to the error correction parameters to obtain the corrected sensor data. Since the error type can characterize the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information, the error type can be regarded as an influence weight of the sensor data to be corrected. Based on the error type, the error correction parameters determined by the pre-trained first prediction model have high accuracy. Moreover, the error correction parameters can be quickly determined by the pre-trained first prediction model. Furthermore, the technical solution can achieve effective and accurate correction of temperature sensor data.

[0016] Other features and advantages of the present disclosure will be described in detail in the following detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the present disclosure and constitute a part of the specification. Together with the following specific embodiments, they are used to explain the present disclosure but do not constitute a limitation of the present disclosure. In the accompanying drawings:

[0018] Figure 1 It is a structural block diagram of a sensor data error correction system based on big data according to an exemplary embodiment.

[0019] Figure 2 The present invention is a flowchart of a sensor data error correction method based on big data according to an exemplary embodiment.

[0020] Figure 3 The figure is a flowchart of the training process of the first prediction model and the second prediction model according to an exemplary embodiment.

[0021] Figure 4 It is a structural block diagram of a sensor data error correction device based on big data according to an exemplary embodiment.

[0022] Figure 5 It is a block diagram of an electronic device according to an exemplary embodiment. DETAILED DESCRIPTION

[0023] The specific implementation of the present disclosure is described in detail below in conjunction with the accompanying drawings. It should be understood that the specific implementation described herein is only used to illustrate and explain the present disclosure, and is not used to limit the present disclosure.

[0024] The technical solution of the embodiment of the present disclosure can be applied to various application scenarios involving temperature sensors. In these application scenarios, temperature sensors are used to collect temperature data.

[0025] In some application scenarios, after obtaining batches of temperature sensor data, the batches of temperature sensor data are corrected. In other application scenarios, after obtaining real-time temperature sensor data, the real-time temperature sensor data are corrected.

[0026] It is understandable that when the temperature sensor collects temperature data, it will be affected by many factors, so that the temperature data may have errors. Therefore, it is necessary to correct the errors of the temperature data collected by the temperature sensor. At present, the data correction of the temperature sensor has the problems of low efficiency and difficulty in ensuring accuracy.

[0027] Based on this, an embodiment of the present disclosure provides a technical solution, in which, based on the error type that characterizes the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information, a pre-trained first prediction model is used to quickly and accurately determine the error correction parameter, thereby achieving effective and accurate correction of the temperature sensor data.

[0028] Figure 1 is a structural block diagram of a sensor data error correction system based on big data according to an exemplary embodiment. Figure 1 As shown, the system includes: a temperature sensor and a terminal.

[0029] The temperature sensor is connected to the terminal, and is used to collect temperature data of the measured object, and transmit the temperature data of the measured object as sensor data to be corrected to the terminal. The terminal corrects the sensor data to be corrected and applies the corrected sensor data.

[0030] Regarding the measurement object, it can be a measurement object of different types, for example, it can be a measurement object of different shapes, different states and different sizes.

[0031] Regarding the temperature sensor, the number thereof may be one or more, which is not limited here. Also, the temperature sensor may be of different types, for example, different protective tubes, different temperature sensing elements, and different temperature measurement methods may be used.

[0032] In some embodiments, the system may be part of an Internet of Things system, and the system may serve as a data collection part of the Internet of Things system to achieve accurate data collection in the Internet of Things system.

[0033] For example, some scenarios involving temperature collection: equipment monitoring scenario, it is necessary to detect the temperature of the equipment to detect whether the equipment is safe; feedback control scenario, it is necessary to detect temperature data to perform feedback control based on the temperature data, etc.

[0034] The terminal may be any electronic device with data processing function.

[0035] Figure 2 is a flow chart of a sensor data error correction method based on big data according to an exemplary embodiment. Figure 2 As shown, the method comprises the following steps:

[0036] Step 201 : acquiring sensor data to be corrected, temperature sensor information corresponding to the sensor data to be corrected, and measurement object information corresponding to the sensor data to be corrected.

[0037] Step 202: determining an error type according to the temperature sensor information and the measurement object information, where the error type is used to characterize the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information.

[0038] Step 203: Determine error correction parameters according to the error type, the sensor data to be corrected, and the pre-trained first prediction model.

[0039] Step 204 , correcting the sensor data to be corrected according to the error correction parameter to obtain corrected sensor data.

