An RFID-based liquid leakage detection method and detection system

Abstract

The present invention relates to the technical field of radio frequency identification applications, and specifically relates to a liquid leakage detection method and detection system based on RFID. The method includes using an RFID reader to send a radio frequency signal to an RFID liquid-resistant tag attached to a target object, and acquiring the signal data backscattered by the RFID liquid-resistant tag after being activated by the radio frequency signal. Selecting multiple sample data to extract feature vectors including signal strength and phase, defining class labels including leakage mode, leakage liquid type, and leakage volume for the feature vectors, constructing a liquid leakage detection model based on the feature vectors with defined class labels, inputting the real-time acquired signal data into the liquid leakage detection model, and outputting the liquid leakage detection result. The present invention can quickly identify the leakage situation, leakage location, leakage mode, leakage liquid type, and leakage volume of the target object, and can achieve high-precision dynamic monitoring of the liquid leakage situation of the target object.

Description

Technical Field

[0001] The present invention relates to the technical field of radio frequency identification applications, and particularly to a liquid leakage detection method and detection system based on RFID. Background Art

[0002] Currently, some important equipment and facilities often need to be protected against liquid leakage, such as hospital medical equipment, data center equipment, liquid transmission pipelines, etc. Once a liquid leakage accident occurs, it will cause certain harm. Therefore, it is necessary to detect liquid leakage in important facilities to avoid the situation from deteriorating. With the continuous development of the Internet of Things technology, the application of radio frequency identification (RFID) technology in liquid environments has become increasingly popular.

[0003] In the prior art, the invention patent with the publication number CN116933167A discloses a liquid perception method based on RFID in a mobile scenario, and proposes a liquid perception method based on the coupling effect of materials and tags and the relationship between phase and RSSI. Through four parts: data acquisition and preprocessing, phase and signal strength model construction, signal feature extraction, and liquid type identification, liquid perception is finally realized. However, this method only judges the liquid type and does not involve the liquid leakage detection part; the invention patent with the publication number CN118032239A discloses a non-destructive liquid leakage detection method based on RFID, and proposes to densely arrange the liquid containers on the conveyor belt in a uniform grid form of M×N and pass them through the detection box to judge in real time whether there is a liquid leakage in the liquid container. However, this method does not involve the key elements in liquid leakage detection, such as the identification ability of leakage type, leakage mode, and leakage volume; the utility model patent with the publication number CN217827214U discloses a hazardous chemical management cabinet based on RFID, biometric identification, and status monitoring, and proposes to integrate RFID technology, biometric identification technology, and various status monitoring sensors to achieve efficient management and real-time monitoring of hazardous chemicals. However, this technology is mainly applicable to the safety management of hazardous chemicals in laboratory scenarios, lacks applicability in multiple scenarios, and does not conduct a detailed analysis of liquid leakage.

[0004] Therefore, it is necessary to accurately identify and detect the liquid leakage type, leakage mode, and leakage volume for liquid leakage detection. Summary of the Invention

[0005] The technical problem to be solved in the embodiments of the present invention is to provide a liquid leakage detection method and detection system based on RFID to solve the problem in the prior art that the liquid leakage type, leakage mode, and leakage volume cannot be accurately identified and detected.

[0006] The present invention discloses a method for detecting liquid leakage based on RFID, including:

[0007] Using an RFID reader / writer to send a radio frequency signal to an RFID liquid-resistant tag attached to a target object, and obtaining signal data backscattered by the RFID liquid-resistant tag after being activated by the radio frequency signal;

[0008] Selecting a part of the obtained signal data as sample data, extracting feature vectors including signal strength and phase from each of the sample data, and storing the extracted feature vectors in a database;

[0009] Defining a class label including leakage mode, leakage liquid type, and leakage volume for each feature vector in the database;

[0010] Constructing a liquid leakage detection model based on the feature vectors with class labels defined in the database, inputting the signal data obtained in real time into the liquid leakage detection model, and outputting a liquid leakage detection result including the leakage situation, leakage location, leakage mode, leakage liquid type, and leakage volume of the target object.

