CN111753686A - CSI-based people number identification method, device, equipment and computer storage medium - Google Patents
CSI-based people number identification method, device, equipment and computer storage medium Download PDFInfo
- Publication number
- CN111753686A CN111753686A CN202010533923.9A CN202010533923A CN111753686A CN 111753686 A CN111753686 A CN 111753686A CN 202010533923 A CN202010533923 A CN 202010533923A CN 111753686 A CN111753686 A CN 111753686A
- Authority
- CN
- China
- Prior art keywords
- data
- phase difference
- amplitude
- target
- channel state
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 42
- 239000013598 vector Substances 0.000 claims abstract description 111
- 239000011159 matrix material Substances 0.000 claims description 60
- 238000012549 training Methods 0.000 claims description 57
- 230000006870 function Effects 0.000 claims description 28
- 238000012545 processing Methods 0.000 claims description 26
- 238000000605 extraction Methods 0.000 claims description 23
- 238000007781 pre-processing Methods 0.000 claims description 19
- 238000001914 filtration Methods 0.000 claims description 15
- 238000000513 principal component analysis Methods 0.000 claims description 12
- 230000017105 transposition Effects 0.000 claims description 11
- 230000003247 decreasing effect Effects 0.000 claims description 5
- 238000012706 support-vector machine Methods 0.000 description 11
- 238000001514 detection method Methods 0.000 description 8
- 238000004891 communication Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 6
- 230000002159 abnormal effect Effects 0.000 description 5
- 238000011160 research Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000007704 transition Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 230000007613 environmental effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005562 fading Methods 0.000 description 1
- 230000009545 invasion Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000011218 segmentation Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 238000000844 transformation Methods 0.000 description 1
- 230000001960 triggered effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a method, a device and equipment for identifying the number of people based on CSI (channel state information), and a computer storage medium, wherein the method comprises the following steps: acquiring channel state information data in a scene to be detected; extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector; inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier; and determining the number information under the scene to be detected based on the number prediction label. Therefore, the amplitude characteristic vector and the phase difference characteristic vector of the channel state information data are used as characteristic information, and the people number identification accuracy based on the channel state information is improved.
Description
Technical Field
The invention relates to the technical field of signal processing, in particular to a method, a device and equipment for identifying people based on CSI (channel state information), and a computer storage medium.
Background
People number identification is a process for estimating the number of people in a specific environment, and is the basis of important tasks such as modern security, collection of business information, intelligent management and the like. People number recognition also plays an important role in safety management, and people number recognition technology can provide reference for control departments in places where people stream density needs to be controlled. The invention is particularly applied to an underground corridor with poor personnel mobility and fewer people. In such underground corridors, if personnel safety problems occur, rescue cannot be carried out in time. Under general conditions, all personnel in the underground corridor are in the coverage area of the number-of-people identification signal during operation, so that once the number of people is reduced, an alarm is triggered to avoid potential safety hazards, and then whether the personnel lacking in the underground corridor are in dangerous conditions or not can be checked.
At present, due to the common deployment of wireless networks, RSSI (received signal strength indicator) of WIFI signals is largely applied to this research, however, multipath fading and an overlay effect after multipath propagation of signals are measured by RSSI, and the effect of people number identification based on RSSI is limited, whereas the current Intel5300 commercial network card can acquire Channel State Information (CSI) in an MIMO system under the 802.11 standard, the channel state information is an estimation of channel characteristics of a communication link, the channel state information acquired from the MIMO system describes the measured value of amplitude and phase of each subcarrier, and compared with the above, the channel state information has a more refined expression in describing infinite signal multipath effect, and is adopted by more and more researchers, however, the accuracy of the people number identification based on the channel state information still needs to be improved.
Disclosure of Invention
The invention provides a method, a device and equipment for identifying the number of people based on CSI (channel state information), and a computer storage medium, and aims to solve the technical problem that the accuracy rate of identifying the number of people based on the CSI is not high at present.
In order to achieve the above object, the present invention provides a CSI-based person number recognition method, including:
acquiring channel state information data in a scene to be detected;
extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector;
inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier;
and determining the number information under the scene to be detected based on the number prediction label.
Preferably, amplitude data and phase data corresponding to the channel state information data are acquired;
preprocessing the amplitude data to obtain target amplitude data, and performing feature extraction on the target amplitude data to obtain an amplitude feature vector;
and preprocessing the phase data to obtain target phase difference data, and performing feature extraction on the target phase difference data to obtain a phase difference feature vector.