[0040] In step 201, the sensor data to be corrected is temperature sensor data. The sensor data to be corrected may be real-time temperature sensor data or non-real-time temperature sensor data. The sensor data to be corrected may include one or more temperature sensor data. If there are multiple temperature sensor data, each temperature sensor data may be corrected in the same error correction method.

[0041] The temperature sensor information can be understood as the relevant information of the temperature sensor. The measurement object information can be understood as the relevant information of the object measured by the temperature sensor.

[0042] As an optional implementation, the temperature sensor information includes: protection tube information, temperature sensing element information and measurement information, the measurement information is used to characterize the measurement method of the temperature sensor; the measurement object information includes: measurement object shape, measurement object state and measurement object size.

[0043] For example, the protection tube information may include: protection tube material, protection tube attachments, and protection tube thickness. The temperature sensing element information may include: the diameter of the temperature sensing element measurement end. The measurement information is used to characterize the measurement method of the temperature sensor, such as insertion measurement, contact measurement, and non-contact measurement.

[0044] Among them, the material and thickness of the protection tube are fixed parameters and can be directly obtained. As for the attachments of the protection tube, detection can be achieved through image detection and other methods. It can be understood that these protection tube information will affect the thermal conductivity of the protection tube and cause measurement errors. For example, when the thickness of the attachments of the protection tube is thick, the thickness of the protection tube will increase, reducing the thermal conductivity of the protection tube and causing measurement errors.

[0045] The temperature sensing element information and measurement information are both fixed information and can be directly obtained. It can be understood that the smaller the diameter of the temperature sensing element, the shorter the thermal response time, and the smaller the possibility of causing errors. Different measurement methods may lead to different errors. For example, insertion measurement may cause errors due to improper insertion depth; contact measurement may cause errors due to slow thermal response speed; non-contact measurement may cause errors due to relative position relationship. Therefore, both the temperature sensing element information and the measurement information may cause errors in the temperature sensor data.

[0046] For example, the form of the measured object may include: gaseous, liquid and solid, etc. The state of the measured object may include: static and dynamic, etc. The size of the measured object may be represented by different parameters according to the different measured objects, for example, it may be represented by diameter, length, width and area, etc.

[0047] It is understandable that different shapes and states of the measured object may cause the temperature of the measured object to be inconsistent with that of the temperature sensor, thus causing measurement errors. Also, different sizes of the measured object may affect the thermal response time of the temperature sensor, thus causing measurement errors.

[0048] Therefore, in the embodiment of the present disclosure, the temperature sensor information and the measurement object information are first quantified to determine the error type, which can be used as indicative information of the first prediction model to assist the first prediction model in determining the error correction parameters more quickly and accurately.

[0049] Therefore, in step 202, the error type characterizing the error influence weight of the temperature sensor information and the error influence weight of the measurement object information is determined according to the temperature sensor information and the measurement object information.

[0050] As an optional implementation, step 202 includes: if the measurement information indicates that the measurement method of the temperature sensor is insertion measurement, determining the error influence weight of the temperature sensor information according to the insertion depth and the protection tube information in the measurement information, and determining the error influence weight of the measurement object information according to the shape of the measurement object and the state of the measurement object; if the measurement information indicates that the measurement method of the temperature sensor is contact measurement, determining the error influence weight of the temperature sensor information according to the protection tube information and the temperature sensing element information, and determining the error influence weight of the measurement object information according to the shape of the measurement object and the size of the measurement object; determining the error type according to the error influence weight of the temperature sensor information and the error influence weight of the measurement object information.

[0051] In this implementation, the error influence weight of the temperature sensor information and the error influence weight of the measurement object information are first determined, and then the error type is determined based on the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information.

[0052] It is understandable that when using common insertion temperature sensors, the selection of the temperature measurement point is very important. For the production process, the location of the temperature measurement should be typical and representative, otherwise the measurement is meaningless. When the temperature sensor is inserted into the measurement object, some heat flow will be generated along the length of the sensor. When the ambient temperature is relatively low, there will be heat loss, resulting in inconsistent temperatures between the temperature sensor and the measurement object, which will cause temperature measurement errors.

[0053] Therefore, the measurement error caused by heat conduction is related to the insertion depth, which in turn is related to the material of the protection tube.

[0054] For different materials, the metal protection tube has good thermal conductivity, so its insertion depth should be deeper, while the ceramic material has better thermal insulation performance and can be inserted shallower. In some aspects, for engineering temperature measurement, the insertion depth is also related to whether the measured object is static or flowing. For example, the measurement of the temperature of flowing liquid or high-speed airflow will not be subject to the above restrictions, and the insertion depth can also be shallower. The specific value can be determined by experiment.