[0011] Optionally, the step of using an RFID reader / writer to send a radio frequency signal to an RFID liquid-resistant tag attached to a target object and obtaining signal data backscattered by the RFID liquid-resistant tag after being activated by the radio frequency signal includes:

[0012] Based on RFID liquid-resistant tags pre-arranged at different positions on the target object, the RFID reader / writer sends a radio frequency signal to the RFID liquid-resistant tag at a set frequency and constant power to activate the RFID liquid-resistant tag;

[0013] When the RFID liquid-resistant tag is activated by radio frequency, the RFID reader / writer collects in real time the signal data backscattered by each RFID liquid-resistant tag.

[0014] Optionally, before storing the extracted feature vectors in the database, a method for preprocessing the feature vectors is further included, including:

[0015] According to the feature vectors including signal strength and phase extracted from the sample data, standardizing and unifying the scale of the feature vectors to obtain standardized feature data, and the function expression of standardization is:

[0016] [[ID=3l]]

[0017] In the formula, is the standard value of the feature vector, is the value of the feature vector, is the mean value of the feature vector, is the standard deviation of the eigenvector;

[0018] Fill in the missing values of the standardized feature data with the average value to obtain the preprocessed feature data. The functional expression for filling with the average value is:

[0019]

[0020] In the formula, is the average value of non-missing values, is the non-missing value of the i-th feature data, and n is the number of non-missing values;

[0021] Judge whether there are outliers in the preprocessed feature data according to the preset conditions. The functional expression of the preset conditions is:

[0022]

[0023] In the formula, k is the condition coefficient;

[0024] If there are outliers in the preprocessed feature data, re-preprocess the feature data with outliers.

[0025] Optionally, constructing the liquid leakage detection model according to the eigenvector with class labels defined in the database includes:

[0026] Establish a training set according to the preprocessed feature data. The functional expression of the training set is:

[0027]

[0028] In the formula, is the signal strength and phase after preprocessing of the n-th sample data, is the class label including the leakage mode, leakage liquid type, and leakage volume defined for the n-th sample data;

[0029] Select the KNN algorithm to establish a basic classification model, and input the training set into the basic classification model for training to obtain the liquid leakage detection model;

[0030] Select a sample data without a defined class label from the database, and input the sample data without a defined class label into the liquid leakage detection model for testing to update the model parameters of the liquid leakage detection model.

[0031] Optionally, after inputting the real-time acquired signal data into the liquid leakage detection model, the method for optimizing the liquid leakage detection model further includes:

[0032] Input the signal data obtained in real time into the liquid leakage detection model;

[0033] If the liquid leakage detection model outputs an anomaly, extract a real-time feature vector including signal strength and phase from the signal data collected in real time;

[0034] Store the real-time feature vector in the database and update the database by defining a class label for the real-time feature vector;

[0035] Input the real-time feature vector with a defined class label into the liquid leakage detection model, update the model parameters of the liquid leakage detection model, and obtain the optimized liquid leakage detection model.

[0036] Optionally, the liquid leakage detection method further includes a method for leakage warning after the liquid leakage detection model outputs a liquid leakage detection result, including:

[0037] Generate a liquid leakage detection report according to the liquid leakage detection result output by the liquid leakage detection model;

[0038] In response to an abnormal leakage judged in the liquid leakage detection report, trigger a warning.

[0039] The present invention also discloses a detection system, which adopts the above-mentioned RFID-based liquid leakage detection method, and is characterized in that the detection system includes:

[0040] A radio frequency signal processing module, configured to send a radio frequency signal to an RFID liquid-resistant tag attached to a target object by using an RFID reader and obtain signal data backscattered after the RFID liquid-resistant tag is activated by the radio frequency signal;

[0041] A feature extraction module, configured to select part of the obtained signal data as sample data, extract a feature vector including signal strength and phase from each sample data, and store the extracted feature vector in a database;

[0042] A label definition module, configured to define a class label including a leakage mode, a leakage liquid type, and a leakage volume for each feature vector in the database;

[0043] A model detection module, configured to construct a liquid leakage detection model according to the feature vectors with defined class labels in the database, input the signal data obtained in real time into the liquid leakage detection model, and output a liquid leakage detection result including the leakage situation, leakage location, leakage mode, leakage liquid type, and leakage volume of the target object.