Preferably, the amplitude data is subjected to butterworth filtering processing to obtain target amplitude data;
dividing the target amplitude data into a preset number of short sequence amplitude data by using a time window;
solving first-order differential amplitude data corresponding to the short sequence amplitude data in each time window;
performing principal component analysis on the first-order difference amplitude data to obtain a first amplitude principal component and a second amplitude principal component corresponding to the short sequence amplitude data in each time window;
and obtaining an amplitude characteristic vector according to the first amplitude principal component and the second amplitude principal component.
Preferably, an amplitude matrix corresponding to the first amplitude principal component and the second amplitude principal component is constructed;
obtaining an amplitude transpose corresponding to the amplitude matrix;
acquiring a target amplitude matrix based on the amplitude transposition and the amplitude matrix;
and obtaining a magnitude eigenvector based on the target magnitude matrix.
Preferably, the phase data is sequentially subjected to phase compensation and phase linearization processing to obtain linear phase data;
determining phase difference data corresponding to the linear phase data, and performing Hampel filtering processing on the phase difference data to obtain target phase difference data;
dividing the target phase difference data into a preset number of short sequence phase difference data by using a time window;
solving first-order differential phase difference data corresponding to the short-sequence phase difference data in each time window;
performing principal component analysis on the first-order differential phase difference data to obtain a first phase difference principal component and a second phase difference principal component corresponding to the short-sequence phase difference data in each time window;
and acquiring a phase difference characteristic vector according to the first phase difference main component and the second phase difference main component.
Preferably, a phase difference matrix corresponding to the first phase difference principal component and the second phase difference principal component is constructed;
acquiring a phase difference transpose corresponding to the phase difference matrix;
acquiring a target phase difference matrix based on the phase difference transposition and the phase difference matrix;
and acquiring a phase difference characteristic vector based on the target phase difference matrix.
Preferably, acquiring channel state information training data, and determining a real person number label corresponding to the channel state information training data;
acquiring a predictive population label corresponding to the channel state information training data by using an initial classifier;
calculating a loss function corresponding to the initial classifier based on the predictive population number label and the real population number label corresponding to the channel state information training data;
updating parameters of the initial classifier in a gradient decreasing manner based on the loss function;
and if the loss function reaches the convergence condition, stopping updating, and storing the target parameters corresponding to the loss function reaching the convergence condition to obtain the target classifier.
In addition, to achieve the above object, the present invention provides a CSI-based person number recognition apparatus, including:
the acquisition module is used for acquiring channel state information data in a scene to be detected;
the extraction module is used for extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector;
the output module is used for inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier;
and the determining module is used for determining the number information under the scene to be detected based on the number prediction label.
In addition, to achieve the above object, the present invention also provides a CSI-based person number recognition apparatus, which includes a processor, a memory, and a CSI-based person number recognition program stored in the memory, and when the CSI-based person number recognition program is executed by the processor, the steps of the CSI-based person number recognition method as described above are implemented.
In addition, to achieve the above object, the present invention also provides a computer storage medium having stored thereon a CSI-based person number recognition program, which when executed by a processor, implements the steps of the CSI-based person number recognition method as described above.
Compared with the prior art, the invention discloses a method, a device, equipment and a computer storage medium for identifying the number of people based on CSI, which are used for acquiring channel state information data under a scene to be detected, performing feature extraction on the channel state information data to acquire an amplitude feature vector and a phase difference feature vector, inputting the amplitude feature vector and the phase difference feature vector into a target classifier, outputting a corresponding number of people prediction label by the target classifier, and determining the number of people under the scene to be detected based on the number of people prediction label. Therefore, the amplitude characteristic vector and the phase difference characteristic vector of the channel state information data are used as characteristic information, and the people number identification accuracy based on the channel state information is improved.
Drawings
Fig. 1 is a hardware configuration diagram of a CSI-based person number recognition apparatus according to embodiments of the present invention;
FIG. 2 is a schematic flow chart of a first embodiment of the CSI-based person number identification method according to the present invention;
FIG. 3 is a flowchart illustrating a second embodiment of the CSI-based person number identification method according to the present invention;
fig. 4 is a functional block diagram of a first embodiment of the CSI-based person number recognition apparatus according to the present invention.
The implementation, functional features and advantages of the objects of the present invention will be further explained with reference to the accompanying drawings.