[0055] Therefore, when the measurement information indicates that the measurement method of the temperature sensor is insertion measurement, the error influence weight of the temperature sensor information can be determined according to the insertion depth and protection tube information in the measurement information.

[0056] In some embodiments, an error influence weight table can be generated in advance through experiments, and the error influence weight table includes influence weights corresponding to different insertion depths and different protection tube information. By looking up the table, the error influence weights corresponding to the insertion depth and the protection tube information can be determined, and then the error influence weights corresponding to the insertion depth and the protection tube information are weighted and summed to determine the error influence weight of the temperature sensor information. Among them, the weighted weights of the insertion depth and the protection tube information can be determined according to the actual situation, and their sum can be 1.

[0057] Regarding the error influence weight of the measurement object information, it is also possible to generate an error influence weight table in advance through experiments, and then determine the error influence weight of the measurement object information by looking up the table.

[0058] In some embodiments, the temperature sensor information and the measurement object information may correspond to different error influence weight tables respectively, or may be integrated into the same error influence weight table, and the latter has higher accuracy than the former.

[0059] Therefore, in the embodiment of the present disclosure, in order to determine the error impact weight, it is necessary to analyze a large amount of data, that is, to implement it based on big data.

[0060] It can be understood that the basic principle of contact temperature measurement in temperature sensors is that the temperature measuring element must reach a thermal balance with the measured object. Therefore, when measuring temperature, the two need to be kept in contact for a certain period of time to allow the two to reach a thermal balance. The length of the contact time is related to the thermal response time of the temperature measuring element. The thermal response time mainly depends on the structure of the temperature sensor and the measurement conditions, which vary greatly. Therefore, the temperature sensor may not be able to keep up with the temperature change speed of the measured object, or it may also cause some measurement errors due to failure to achieve thermal balance.

[0061] In addition, for temperature sensors, in addition to the influence of the protective tube, the diameter of the measuring end of the temperature sensing element is also a major factor. That is, the thinner the temperature sensing element and the smaller the diameter of the measuring end, the shorter its thermal response time will be.

[0062] In addition, for temperature sensors used at high temperatures, if the medium of the measured object is in gaseous form, some dust deposited on the surface of the protection tube will melt on the surface of the protection tube, making the protection tube thicker and increasing its thermal impedance.

[0063] If the measured object is in the form of a melt, slag will be deposited on it during use, which will not only increase the response time of the temperature sensor, but also make the indicated temperature lower, resulting in deviation.

[0064] Therefore, when the measurement information indicates that the measurement method of the temperature sensor is contact measurement, the error influence weight of the temperature sensor information is determined according to the protection tube information and the temperature sensing element information. Also, the error influence weight of the measurement object information is determined according to the measurement object shape and the measurement object size.

[0065] The implementation method for determining the error influence weight of the temperature sensor information and the error influence weight of the measurement object information may refer to the above embodiment, that is, it may also be determined by looking up a table.

[0066] In some embodiments, if the measurement information indicates that the measurement method of the temperature sensor is non-contact measurement, the error influence weight of the measurement object information may be 0. The error influence weight of the temperature sensor information may be determined according to the distance between the temperature sensor and the measurement object. Specifically, the error influence weight may be determined in accordance with the aforementioned error influence weight table, which will not be repeated here.

[0067] Furthermore, determining the error type based on the error influence weight of the temperature sensor information and the error influence weight of the measurement object information may include: if the error influence weight of the temperature sensor information is equal to the error influence weight of the measurement object information, determining the error type as a first error type; if the error influence weight of the temperature sensor information is greater than the error influence weight of the measurement object information, determining the error type as a second error type; if the error influence weight of the temperature sensor information is less than the error influence weight of the measurement object information, determining the error type as a third error type.

[0068] In this embodiment, three error types are divided. The first error type represents that the error influence weights of the two errors are equal; the second error type represents that the error influence weight of the temperature sensor information is greater than the error influence weight of the measurement object information; and the third error type represents that the error influence weight of the temperature sensor information is less than the error influence weight of the measurement object information.

[0069] In some embodiments, the three error types may be represented by corresponding identifiers, which may be vectors or characters. For example, the first error type is represented by 0, the second error type is represented by 1, and the third error type is represented by 2. For another example, the first error type is represented by [0,0,0], the second error type is represented by [1,1,1], and the third error type is represented by [2,2,2].