[0044] Optionally, the radio frequency signal processing module is connected to a reading and writing device, which includes a reading and writing unit and a detection unit. The reading and writing unit includes an RFID reader and an antenna electrically connected to the RFID reader. The detection unit includes a liquid absorption medium and two RFID liquid-resistant tags. The liquid absorption medium includes a first medium body and a second medium body integrally formed side by side, and the shapes of the first medium body and the second medium body are fitted to the surface to be detected of the target object;

[0045] A partition layer is provided between the first medium body and the second medium body. One of the RFID liquid-resistant tags is arranged inside the first medium body, and the other RFID liquid-resistant tag is arranged on the outer side wall of the second medium body.

[0046] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. The computer program, when executed by a processor, implements the steps of the above-mentioned RFID-based liquid leakage detection method.

[0047] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. The computer program, when executed by the processor, implements the steps of the above-mentioned RFID-based liquid leakage detection method.

[0048] Compared with the prior art, the beneficial effects of the RFID-based liquid leakage detection method and detection system provided by the embodiments of the present invention are as follows:

[0049] By using RFID technology, a signal data set backscattered by an RFID liquid-resistant tag attached to a potential leakage area of a target object after being activated by a radio frequency signal is obtained, and a feature vector including signal intensity and phase is extracted from the obtained signal data. By defining a class label including leakage mode, leakage liquid type, and leakage volume for each feature vector, a liquid leakage detection model is trained and constructed. Thus, the real-time obtained signal data set is detected by the liquid leakage detection model to identify the leakage situation, leakage location, leakage mode, leakage liquid type, and leakage volume of the target object, so as to achieve high-precision dynamic monitoring of the liquid leakage situation of the target object, with the advantages of strong real-time performance and wide application range. Description of the Drawings

[0050] The technical solutions of the present invention will be further described in detail below in conjunction with the drawings and embodiments. In the drawings:

[0051] Figure 1 It is a schematic block diagram of the steps of the RFID-based liquid leakage detection method provided by the embodiments of the present invention;

[0052] Figure 2 Schematic diagram of the principle of the radio frequency signal processing module provided by the embodiment of the present invention;

[0053] Figure 3 Schematic diagram of the structure of the liquid absorption medium provided by the embodiment of the present invention;

[0054] Figure 4 Line graph showing the change of signal intensity value when different volumes of water leak provided by the embodiment of the present invention;

[0055] Figure 5 Line graph showing the change of signal intensity value when different volumes of oil leak provided by the embodiment of the present invention;

[0056] Figure 6 Line graph showing the change of signal intensity value when different volumes of milk leak provided by the embodiment of the present invention.

[0057] Each reference numeral in the figure is:

[0058] 1, RFID reader / writer; 2, antenna; 3, liquid absorption medium; 31, first medium body; 32, second medium body; 33, isolation layer; 4, RFID liquid-resistant tag; 5, host computer software. Detailed implementation manners

[0059] It should be noted that, without conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. Now, with reference to the accompanying drawings, the preferred embodiments of the present invention will be described in detail.

[0060] The present invention discloses a liquid leakage detection method based on RFID, as Figure 1 shown, including:

[0061] S1. Use the RFID reader / writer 1 to send a radio frequency signal to the RFID liquid-resistant tag 4 attached to the target object, and obtain the signal data backscattered by the RFID liquid-resistant tag 4 after being activated by the radio frequency signal;

[0062] S2. Select part of the obtained signal data as sample data, extract the feature vectors including signal intensity and phase from each sample data, and store the extracted feature vectors in the database;

[0063] S3. Define a class label including the leakage mode, the type of leakage liquid, and the leakage volume for each feature vector in the database;

[0064] S4. Construct a liquid leakage detection model according to the feature vectors defined with class labels in the database, input the signal data obtained in real time into the liquid leakage detection model, and output the liquid leakage detection result including the leakage situation, leakage position, leakage mode, type of leakage liquid, and leakage volume of the target object.