Detailed Description
It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Referring to fig. 1, fig. 1 is a hardware configuration diagram of a CSI-based person number recognition apparatus according to embodiments of the present invention. In this embodiment of the present invention, the CSI-based person identification apparatus may include a processor 1001 (e.g., a Central Processing Unit, CPU), a communication bus 1002, an input port 1003, an output port 1004, and a memory 1005. The communication bus 1002 is used for realizing connection communication among the components; the input port 1003 is used for data input; the output port 1004 is used for data output, the memory 1005 may be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as a magnetic disk memory, and the memory 1005 may optionally be a storage device independent of the processor 1001. Those skilled in the art will appreciate that the hardware configuration depicted in FIG. 1 is not intended to be limiting of the present invention, and may include more or less components than those shown, or some components in combination, or a different arrangement of components.
With continued reference to fig. 1, the memory 1005 of fig. 1, which is one type of readable storage medium, may include an operating system, a network communication module, an application module, and a CSI-based people recognition program. In fig. 1, the network communication module is mainly used for connecting to a server and performing data communication with the server; and the processor 1001 may call the CSI-based person count recognition program stored in the memory 1005 and perform the following operations:
acquiring channel state information data in a scene to be detected;
extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector;
inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier;
and determining the number information under the scene to be detected based on the number prediction label.
Further, the processor 1001 may be further configured to call the CSI-based person count recognition program stored in the memory 1005, and perform the following steps:
acquiring amplitude data and phase data corresponding to the channel state information data;
preprocessing the amplitude data to obtain target amplitude data, and performing feature extraction on the target amplitude data to obtain an amplitude feature vector;
and preprocessing the phase data to obtain target phase difference data, and performing feature extraction on the target phase difference data to obtain a phase difference feature vector.
Further, the processor 1001 may be further configured to call the CSI-based person count recognition program stored in the memory 1005, and perform the following steps:
performing Butterworth filtering processing on the amplitude data to acquire target amplitude data;
dividing the target amplitude data into a preset number of short sequence amplitude data by using a time window;
solving first-order differential amplitude data corresponding to the short sequence amplitude data in each time window;
performing principal component analysis on the first-order difference amplitude data to obtain a first amplitude principal component and a second amplitude principal component corresponding to the short sequence amplitude data in each time window;
and obtaining an amplitude characteristic vector according to the first amplitude principal component and the second amplitude principal component.
Further, the processor 1001 may be further configured to call the CSI-based person count recognition program stored in the memory 1005, and perform the following steps:
constructing an amplitude matrix corresponding to the first amplitude principal component and the second amplitude principal component;
obtaining an amplitude transpose corresponding to the amplitude matrix;
acquiring a target amplitude matrix based on the amplitude transposition and the amplitude matrix;
and obtaining a magnitude eigenvector based on the target magnitude matrix.
Further, the processor 1001 may be further configured to call the CSI-based person count recognition program stored in the memory 1005, and perform the following steps:
sequentially carrying out phase compensation and phase linearization processing on the phase data to obtain linear phase data;
determining phase difference data corresponding to the linear phase data, and performing Hampel filtering processing on the phase difference data to obtain target phase difference data;
dividing the target phase difference data into a preset number of short sequence phase difference data by using a time window;
solving first-order differential phase difference data corresponding to the short-sequence phase difference data in each time window;
performing principal component analysis on the first-order differential phase difference data to obtain a first phase difference principal component and a second phase difference principal component corresponding to the short-sequence phase difference data in each time window;
and acquiring a phase difference characteristic vector according to the first phase difference main component and the second phase difference main component.
Further, the processor 1001 may be further configured to call the CSI-based person count recognition program stored in the memory 1005, and perform the following steps:
constructing a phase difference matrix corresponding to the first phase difference principal component and the second phase difference principal component;
acquiring a phase difference transpose corresponding to the phase difference matrix;
acquiring a target phase difference matrix based on the phase difference transposition and the phase difference matrix;
and acquiring a phase difference characteristic vector based on the target phase difference matrix.
Further, the processor 1001 may be further configured to call the CSI-based person count recognition program stored in the memory 1005, and perform the following steps:
acquiring channel state information training data, and determining a real person number label corresponding to the channel state information training data;
acquiring a predictive population label corresponding to the channel state information training data by using an initial classifier;
calculating a loss function corresponding to the initial classifier based on the predictive population number label and the real population number label corresponding to the channel state information training data;
updating parameters of the initial classifier in a gradient decreasing manner based on the loss function;
and if the loss function reaches the convergence condition, stopping updating, and storing the target parameters corresponding to the loss function reaching the convergence condition to obtain the target classifier.
Based on the above structure, various embodiments of the CSI-based person number recognition method of the present invention are proposed.
Referring to fig. 2, fig. 2 is a flowchart illustrating a first embodiment of the CSI-based person number recognition method according to the present invention.