[0070] In step 203, error correction parameters are determined according to the error type, the sensor data to be corrected, and the pre-trained first prediction model.

[0071] In some embodiments, the pre-trained first prediction model is used to output error correction parameters according to the input error type and sensor data to be corrected.

[0072] In some embodiments, the sensor data to be corrected may be first converted into a vector form and then input into the first prediction model.

[0073] In some embodiments, the first prediction model may be a large language model or a neural network model, etc.

[0074] In some embodiments, the training process of the first prediction model may include: obtaining a first training data set, the first training data set including: first sample sensor data, error correction parameters corresponding to the first sample sensor data, and error types corresponding to the first sample sensor data; training the first prediction model to be trained according to the first training data set to obtain a pre-trained first prediction model.

[0075] In this implementation, the first sample sensor data may be historically collected temperature sensor data, and the historically collected temperature sensor data is corrected by the error correction parameter corresponding to the first sample sensor data, and the error correction parameter corresponding to the first sample sensor data may be determined manually or by other mature error correction parameter determination methods. Also, the error type corresponding to the first sample sensor data may be determined by the error type determination method in the aforementioned embodiment, or may be determined manually or by other mature determination methods.

[0076] Furthermore, the first prediction model to be trained is trained by the first training data set to obtain a pre-trained first prediction model, so that the first prediction model can output the pre-trained first prediction model based on the sensor data to be corrected under the indication of the error type.

[0077] Furthermore, in step 204, the sensor data to be corrected is corrected according to the error correction parameter to obtain the corrected sensor data.

[0078] In some embodiments, the correction method of the error correction parameter may be a default method, for example, adding the error correction parameter to the sensor data to be corrected, or multiplying the error correction parameter to complete the correction.

[0079] In the disclosed embodiment, the correction effect of the error correction parameter can also be evaluated. Therefore, as an optional implementation, the method further includes: determining an error correction evaluation result according to the error type, the error correction parameter, the corrected sensor data and the pre-trained second prediction model, the error correction evaluation result is used to characterize the error correction effect of the error correction parameter; and feeding back the corrected sensor data and the error correction evaluation result.

[0080] In this embodiment, the error correction effect is evaluated at the level of corrected sensor data in combination with the error type and the error correction parameter.

[0081] In some embodiments, the first prediction model and the second prediction model may adopt different model types; or they may adopt the same model type.

[0082] Regarding the second prediction model, the corresponding training data set may include: second sample sensor data, error types of original sample sensor data corresponding to the second sample sensor data, target error parameters, and error correction evaluation labels.

[0083] In some embodiments, the second sample sensor data is correction data corresponding to historically collected temperature sensor data.

[0084] In the embodiment of the present disclosure, since the first prediction model and the second prediction model are used, in order to improve the training efficiency of the model and reduce the amount of data collection in the early stage, the training data set of the second prediction model can be obtained based on the training data set of the first prediction model.

[0085] Therefore, as an optional implementation, the training process of the second prediction model includes: obtaining a second training data set, the second training data set including: second sample sensor data, error type, target error parameter and error correction evaluation label of original sample sensor data corresponding to the second sample sensor data, wherein the original sample sensor data is sensor data selected from the first sample sensor data; training the second prediction model to be trained according to the second training data set to obtain a pre-trained second prediction model.

[0086] In this embodiment, the original sample sensor data is sensor data selected from the first sample sensor data, and accordingly, the error type, target error parameter and error correction evaluation label of the original sample sensor data corresponding to the second sample sensor data can be determined in combination with the first prediction model.

[0087] Therefore, as an optional implementation, obtaining a second training data set includes: determining original sample sensor data from the first sample sensor data; inputting the original sample sensor data and the error type corresponding to the original sample sensor data into a pre-trained first prediction model to obtain prediction error correction parameters output by the pre-trained first prediction model; determining target error correction parameters and a first error correction evaluation label based on the error correction parameters and the prediction error correction parameters corresponding to the original sample sensor data; determining the second sample sensor data based on the target error correction parameters and the original sample sensor data; determining the error correction evaluation label based on the first error correction evaluation label; generating a second training data set based on the second sample sensor data and the error type, target error correction parameters and the error correction evaluation label of the original sample sensor data corresponding to the second sample sensor data.

[0088] The original sample sensor data may be sampled and acquired from the first sample sensor data, and the sampling method may refer to the mature technology in the art.

[0089] After the original sample sensor data is acquired, the original sample sensor data and the error type corresponding to the original sample sensor data may be input into a pre-trained first prediction model to obtain a prediction error correction parameter output by the pre-trained first prediction model.