[0065] Through the implementation of the above embodiments of the liquid leakage detection method, based on the RFID technology, the RFID reader 1 is used to obtain the signal data backscattered after the RFID liquid-resistant tag 4 attached to the potential leakage area of the target object is activated by the radio frequency signal, and a data set is formed. Feature vectors including signal strength and phase are extracted from the obtained signal data. Among them, the obtained signal data set usually includes characteristic information such as time stamps, signal strength, phase, and Doppler frequency. The signal strength is used to quickly judge whether leakage occurs, and the phase is used to accurately locate the leakage position and analyze the leakage characteristics. By extracting the signal strength and phase, the feature vectors including signal strength and phase are jointly analyzed to realize the all-round monitoring of the liquid leakage situation. Then, by defining a class label including the leakage mode, the type of leaked liquid, and the leakage volume for each feature vector, the feature vectors with defined class labels are used to train and construct a liquid leakage detection model. Thus, the liquid leakage detection model can detect the real-time obtained signal data set to accurately identify the leakage situation, leakage position, leakage mode, type of leaked liquid, and leakage volume of the target object, so as to realize the high-precision dynamic monitoring of the liquid leakage situation of the target object, with the advantages of strong real-time performance and wide application range, and can be widely applied in fields such as chemical industry, environmental protection, and industrial safety. Among them, the leakage mode preferably includes internal leakage and external immersion of the target object; the type of leaked liquid is applicable to any non-corrosive liquid; the leakage volume can be classified as a large amount, a small amount, or a normal range according to the actual situation.

[0066] Further, the RFID reader 1 is used to send a radio frequency signal to the RFID liquid-resistant tag 4 attached to the target object, and the signal data backscattered after the RFID liquid-resistant tag 4 is activated by the radio frequency signal is obtained, including:

[0067] Based on the RFID liquid-resistant tags 4 pre-arranged at different positions on the target object, the RFID reader 1 sends a radio frequency signal to the RFID liquid-resistant tag 4 at a set frequency and constant power to activate the RFID liquid-resistant tag 4;

[0068] When the RFID liquid-resistant tag 4 is activated by radio frequency, the RFID reader 1 real-time collects the signal data sets backscattered by each RFID liquid-resistant tag 4.

[0069] By implementing the above embodiments of the liquid leakage detection method, radio frequency signals are sent to the RFID liquid-resistant tags 4 attached to different positions of the potential leakage area of the target object, which are used to activate the RFID liquid-resistant tags 4 and provide electromagnetic energy for them, enabling the RFID liquid-resistant tags 4 to reflect signals. Thus, accurate liquid leakage conditions can be monitored and obtained based on multiple positions in the potential leakage area of the target object. Among them, sending radio frequency signals at a set frequency helps the RFID liquid-resistant tags 4 to respond within a specific frequency range, thereby reducing interference and improving positioning accuracy. Since the radio frequency signals are sent in real time and the response of the RFID liquid-resistant tags 4 is also in real time, the liquid leakage of the target object can be monitored in real time, enabling a rapid response. And sending radio frequency signals with a constant power can ensure that the signals are evenly propagated within a predetermined coverage range, reduce signal attenuation, and improve the reliability of detection.

[0070] Further, before storing the extracted feature vectors in the database, there is also a method for preprocessing the feature vectors, including:

[0071] According to the feature vectors containing signal strength and phase extracted from the sample data, the scales of the feature vectors are standardized and unified to obtain standardized feature data. The functional expression of the standardization is:

[0072]

[0073] In the formula, is the standard value of the feature vector, representing the multiple of the standard deviation by which the value of the feature vector deviates from the mean, is the value of the feature vector, is the mean of the feature vector, is the standard deviation of the feature vector;

[0074] Perform mean value filling processing on the missing values of the standardized feature data to obtain the preprocessed feature data. The functional expression of the mean value filling is:

[0075]

[0076] In the formula, is the mean value of the non-missing values, is the non-missing value of the i-th feature data, and n is the number of non-missing values;

[0077] Judge whether there are outliers in the preprocessed feature data according to the preset conditions. The functional expression of the preset conditions is:

[0078]

[0079] In the formula, k is a conditional coefficient used to determine whether the feature data is considered an outlier, that is, to define the distance between the feature data and the mean value, in units of standard deviation, and preferably takes a value of 2 or 3;

[0080] If there are outliers in the preprocessed feature data, the feature data with outliers is preprocessed again.