In this embodiment, the CSI-based person number recognition method is applied to a CSI-based person number recognition device, and the method includes:
step S10: acquiring channel state information data in a scene to be detected;
in this embodiment, with the development of technology and the popularization of WIFI router devices, the WIFI wireless lan technology is applied in the field of intrusion detection by people, which becomes the research direction of more and more students in China and abroad, because the devices such as routers are cheap and easy to obtain, and have high popularity, and people can perform intrusion detection only by relying on the acquired WIFI signals without requiring the detected people to carry any additional devices for detection, the convenience and practicability of the technology are greatly improved, most researchers currently adopt a wireless Signal Received strength indication parameter (RSSI) as a research parameter for intrusion detection by people, however, only one piece of RSSI information in one data packet results in large research difficulty, so in this embodiment, wireless Channel State Information (CSI) is adopted to replace the traditional wireless Signal Received strength indication parameter, the problem that the traditional method depends on signal power can be solved, specifically, wireless links can be affected when people invade, and the fluctuation of amplitude and phase is reflected on CSI information, so that whether people invade can be detected through the amplitude and the phase.
In specific application, channel state information data received by multiple antennas at a receiving end is obtained, for example, channel state information data corresponding to three antennas at a wireless signal receiving end in a scene to be detected is obtained, where a transmitting end for transmitting a wireless signal and a receiving end for receiving a wireless signal are both common commercial devices, for example, a router can be used as a transmitter, a notebook can be used as a receiver, but not limited to these two devices, and optionally, an Intel5300 network card is used to obtain data of 30 subcarrier clusters in the scene to be detected, and then 30 channel state information data corresponding to the data of the 30 subcarrier clusters are obtained.
Step S20: extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector;
the channel state information includes amplitude information and phase information, wherein the amplitude information has better stability than the phase information, but the phase information is more precise than the amplitude information in disturbance of a person who is depicted, so in the embodiment, the amplitude information and the phase information corresponding to the channel state information are extracted and used as characteristic information to be researched.
It should be noted that the original channel state information data acquired from the receiving-end antenna cannot be directly used for human intrusion identification, because there are many interferences, such as environmental noise, electrical noise, and other interference of nearby WIFI signals, and these interference factors may cause severe abnormal fluctuation of the acquired original channel state information data, thereby affecting the discrimination of our intrusion detection, so that we need to perform preprocessing, such as filtering, on the acquired data information to remove noise and abnormal data points of the original channel state information data, and perform feature extraction on the preprocessed channel state information data after preprocessing.
It should be noted that, because the phase difference often has a larger variance, the phase difference is used as the feature information in this embodiment, specifically, the amplitude data and the phase data corresponding to the channel state information data are extracted, then the phase difference data corresponding to the three antennas at the receiving end is calculated, and finally, the feature extraction is performed on the amplitude data and the phase difference data respectively to obtain the amplitude feature vector and the phase difference feature vector.
Step S30: inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier;
step S40: and determining the number information under the scene to be detected based on the number prediction label.
In this embodiment, the classifier may select a Support Vector Machine (SVM) to output the people number prediction labels corresponding to the amplitude feature Vector and the phase difference feature Vector, and further, before performing the feature Vector on the amplitude feature Vector and the phase difference feature Vector input each time by using the SVM classifier, the SVM classifier needs to be trained by using training data to obtain a trained target classifier, so as to output the corresponding people number prediction labels to the amplitude feature Vector and the phase difference feature Vector input each time based on the target classifier.
Specifically, before the step of inputting the amplitude feature vector and the phase difference feature vector into the target classifier, the method further includes:
step a: acquiring channel state information training data, and determining a real person number label corresponding to the channel state information training data;
step b: acquiring a predictive population label corresponding to the channel state information training data by using an initial classifier;
step c: calculating a loss function corresponding to the initial classifier based on the predictive population number label and the real population number label corresponding to the channel state information training data;
step d: updating parameters of the initial classifier in a gradient decreasing manner based on the loss function;
step e: and if the loss function reaches the convergence condition, stopping updating, and storing the target parameters corresponding to the loss function reaching the convergence condition to obtain the target classifier.
In this step, a large amount of channel state information training data in manned scenes are prepared at random, wherein the channel state information training data in manned scenes can include channel state information training data corresponding to different numbers of testers in scenes, channel state information training data corresponding to different gender testers in scenes, and channel state information training data corresponding to different height and thinness testers in scenes, so that the diversity of the training data is enriched, and the classification accuracy of the classifier is improved.