[0090] Therefore, in the embodiment of the present disclosure, the first prediction model needs to be trained first, and then the second prediction model is trained. Also, it can be understood that since the original sample sensor data is determined from the first sample sensor data, the error type corresponding to the original sample sensor data is also known and can be directly obtained.

[0091] Based on the error correction parameters corresponding to the original sample sensor data and the predicted error correction parameters, the target error correction parameters and the first error correction evaluation label may be determined.

[0092] For example, if the difference between the error correction parameter corresponding to the original sample sensor data and the predicted error correction parameter is large, the error correction parameter corresponding to the original sample sensor data is determined as the target error correction parameter. If the difference between the error correction parameter corresponding to the original sample sensor data and the predicted error correction parameter is small, the predicted error parameter is determined as the target error correction parameter. If they are equal, the target error correction parameter can be any error correction parameter.

[0093] If the difference (absolute value) between the error correction parameter corresponding to the original sample sensor data and the predicted error correction parameter is greater than the preset difference, the difference is considered to be large. Otherwise, the difference is considered to be small. The preset difference can be determined according to the accuracy of the temperature sensor. For example, if the accuracy of the sensor is ±1, the preset difference can be 1.

[0094] In some embodiments, the first error correction evaluation tag is determined based on a difference between an error correction parameter corresponding to the original sample sensor data and a predicted error correction parameter.

[0095] For example, based on the accuracy of the sensor, it can be subdivided. For example, for the accuracy of +1, it can be divided into 0-0.5, 0.5-0.8, 0.8-1 and 1-∞. Then, the first error correction evaluation label corresponding to 0-0.5 can be 3, the first error correction evaluation label corresponding to 0.5-0.8 can be 2, the first error correction evaluation label corresponding to 0.8-1 can be 1, and the first error correction evaluation label corresponding to 1-∞ can be 0. Moreover, the larger the first error correction evaluation label, the better the error correction evaluation effect.

[0096] In some embodiments, the first error correction evaluation label may be directly determined as the error correction evaluation label.

[0097] In other embodiments, determining an error correction evaluation label based on a first error correction evaluation label includes: feeding back target error correction parameters and original sample sensor data; receiving a second error correction evaluation label input by a user based on the fed-back target error correction parameters, original sample sensor data, and a preset error correction parameter range; and determining the error correction evaluation label based on the first error correction evaluation label and the second error correction evaluation label.

[0098] In this embodiment, the target error correction parameter and the original sample sensor data may also be fed back to the user, and the user may input a second error correction evaluation label based on the fed-back target error correction parameter, the original sample sensor data and the preset error correction parameter range.

[0099] The preset error correction parameter range is a preset error correction parameter range that meets the requirements, which can constrain the error correction. For example, if the error correction parameter is too large or too small, the error correction is meaningless.

[0100] Therefore, the user can evaluate the error correction effect based on experience and a preset error correction parameter range, and input a second error correction evaluation tag.

[0101] Then, the first error correction evaluation label and the second error correction evaluation label are weighted and summed to determine the final error correction evaluation label, wherein the weight of the second error correction evaluation label is greater than the weight of the first error correction evaluation label.

[0102] Furthermore, a second training data set may be generated according to the second sample sensor data, the error type of the original sample sensor data corresponding to the second sample sensor data, the target error correction parameter, and the error correction evaluation label.

[0103] In addition, the first prediction model can also be optimized based on the error correction parameters and predicted error correction parameters corresponding to the original sample sensor data. For example, the difference between the error correction parameters and the corresponding predicted error correction parameters corresponding to multiple original sample sensor data is determined respectively to obtain multiple differences. If the average value of the multiple differences is less than the preset difference, no optimization is required. If it is greater than or equal to the preset difference, the first prediction model needs to be optimized. The optimization method can be: training with a new training data set; adjusting model parameters, etc.

[0104] Figure 3 is a flowchart of the training process of the first prediction model and the second prediction model according to an exemplary embodiment. Figure 3 As shown, the first prediction model is trained based on the first training data set.

[0105] The first prediction model is trained using the second training data set, and the second training data set is generated based on the first training data set. Specifically, a portion of the training data is selected from the first training data set, and then the target error correction parameter and the error correction evaluation label are determined. The determination of the error correction evaluation label also introduces manual labeling. And the basis of the manual labeling is the target error correction parameter and the original sample sensor data.