[0081] Through the implementation of the above embodiments of the liquid leakage detection method, the scales of the signal strength and the phase can be unified, which can ensure that the importance of different feature vectors is fairly treated in the subsequent model training process, that is, it can prevent some feature vectors from having a disproportionate impact on the model training due to their large magnitudes. Secondly, using the average value to fill the missing values in the feature data can reduce the loss of information in the feature data set, which helps to maintain the integrity and continuity of the feature data set. When there are outliers in the feature data set, by identifying and removing them, the detection accuracy of the model training can be improved. Thus, the signal strength, phase, etc. are stored in the database in a structured form through data preprocessing, providing high-quality data input for the subsequent liquid leakage detection model. Preferably, cleaning, denoising, and normalization processing of the feature vectors can also be considered.

[0082] Further, a liquid leakage detection model is constructed according to the feature vectors defined with class labels in the database, including:

[0083] A training set is established according to the preprocessed feature data, and the functional expression of the training set is:

[0084]

[0085] In the formula, is the signal strength and phase after preprocessing of the nth sample data, is the class label including the leakage mode, the type of leaked liquid, and the leakage volume defined for the nth sample data;

[0086] The KNN algorithm is selected to establish a basic classification model, and the training set is input into the basic classification model for training to obtain a liquid leakage detection model;

[0087] A sample data without a defined class label is selected from the database, and the sample data without a defined class label is input into the liquid leakage detection model for testing to update the model parameters of the liquid leakage detection model.

[0088] Through the implementation of the embodiments of the above liquid leakage detection method, the K-Nearest Neighbors (KNN) algorithm determines new data points belonging to a certain category by measuring the distances between different feature values. That is, for sample data with undefined class labels, calculate its Euclidean distance from all sample data in the training set, and sort all the calculated distances in ascending order. Select K nearest neighbor points, that is, select the K samples with the smallest distances from the sorted distances, and set these samples as the K nearest neighbor points of the sample data with undefined class labels. Based on the selected K nearest neighbor points for classification decision-making, the detection category of the sample with undefined class labels can be selected by counting the occurrence times of each class label among the K nearest neighbors and choosing the class label with the most occurrences.

[0089] Further, after inputting the signal data obtained in real time into the liquid leakage detection model, there is also a method for optimizing the liquid leakage detection model, including:

[0090] Input the signal data obtained in real time into the liquid leakage detection model;

[0091] If the liquid leakage detection model outputs an anomaly, extract the real-time feature vector containing signal intensity and phase from the signal data set collected in real time;

[0092] Store the real-time feature vector in the database and define a class label for the real-time feature vector to update the database;

[0093] Input the real-time feature vector with a defined class label into the liquid leakage detection model, update the model parameters of the liquid leakage detection model, and obtain an optimized liquid leakage detection model.

[0094] Through the implementation of the embodiments of the above liquid leakage detection method, by obtaining the signal data backscattered by the target object in real time, the real-time performance and response speed of liquid leakage detection can be improved. When the detection model outputs an anomaly, the liquid leakage detection model can automatically extract the real-time feature vector. By storing the real-time feature vector in the database and updating the class label, the system can continuously accumulate data and provide support for the continuous optimization of the model. By inputting the real-time feature vector with a defined class label into the model for training, the model parameters are updated, thereby improving the accuracy and reliability of the liquid leakage detection model. Furthermore, by real-time updating and optimizing the model, false alarms and missed detections can be reduced, and the detection accuracy and response speed of the liquid leakage detection model can be improved to adapt to changes in different environments and leakage conditions.

[0095] Further, the liquid leakage detection method also includes a method for leakage warning after the liquid leakage detection model outputs the liquid leakage detection result, including:

[0096] Generate a liquid leakage detection report based on the liquid leakage detection result output by the liquid leakage detection model;

[0097] In response to an abnormal leakage determined in the liquid leakage detection report, trigger an early warning.

[0098] Through the implementation of the above liquid leakage detection method embodiments, the detection report can be used to detail the specific information of the liquid leakage, including key data such as the occurrence situation, location, leakage mode, liquid type, and leakage volume of the leakage. And when an abnormal leakage is detected, the system can trigger the early warning mechanism in various ways, such as audible and visual warnings, visual alarms, or information emails and other early warning notifications.