After the channel state information training data is obtained, the number of real people labels corresponding to the channel state information training data is determined, for example, if the training data is the channel state information training data in a 3-person scene, the number of real people labels corresponding to the training data is 3, if the training data is the channel state information training data in a 9-person scene, the number of real people labels corresponding to the training data is 9, and then the channel state information training data is subjected to preprocessing and feature extraction processing. For example, amplitude training data and phase training data corresponding to channel state information training data are extracted, phase compensation and linearization are performed on the phase training data to obtain linear phase training data corresponding to the phase training data, phase difference training data corresponding to the linear phase training data is obtained, and filtering is performed on the amplitude training data and the phase difference training data to perform denoising processing on the amplitude training data and the phase difference training data. After filtering and denoising, feature extraction is carried out on the amplitude training data and the phase difference training data to obtain corresponding amplitude feature training vectors and phase difference feature training vectors, and finally an initial classifier is trained according to the amplitude feature training vectors and the phase difference feature training vectors to obtain a target classifier.
Specifically, the amplitude feature training vector and the phase difference feature training vector are input into an initial classifier, such as an initial SVM classifier, so that a corresponding number of predicted people label is output by the initial SVM classifier, and then a loss function corresponding to the initial SVM classifier is calculated according to the number of predicted people label and a corresponding real number label thereof, optionally, a cross quotient loss function corresponding to the number of predicted people label and the corresponding real number label thereof is calculated in this embodiment.
After the cross entropy loss function of the initial SVM classifier is obtained, the gradient corresponding to each parameter in the initial SVM classifier is calculated according to the cross entropy loss function, each parameter is correspondingly updated according to the gradient of each parameter, and each parameter of the initial SVM classifier is adjusted. Here, the process of updating the model parameters according to the cross entropy loss function is similar to the existing process of updating the model parameters, and details are not repeated here until the cross entropy loss function reaches the convergence condition, the updating of each parameter in the initial SVM classifier is stopped, and each corresponding parameter value when the cross entropy loss function reaches the convergence condition is confirmed and stored, so as to obtain the target classifier.
After the target classifier is obtained, the amplitude characteristic vector and the phase difference characteristic vector corresponding to the channel state information data in the scene to be detected are simultaneously input to the target classifier, the target classifier outputs corresponding people number prediction labels, the target classifier outputs several different predicted people numbers and corresponding prediction probabilities, and after the people number prediction labels corresponding to the target classifier are obtained, the predicted people number with the highest prediction probability in the output result is determined as the people number information in the scene to be detected.
According to the scheme, the channel state information data under the scene to be detected is obtained; extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector; inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier; and determining the number information under the scene to be detected based on the number prediction label. Therefore, the amplitude characteristic vector and the phase difference characteristic vector of the channel state information data are used as characteristic information, and the people number identification accuracy based on the channel state information is improved.
A third embodiment of the present invention is proposed based on the second embodiment shown in fig. 2 described above. As shown in fig. 4, fig. 4 is a flowchart illustrating a second embodiment of the CSI-based person number identification method according to the present invention.
The step of extracting the characteristics of the channel state information data to obtain the amplitude characteristic vector and the phase difference characteristic vector comprises the following steps:
step S201: acquiring amplitude data and phase data corresponding to the channel state information data;
it is understood that the channel state information includes both amplitude and phase information, and thus, after the channel state information data is acquired, amplitude data and phase data corresponding to the channel state information data are extracted.
Step S202: preprocessing the amplitude data to obtain target amplitude data, and performing feature extraction on the target amplitude data to obtain an amplitude feature vector;
it should be noted that the original channel state information data acquired from the receiving-end antenna cannot be directly used for human intrusion identification, because there are many interferences, such as environmental noise, electrical noise, and other interference of nearby WIFI signals, and these interference factors may cause severe abnormal fluctuation of the acquired original channel state information data, thereby affecting the discrimination of our intrusion detection, so we need to perform preprocessing such as filtering on the acquired data information, that is, after removing noise points and abnormal data points of the original channel state information data, perform feature extraction on the data, so as to obtain a feature vector.
Further, since the channel state information includes two kinds of information, namely amplitude and phase, and there is a certain difference between the two kinds of information, the amplitude information and the phase information corresponding to the channel state information are extracted and then processed respectively, so that the accuracy of data is improved, and further the detection result is improved.