[0106] In addition, based on the prediction error correction parameter, the first prediction model may also be optimized.

[0107] In this way, not only can the training of the first prediction model and the second prediction model be completed, but the accuracy of the two models can also be guaranteed, and the amount of external data obtained can be reduced, thereby reducing the difficulty of model training.

[0108] In the embodiment of the present disclosure, another implementation method for determining the error correction parameter is also provided. In this implementation method, it is necessary to first determine the data input into the first prediction model according to the error type.

[0109] Therefore, as an optional implementation, step 203 includes: if the error type is the first error type, inputting the sensor data to be corrected and the error type into the pre-trained first prediction model to obtain the error correction parameters output by the pre-trained first prediction model; if the error type is the second error type, determining the first sensor data to be corrected from the sensor data to be corrected, inputting the first sensor data to be corrected and the error type into the pre-trained first prediction model to obtain the error correction parameters output by the pre-trained first prediction model, wherein the acquisition time of the first sensor data to be corrected covers the target acquisition period; if the error type is the third error type, determining the second sensor data to be corrected from the sensor data to be corrected, inputting the second sensor data to be corrected and the error type into the pre-trained first prediction model to obtain the error correction parameters output by the pre-trained first prediction model, wherein the acquisition time of the second sensor data to be corrected covers multiple acquisition periods.

[0110] In this implementation, if the error type is the first error type, the sensor data to be corrected may be input into the first prediction model to obtain error correction parameters.

[0111] If the error type is the second error type, the sensor data to be corrected whose acquisition time covers the target acquisition period can be input into the first prediction model, wherein the target acquisition period can be the first acquisition period, or the first acquisition period and the second acquisition period.

[0112] It can be understood that if the error impact weight of the temperature sensor information is greater than the error impact weight of the measured object information, it means that the error caused by the temperature sensor has a greater impact. However, the error caused by the temperature sensor is usually only for the data at the beginning of the collection, and the error of the data collected later is very small. Therefore, it is only necessary to determine the error correction parameters for the sensor data collected in the earlier collection cycle.

[0113] If the error type is the third error type, the sensor data to be corrected covering multiple acquisition cycles may be input into the first prediction model, wherein covering multiple acquisition cycles may be understood as acquiring a portion of sensor data in each acquisition cycle.

[0114] It can be understood that if the error impact weight of the temperature sensor information is less than the error impact weight of the measurement object information, it means that the error caused by the measurement object has a greater impact, and the error caused by the measurement object usually runs through the entire acquisition cycle. Therefore, it is necessary to cover sensor data of multiple acquisition cycles to determine the error correction parameters.

[0115] It can be understood that since multiple acquisition cycles may be involved, this implementation is more suitable for non-real-time temperature sensor data.

[0116] It can be seen from the introduction of the embodiments of the present disclosure that the error type is first determined using the temperature sensor information and the measurement information object, and then the error correction parameters are determined using the error type, the sensor data to be corrected, and the pre-trained first prediction model. Then, the sensor data to be corrected is corrected according to the error correction parameters to obtain the corrected sensor data. Since the error type can characterize the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information, the error type can be regarded as an influence weight of the sensor data to be corrected. Based on the error type, the error correction parameters determined using the pre-trained first prediction model have high accuracy. Moreover, the error correction parameters can be quickly determined using the pre-trained first prediction model. Furthermore, this technical solution can achieve effective and accurate correction of temperature sensor data.

[0117] Figure 4 is a structural block diagram of a sensor data error correction device 400 based on big data according to an exemplary embodiment. Figure 4 As shown, the device comprises:

[0118] The acquisition module 401 is used to acquire sensor data to be corrected, temperature sensor information corresponding to the sensor data to be corrected, and measurement object information corresponding to the sensor data to be corrected.

[0119] The correction module 402 is used to determine the error type according to the temperature sensor information and the measurement object information, wherein the error type is used to characterize the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information; determine the error correction parameter according to the error type, the sensor data to be corrected and the pre-trained first prediction model; and correct the sensor data to be corrected according to the error correction parameter to obtain the corrected sensor data.

[0120] Regarding the device in the above embodiment, the specific manner in which each module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0121] Figure 5 FIG. 5 is a block diagram of an electronic device 500 according to an exemplary embodiment. Figure 5 As shown, the electronic device 500 may include: a processor 501 and a memory 502. The electronic device 500 may also include one or more of a multimedia component 503, an input / output (I / O) interface 504, and a communication component 505.