[0099] The present invention also discloses a detection system, which adopts the above RFID-based liquid leakage detection method, and is characterized in that the detection system includes:

[0100] A radio frequency signal processing module, configured to send a radio frequency signal to the RFID liquid-resistant tag 4 attached to the target object by using the RFID reader 1, and obtain the signal data backscattered by the RFID liquid-resistant tag 4 after being activated by the radio frequency signal;

[0101] A feature extraction module, configured to select part of the obtained signal data as sample data, extract feature vectors including signal strength and phase from each sample data, and store the extracted feature vectors in a database;

[0102] A tag definition module, configured to define category tags including leakage mode, leakage liquid type, and leakage volume for each feature vector in the database;

[0103] A model detection module, configured to construct a liquid leakage detection model according to the feature vectors defined with category tags in the database, input the real-time obtained signal data into the liquid leakage detection model, and output a liquid leakage detection result including the leakage situation, leakage location, leakage mode, leakage liquid type, and leakage volume of the target object.

[0104] Further, as shown in Figure 2 and Figure 3 The radio frequency signal processing module is connected to a reading and writing device, the reading and writing device includes a reading and writing unit and a detection unit, the reading and writing unit includes the RFID reader 1 and an antenna 2 electrically connected to the RFID reader 1, the detection unit includes a liquid absorption medium 3 and two RFID liquid-resistant tags 4, the liquid absorption medium 3 includes a first medium body 31 and a second medium body 32 integrally formed side by side, and the shapes of the first medium body 31 and the second medium body 32 are adapted to fit the surface to be detected of the target object;

[0105] A separation layer 33 is provided between the first dielectric body 31 and the second dielectric body 32. One of the RFID liquid-resistant tags 4 is disposed inside the first dielectric body 31, and the other RFID liquid-resistant tag 4 is disposed on the outer sidewall of the second dielectric body 32.

[0106] Through the implementation of the above-described embodiment of the detection system, its hardware part is composed of a radio frequency signal processing module, and software parts such as a feature extraction module, a tag definition module, and a model detection module perform data processing by a computing server and are displayed in real time through a host computer software 5. The radio frequency signal processing module is composed of a liquid absorption medium 3 and two RFID liquid-resistant tags 4. One of the RFID liquid-resistant tags 4 is disposed inside the first dielectric body 31, and the other RFID liquid-resistant tag 4 is disposed on the outer sidewall of the second dielectric body 32 to obtain multi-dimensional signal features. In addition, a separation layer 33 is added to the dielectric body between the two RFID liquid-resistant tags 4, which can be used to block the diffusion of the liquid-impregnated side to the other side, avoid the liquid from affecting the signal judgment of the other RFID liquid-resistant tag 4, ensure that the two RFID liquid-resistant tags 4 independently reflect the leakage state, and improve the detection accuracy and the reliability of the judgment. Preferably, at least one RFID liquid-resistant tag 4 is respectively provided on the first dielectric body 31 and the second dielectric body 32.

[0107] Preferably, the liquid absorption medium 3 is selected as a medium that matches the material and specifications of the target object, and is used to closely adhere to the area where leakage may occur on the target object, such as joints, valve connections, or pipe joints. The liquid absorption medium 3 needs to ensure good adhesion to the container surface, and can be fixed by magnetic attraction, adhesive, or mechanical fixture. At the same time, the antenna direction of the RFID liquid-resistant tag 4 is adjusted to optimize the signal reflection effect. During the placement of the liquid absorption medium 3, the signal strength and reflection effect need to be tested to ensure that the radio frequency signal processing module functions properly and adapts to the liquid characteristics and environmental conditions.