Specifically, step S202 includes:
step S202 a: performing Butterworth filtering processing on the amplitude data to acquire target amplitude data;
step S202 b: dividing the target amplitude data into a preset number of short sequence amplitude data by using a time window;
step S202 c: solving first-order differential amplitude data corresponding to the short sequence amplitude data in each time window;
step S202 d: performing principal component analysis on the first-order difference amplitude data to obtain a first amplitude principal component and a second amplitude principal component corresponding to the short sequence amplitude data in each time window;
step S202 e: and obtaining an amplitude characteristic vector according to the first amplitude principal component and the second amplitude principal component.
In the step, after obtaining the amplitude data, performing butterworth filtering processing on the amplitude data to delete the abnormal amplitude data and obtain the target amplitude data, and further, research finds that the curve fluctuation after the first-order difference of the corresponding channel state information data is more severe than the curve fluctuation before the first-order difference when someone invades, that is, the first-order difference can amplify the influence of the person invading the channel state information data so as to detect and judge the invasion, so the embodiment performs the first-order difference processing on the amplitude data, and further, in order to improve the data processing efficiency, in the embodiment, firstly, the target amplitude data is divided into a preset number of short-sequence amplitude data by using a time window, then, the first-order difference amplitude data corresponding to the short-sequence amplitude data in each time window is simultaneously solved, then, the main component analysis dimensionality reduction is performed on the short-sequence amplitude data in each time window, and finally, obtaining an amplitude feature vector according to the first amplitude principal component and the second amplitude principal component.
Specifically, the step of obtaining the amplitude feature vector according to the first amplitude principal component and the second amplitude principal component includes:
constructing an amplitude matrix corresponding to the first amplitude principal component and the second amplitude principal component;
obtaining an amplitude transpose corresponding to the amplitude matrix;
acquiring a target amplitude matrix based on the amplitude transposition and the amplitude matrix;
and obtaining a magnitude eigenvector based on the target magnitude matrix.
In this step, after obtaining the first amplitude principal component and the second amplitude principal component, an amplitude matrix corresponding to the first amplitude principal component and the second amplitude principal component is constructed, for example: and finally, multiplying the amplitude matrix of m by n with the amplitude transpose matrix of n by m to obtain a target amplitude matrix of m, and obtaining the amplitude feature vector through the target amplitude matrix of m by m.
Step S203: and preprocessing the phase data to obtain target phase difference data, and performing feature extraction on the target phase difference data to obtain a phase difference feature vector.
In this step, in order to avoid the influence of the external interference factors on the detection result, the phase data needs to be preprocessed to obtain the target phase difference data, and then the feature extraction is performed on the target phase difference data to obtain the phase difference feature vector.
Specifically, step S203 includes:
step S203 a: sequentially carrying out phase compensation and phase linearization processing on the phase data to obtain linear phase data;
step S203 b: determining phase difference data corresponding to the linear phase data, and performing Hampel filtering processing on the phase difference data to obtain target phase difference data;
step S203 c: dividing the target phase difference data into a preset number of short sequence phase difference data by using a time window;
step S203 d: solving first-order differential phase difference data corresponding to the short-sequence phase difference data in each time window;
step S203 e: performing principal component analysis on the first-order differential phase difference data to obtain a first phase difference principal component and a second phase difference principal component corresponding to the short-sequence phase difference data in each time window;
step S203 f: and acquiring a phase difference characteristic vector according to the first phase difference main component and the second phase difference main component.
In this step, since the measurement phase pairs which are randomly distributed are obtained from the Intel5300 network card, and the CSI original phase data in the data packet are distributed in a disordered curve, it is not possible to detect disturbance generated to the wireless network environment when a person invades, that is, it is not possible to sense and detect the person by using phase information, therefore, in this embodiment, after extracting phase data, phase compensation processing is performed on the phase data to obtain complete phase data, where there are two transition directions of the original CSI phase, that is, transition from-pi to + pi and transition from + pi to-pi, and different compensations are performed on the CSI phase according to different phase transition directions, where the compensated phase size is 2 pi or-2 pi.
Further, since the phase difference tends to have a larger variance, the phase difference is used as the characteristic information in the present embodiment, so in the present embodiment, in order to conveniently and accurately obtain the characteristic information of the phase difference, after performing phase compensation, phase linearization processing is performed on the phase data to obtain linear phase data, and further more accurate phase difference data is obtained according to the linear phase data.
Further, like the amplitude data, in order to obtain features with larger differences and improve data processing efficiency, in this embodiment, after the target phase difference data is obtained, the target phase difference data is first divided into a preset number of short-sequence phase difference data by using a time window, then the first-order difference processing is performed on the preset number of short-sequence phase difference data, and finally, a phase difference feature vector is obtained based on the phase difference data after the first-order difference processing, which is not described herein again.