[0122] The processor 501 is used to control the overall operation of the electronic device 500 to complete all or part of the steps in the above-mentioned sensor data error correction method based on big data. The memory 502 is used to store various types of data to support the operation of the electronic device 500, and these data may include, for example, instructions for any application or method used to operate on the electronic device 500, and application-related data, such as contact data, sent and received messages, pictures, audio, video, etc. The memory 502 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, referred to as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, referred to as EEPROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, referred to as EPROM), programmable read-only memory (Programmable Read-Only Memory, referred to as PROM), read-only memory (Read-Only Memory, referred to as ROM), magnetic memory, flash memory, disk or optical disk. The multimedia component 503 may include a screen and an audio component. The screen may be, for example, a touch screen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone, which is used to receive external audio signals. The received audio signal may be further stored in the memory 502 or sent through the communication component 505. The audio component also includes at least one speaker for outputting audio signals. The I / O interface 504 provides an interface between the processor 501 and other interface modules, and the above-mentioned other interface modules may be keyboards, mice, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 505 is used for wired or wireless communication between the electronic device 500 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G or 4G, or a combination of one or more of them, so the corresponding communication component 505 may include: Wi-Fi module, Bluetooth module, NFC module.

[0123] In an exemplary embodiment, the electronic device 500 can be implemented by one or more application specific integrated circuits (ASIC), digital signal processors (DSP), digital signal processing devices (DSPD), programmable logic devices (PLD), field programmable gate arrays (FPGA), controllers, microcontrollers, microprocessors or other electronic components to execute the above-mentioned big data-based sensor data error correction method.

[0124] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, and when the program instructions are executed by a processor, the steps of the sensor data error correction method based on big data are implemented. For example, the computer-readable storage medium can be the memory 502 including the program instructions, and the program instructions can be executed by the processor 501 of the electronic device 500 to complete the sensor data error correction method based on big data.

[0125] In another exemplary embodiment, a computer program product is also provided. The computer program product includes a computer program that can be executed by a processor. When the computer program is executed by the processor, the steps of the above-mentioned sensor data error correction method based on big data are implemented.

[0126] The preferred embodiments of the present disclosure are described in detail above in conjunction with the accompanying drawings; however, the present disclosure is not limited to the specific details in the above embodiments. Within the technical concept of the present disclosure, a variety of simple modifications can be made to the technical solution of the present disclosure, and these simple modifications all fall within the protection scope of the present disclosure.

[0127] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present disclosure will not further describe various possible combinations.

[0128] In addition, various embodiments of the present disclosure may be arbitrarily combined, and as long as they do not violate the concept of the present disclosure, they should also be regarded as the contents disclosed by the present disclosure.

Claims

1. A sensor data error correction method based on big data, characterized in that: include: Acquire sensor data to be corrected, temperature sensor information corresponding to the sensor data to be corrected, and measurement object information corresponding to the sensor data to be corrected; Determine an error type according to the temperature sensor information and the measurement object information, where the error type is used to characterize the relationship between the error influence weight of the temperature sensor information and the error influence weight of the measurement object information; Determining an error correction parameter according to the error type, the sensor data to be corrected, and a pre-trained first prediction model; The sensor data to be corrected is corrected according to the error correction parameter to obtain corrected sensor data.

2. The sensor data error correction method based on big data according to claim 1, characterized in that: The temperature sensor information includes: protection tube information, temperature sensing element information and measurement information, wherein the measurement information is used to characterize the measurement method of the temperature sensor; the measurement object information includes: measurement object form, measurement object state and measurement object size; the error type is determined according to the temperature sensor information and the measurement object information, including: If the measurement information indicates that the measurement method of the temperature sensor is insertion measurement, determining the error influence weight of the temperature sensor information according to the insertion depth in the measurement information and the protection tube information, and determining the error influence weight of the measurement object information according to the measurement object morphology and the measurement object state; If the measurement information indicates that the measurement method of the temperature sensor is contact measurement, determining the error influence weight of the temperature sensor information according to the protection tube information and the temperature sensing element information, and determining the error influence weight of the measurement object information according to the measurement object shape and the measurement object size; The error type is determined according to the error influence weight of the temperature sensor information and the error influence weight of the measurement object information.

3. The sensor data error correction method based on big data according to claim 2, characterized in that: The determining the error type according to the error influence weight of the temperature sensor information and the error influence weight of the measurement object information includes: If the error influence weight of the temperature sensor information is equal to the error influence weight of the measurement object information, determining the error type as a first error type; If the error influence weight of the temperature sensor information is greater than the error influence weight of the measurement object information, determining the error type as a second error type; If the error influence weight of the temperature sensor information is less than the error influence weight of the measurement object information, the error type is determined as a third error type.