[0108] According to the above-described RFID-based liquid leakage detection method and detection system, the present invention is further illustrated by specific embodiments:

[0109] In the laboratory scenario, the detection of the liquid leakage state of the beaker is selected. The radio frequency signal processing module consists of double-layer gauze and two RFID liquid-resistant tags. One RFID liquid-resistant tag is embedded on the left side inside the gauze, and another RFID liquid-resistant tag is attached to the outside on the right side. A partition layer is set between the two RFID liquid-resistant tags to prevent the liquid from contaminating each other in their respective areas, thus affecting the accuracy of signal feature acquisition and judgment. During the detection process, an RFID reader and a 9dBi gain antenna are selected. At a frequency of 920.625 MHz, a radio frequency signal is sent with a transmission power of 27 dBm, and the reflected signals of the two RFID liquid-resistant tags are read. The RFID reader is connected to the host computer software through a serial port, and information such as the signal strength and phase collected is transmitted to the host computer software in real time for data analysis and processing.

[0110] Among them, the feature data after preprocessing is used to train the liquid leakage detection model. The data sources include the data collected when water, oil, and milk are contained in the beaker container. The training set of the liquid leakage detection model is shown in Table 1, the training part data set table:

[0111] Table 1 Training part data set table

[0112]

[0113] Among them, the leakage methods are divided into internal leakage and external contamination. Each method is divided into a small amount, a large amount, a large quantity, and no leakage according to the different leakage volumes.

[0114] The broken line graph of the change in the signal strength indicator (Received Signal Strength Indicator, RSSI) under different liquid leakage volumes is as Figures 4 - 6 shown. It can be seen from the figure that in the scenario of liquid leakage detection, there are three types of liquids: water, milk, and oil. Each liquid's leakage method includes three categories: no leakage, internal leakage, and external contamination. Among them, no leakage does not involve the judgment of leakage volume and classification; while internal leakage and external contamination are further divided into three situations: a small amount, a large amount, and a large quantity. Considering all possible situations, there are a total of 21 groups of liquid leakage states. The different leakage states are used as the output results of the liquid leakage detection model. When the signal strength read by RFID reader 1 from the radio frequency signal processing module is -45 dBm, the output result of the liquid leakage detection model is: the liquid type is oil, belonging to external large-scale contamination; when the signal strength is -50 dBm, the output result of the liquid leakage detection model is: the liquid type is milk, belonging to internal small leakage.

[0115] The present invention also discloses a computer-readable storage medium, on which a computer program is stored. The computer program, when executed by a processor, implements the steps of the above-mentioned RFID-based liquid leakage detection method.

[0116] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the above-mentioned RFID-based liquid leakage detection method are implemented.

[0117] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to specific embodiments. It should be understood that each flow and / or block in the flowchart and / or block diagram, and the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the functions specified in Figure 1 one or more of the flows Figure 1 one or more of the blocks.

[0118] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the functions specified in Figure 1 one or more of the flows Figure 1 one or more of the blocks.

[0119] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in Figure 1 one or more of the flows Figure 1 one or more of the blocks.

[0120] It should be understood that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. For those skilled in the art, the technical solutions described in the above embodiments can be modified, or some of the technical features can be equivalently replaced; and all such modifications and replacements should fall within the protection scope of the present invention.

Claims

1. A liquid leakage detection method based on RFID, characterized in that, The RFID-based liquid leakage detection method includes: Using an RFID reader to send a radio frequency signal to an RFID liquid-resistant tag attached to a target object, and acquiring the signal data backscattered after the RFID liquid-resistant tag is activated by the radio frequency signal; Selecting a part of the acquired signal data as sample data, extracting feature vectors including signal strength and phase from each of the sample data, and storing the extracted feature vectors in a database; Defining a class label including leakage mode, leakage liquid type, and leakage volume for each feature vector in the database; Constructing a liquid leakage detection model based on the feature vectors with class labels defined in the database, including: Establishing a training set according to the preprocessed feature data, and the functional expression of the training set is: In the formula, is the signal strength and phase after preprocessing of the nth sample data, is the class label defining the leakage mode, leakage liquid type, and leakage volume for the nth sample data; Selecting the KNN algorithm to establish a basic classification model, and inputting the training set into the basic classification model for training to obtain the liquid leakage detection model; Selecting a sample data without a defined class label from the database, inputting the sample data without a defined class label into the liquid leakage detection model for testing, and updating the model parameters of the liquid leakage detection model; Inputting the real-time acquired signal data into the liquid leakage detection model, and optimizing the liquid leakage detection model to output a liquid leakage detection result including the leakage situation, leakage location, leakage mode, leakage liquid type, and leakage volume of the target object; Inputting the real-time acquired signal data into the liquid leakage detection model; If the liquid leakage detection model outputs an anomaly, extracting a real-time feature vector including signal strength and phase from the real-time acquired signal data; Storing the real-time feature vector in the database, and updating the database by defining a class label for the real-time feature vector; Inputting the real-time feature vector with a defined class label into the liquid leakage detection model, updating the model parameters of the liquid leakage detection model, and obtaining the optimized liquid leakage detection model.