Further, the step of obtaining a phase difference eigenvector according to the first phase difference principal component and the second phase difference principal component includes:
constructing a phase difference matrix corresponding to the first phase difference principal component and the second phase difference principal component;
acquiring a phase difference transpose corresponding to the phase difference matrix;
acquiring a target phase difference matrix based on the phase difference transposition and the phase difference matrix;
and acquiring a phase difference characteristic vector based on the target phase difference matrix.
In this step, after the first phase difference principal component and the second phase difference principal component are acquired, a phase difference matrix corresponding to the first phase difference principal component and the second phase difference principal component is constructed, for example: and finally, multiplying the phase difference matrix of p × q with the phase difference transposition matrix of p × q to obtain a target phase difference matrix of p × q, and obtaining a phase difference characteristic vector through the target phase difference matrix of p × q.
According to the scheme, the amplitude data and the phase data corresponding to the channel state information data are obtained, the amplitude data are preprocessed to obtain target amplitude data, the target amplitude data are subjected to feature extraction to obtain the amplitude feature vector, the phase data are preprocessed to obtain the target phase difference data, the target phase difference data are subjected to feature extraction to obtain the phase difference feature vector, and therefore the number of people based on the channel state information is improved by using the amplitude feature vector and the phase difference feature vector of the channel state information data as feature information.
In addition, the embodiment also provides a device for identifying the number of people based on the CSI. Referring to fig. 4, fig. 4 is a functional block diagram of a first embodiment of the CSI-based person number recognition apparatus according to the present invention.
In this embodiment, the CSI-based person number recognition device is a virtual device, and is stored in the memory 1005 of the CSI-based person number recognition apparatus shown in fig. 1, so as to realize all functions of the CSI-based person number recognition program: the method comprises the steps of obtaining channel state information data under a scene to be detected; the device is used for extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector; the system is used for inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding people number prediction label by the target classifier; and the system is used for determining the number information under the scene to be detected based on the number prediction label.
Specifically, the CSI-based person number recognition apparatus includes:
the acquisition module 10 is configured to acquire channel state information data in a scene to be detected;
an extracting module 20, configured to perform feature extraction on the channel state information data to obtain an amplitude feature vector and a phase difference feature vector;
the output module 30 is configured to input the amplitude feature vector and the phase difference feature vector to a target classifier, and the target classifier outputs a corresponding number of people prediction labels;
and the determining module 40 is configured to determine the number information in the scene to be detected based on the number prediction tag.
Further, the extraction module comprises:
the first acquisition unit is used for acquiring amplitude data and phase data corresponding to the channel state information data;
the second acquisition unit is used for preprocessing the amplitude data to acquire target amplitude data and extracting the characteristics of the target amplitude data to acquire an amplitude characteristic vector;
and the third acquisition unit is used for preprocessing the phase data to acquire target phase difference data and extracting the characteristics of the target phase difference data to acquire a phase difference characteristic vector.
Further, the second obtaining unit further includes:
the first preprocessing subunit is used for performing Butterworth filtering processing on the amplitude data to acquire target amplitude data;
a first dividing subunit, configured to divide the target amplitude data into a preset number of short-sequence amplitude data by using a time window;
the first solving subunit is used for solving first-order differential amplitude data corresponding to the short sequence amplitude data in each time window;
the first principal component analysis subunit is used for performing principal component analysis on the first-order difference amplitude data to obtain a first amplitude principal component and a second amplitude principal component corresponding to the short sequence amplitude data in each time window;
and the first obtaining subunit is used for obtaining the amplitude characteristic vector according to the first amplitude principal component and the second amplitude principal component.
Further, the first obtaining subunit is further configured to:
constructing an amplitude matrix corresponding to the first amplitude principal component and the second amplitude principal component;
obtaining an amplitude transpose corresponding to the amplitude matrix;
acquiring a target amplitude matrix based on the amplitude transposition and the amplitude matrix;
and obtaining a magnitude eigenvector based on the target magnitude matrix.
Further, the third obtaining unit further includes:
the second preprocessing subunit is used for sequentially performing phase compensation and phase linearization on the phase data to acquire linear phase data;
the third preprocessing subunit is used for determining phase difference data corresponding to the linear phase data and performing Hampel filtering processing on the phase difference data to acquire target phase difference data;
the second segmentation subunit is used for segmenting the target phase difference data into a preset number of short sequence phase difference data by using a time window;
the second solving subunit is used for solving first-order differential phase difference data corresponding to the short-sequence phase difference data in each time window;
the second principal component analysis subunit is used for performing principal component analysis on the first-order differential phase difference data to obtain a first phase difference principal component and a second phase difference principal component corresponding to the short-sequence phase difference data in each time window;
and the second obtaining subunit is used for obtaining the phase difference characteristic vector according to the first phase difference main component and the second phase difference main component.