4. The sensor data error correction method based on big data according to claim 1, characterized in that: The sensor data error correction method based on big data also includes: Determine an error correction evaluation result according to the error type, the error correction parameter, the corrected sensor data and the pre-trained second prediction model, wherein the error correction evaluation result is used to characterize the error correction effect of the error correction parameter; The corrected sensor data and the error correction evaluation result are fed back.

5. According to the sensor data error correction method based on big data in claim 4, the pre-trained first prediction model is used to output the error correction parameter according to the input error type and the sensor data to be corrected, and the sensor data error correction method based on big data further includes: Acquire a first training data set, the first training data set comprising: first sample sensor data, error correction parameters corresponding to the first sample sensor data, and error types corresponding to the first sample sensor data; The first prediction model to be trained is trained according to the first training data set to obtain the pre-trained first prediction model.

6. According to the sensor data error correction method based on big data in claim 5, the pre-trained second prediction model is used to output the error correction evaluation result according to the input error type, the error correction parameter and the corrected sensor data, and the sensor data error correction method based on big data further includes: Acquire a second training data set, the second training data set comprising: second sample sensor data, error types, target error parameters, and error correction evaluation labels of original sample sensor data corresponding to the second sample sensor data, wherein the original sample sensor data is sensor data selected from the first sample sensor data; The second prediction model to be trained is trained according to the second training data set to obtain the pre-trained second prediction model.

7. The sensor data error correction method based on big data according to claim 6, characterized in that: The obtaining of the second training data set comprises: determining the raw sample sensor data from the first sample sensor data; Inputting the original sample sensor data and the error type corresponding to the original sample sensor data into the pre-trained first prediction model to obtain a prediction error correction parameter output by the pre-trained first prediction model; Determining a target error correction parameter and a first error correction evaluation label according to the error correction parameter corresponding to the original sample sensor data and the predicted error correction parameter; determining the second sample sensor data according to the target error correction parameter and the original sample sensor data; determining the error correction evaluation label according to the first error correction evaluation label; The second training data set is generated according to the second sample sensor data, the error type of the original sample sensor data corresponding to the second sample sensor data, the target error correction parameter and the error correction evaluation label.

8. The sensor data error correction method based on big data according to claim 7, characterized in that: The step of determining the error correction evaluation label according to the first error correction evaluation label comprises: feeding back the target error correction parameter and the original sample sensor data; receiving a second error correction evaluation label input by a user based on the target error correction parameter fed back, the original sample sensor data and a preset error correction parameter range; The error correction evaluation label is determined according to the first error correction evaluation label and the second error correction evaluation label.

9. The sensor data error correction method based on big data according to claim 1, characterized in that: The error type includes a first error type, a second error type and a third error type, the first error type indicates that the error influence weight of the temperature sensor information is equal to the error influence weight of the measurement object information, the second error type indicates that the error influence weight of the temperature sensor information is greater than the error influence weight of the measurement object information, and the third error type indicates that the error influence weight of the temperature sensor information is less than the error influence weight of the measurement object information, and determining the error correction parameter according to the error type, the sensor data to be corrected and the pre-trained first prediction model includes: If the error type is the first error type, inputting the sensor data to be corrected and the error type into the pre-trained first prediction model to obtain an error correction parameter output by the pre-trained first prediction model; If the error type is the second error type, determining first sensor data to be corrected from the sensor data to be corrected, inputting the first sensor data to be corrected and the error type into the pre-trained first prediction model, and obtaining error correction parameters output by the pre-trained first prediction model, wherein the collection time of the first sensor data to be corrected covers the target collection period; If the error type is the third error type, second sensor data to be corrected is determined from the sensor data to be corrected, and the second sensor data to be corrected and the error type are input into the pre-trained first prediction model to obtain the error correction parameters output by the pre-trained first prediction model, wherein the acquisition time of the second sensor data to be corrected covers multiple acquisition cycles.

10. A sensor data error correction system based on big data, characterized in that: include: A temperature sensor, wherein the temperature sensor is used to collect temperature data of a measurement object; A terminal connected to the temperature sensor; The temperature sensor is used to transmit the temperature data of the measured object as the sensor data to be corrected to the terminal, and the terminal is used to execute the sensor data error correction method based on big data as described in any one of claims 1 to 9.

Images