2. The RFID-based liquid leakage detection method according to claim 1, characterized in that The step of using an RFID reader to send a radio frequency signal to an RFID liquid-resistant tag attached to a target object and acquiring the signal data backscattered after the RFID liquid-resistant tag is activated by the radio frequency signal includes: Based on RFID liquid-resistant tags pre-arranged at different positions on the target object, the RFID reader sends a radio frequency signal to the RFID liquid-resistant tag at a set frequency and constant power to activate the RFID liquid-resistant tag; When the RFID liquid-resistant tag is activated by radio frequency, the RFID reader real-time collects the signal data backscattered by each RFID liquid-resistant tag.

3. The RFID-based liquid leakage detection method according to claim 1, characterized in that, Before storing the extracted feature vectors in the database, it also includes a method for preprocessing the feature vectors, including: According to the feature vectors including signal strength and phase extracted from the sample data, normalizing the scale of the feature vectors to obtain normalized feature data, and the functional expression of normalization is: In the formula, is the standard value of the eigenvector, is the value of the eigenvector, is the mean value of the eigenvector, is the standard deviation of the eigenvector; Fill in the missing values of the standardized feature data with the average value to obtain the preprocessed feature data. The functional expression for the average value filling is as follows: Wherein, is the average value of non-missing values, is the non-missing value of the i-th feature data, and n is the number of non-missing values; Judge whether there are outliers in the preprocessed feature data according to the preset conditions. The functional expression for the preset conditions is as follows: In the formula, k is the condition coefficient; If there are outliers in the preprocessed feature data, preprocess the feature data with outliers again.

4. The RFID-based liquid leakage detection method according to claim 1, characterized in that The liquid leakage detection method further includes a method for leakage warning after the liquid leakage detection model outputs a liquid leakage detection result, including: Generate a liquid leakage detection report according to the liquid leakage detection result output by the liquid leakage detection model; In response to the abnormal leakage judged in the liquid leakage detection report, trigger a warning.

5. A detection system that uses the RFID-based liquid leakage detection method according to any one of claims 1-4, characterized in that, The detection system includes: A radio frequency signal processing module, configured to send a radio frequency signal to an RFID liquid-resistant tag attached to a target object by using an RFID reader and obtain signal data backscattered by the RFID liquid-resistant tag after being activated by the radio frequency signal; A feature extraction module, configured to select part of the obtained signal data as sample data, extract feature vectors including signal strength and phase from each of the sample data, and store the extracted feature vectors in a database; A tag definition module, configured to define class tags including leakage mode, leakage liquid type, and leakage volume for each feature vector in the database; A model detection module, configured to construct a liquid leakage detection model according to the feature vectors defined with class tags in the database, input the real-time obtained signal data into the liquid leakage detection model, and output a liquid leakage detection result including the leakage situation, leakage location, leakage mode, leakage liquid type, and leakage volume of the target object.

6. The detection system according to claim 5, wherein: The radio frequency signal processing module is connected to a reading and writing device. The reading and writing device includes a reading and writing unit and a detection unit. The reading and writing unit includes an RFID reader and an antenna electrically connected to the RFID reader. The detection unit includes a liquid absorption medium and two RFID liquid-resistant tags. The liquid absorption medium includes a first medium body and a second medium body integrally formed side by side, and the shapes of the first medium body and the second medium body are fitted to the surface to be detected of the target object; A partition layer is provided between the first medium body and the second medium body. One of the RFID liquid-resistant tags is arranged inside the first medium body, and the other RFID liquid-resistant tag is arranged on the outer side wall of the second medium body.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the RFID-based liquid leakage detection method according to any one of claims 1-4.

8. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the RFID-based liquid leakage detection method according to any one of claims 1-4.

Images