Further, the second obtaining subunit is further configured to:
constructing a phase difference matrix corresponding to the first phase difference principal component and the second phase difference principal component;
acquiring a phase difference transpose corresponding to the phase difference matrix;
acquiring a target phase difference matrix based on the phase difference transposition and the phase difference matrix;
and acquiring a phase difference characteristic vector based on the target phase difference matrix.
Further, the output module further includes:
the third acquisition subunit is used for acquiring channel state information training data and determining a real person number label corresponding to the channel state information training data;
the training subunit is used for acquiring a predictive population label corresponding to the channel state information training data by using an initial classifier;
the calculation subunit is used for calculating a loss function corresponding to the initial classifier based on the forecasted population number label and the real population number label corresponding to the channel state information training data;
an updating subunit, configured to update parameters of the initial classifier in a gradient decreasing manner based on the loss function;
and the fourth obtaining subunit is configured to stop updating if the loss function reaches the convergence condition, and store the target parameter corresponding to the loss function reaching the convergence condition to obtain the target classifier.
In addition, an embodiment of the present invention further provides a computer storage medium, where a CSI-based person number identification program is stored on the computer storage medium, and when the CSI-based person number identification program is executed by a processor, the steps of the CSI-based person number identification method are implemented, which are not described herein again.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or system. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other like elements in a process, method, article, or system that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the above description of the embodiments, those skilled in the art will clearly understand that the method of the above embodiments can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware, but in many cases, the former is a better implementation manner. Based on such understanding, the technical solution of the present invention may be embodied in the form of a software product, which is stored in a storage medium (e.g., ROM/RAM, magnetic disk, optical disk) as described above and includes instructions for causing a terminal device to execute the method according to the embodiments of the present invention.
The above description is only for the preferred embodiment of the present invention and is not intended to limit the scope of the present invention, and all equivalent structures or flow transformations made by the present specification and drawings, or applied directly or indirectly to other related arts, are included in the scope of the present invention.
Claims (10)
1. The CSI-based person number identification method is characterized by comprising the following steps of:
acquiring channel state information data in a scene to be detected;
extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector;
inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier;
and determining the number information under the scene to be detected based on the number prediction label.
2. The method of claim 1, wherein the step of extracting the features of the channel state information data to obtain a magnitude feature vector and a phase difference feature vector comprises:
acquiring amplitude data and phase data corresponding to the channel state information data;
preprocessing the amplitude data to obtain target amplitude data, and performing feature extraction on the target amplitude data to obtain an amplitude feature vector;
and preprocessing the phase data to obtain target phase difference data, and performing feature extraction on the target phase difference data to obtain a phase difference feature vector.
3. The method of claim 2, wherein the step of preprocessing the magnitude data to obtain target magnitude data and performing feature extraction on the target magnitude data to obtain a magnitude feature vector comprises:
performing Butterworth filtering processing on the amplitude data to acquire target amplitude data;
dividing the target amplitude data into a preset number of short sequence amplitude data by using a time window;
solving first-order differential amplitude data corresponding to the short sequence amplitude data in each time window;
performing principal component analysis on the first-order difference amplitude data to obtain a first amplitude principal component and a second amplitude principal component corresponding to the short sequence amplitude data in each time window;
and obtaining an amplitude characteristic vector according to the first amplitude principal component and the second amplitude principal component.
4. The method of claim 3, wherein the step of obtaining a magnitude eigenvector from the first and second magnitude principal components comprises:
constructing an amplitude matrix corresponding to the first amplitude principal component and the second amplitude principal component;
obtaining an amplitude transpose corresponding to the amplitude matrix;
acquiring a target amplitude matrix based on the amplitude transposition and the amplitude matrix;
and obtaining a magnitude eigenvector based on the target magnitude matrix.
5. The method of claim 2, wherein the step of preprocessing the phase data to obtain target phase difference data and performing feature extraction on the target phase difference data to obtain a phase difference feature vector comprises:
sequentially carrying out phase compensation and phase linearization processing on the phase data to obtain linear phase data;
determining phase difference data corresponding to the linear phase data, and performing Hampel filtering processing on the phase difference data to obtain target phase difference data;
dividing the target phase difference data into a preset number of short sequence phase difference data by using a time window;
solving first-order differential phase difference data corresponding to the short-sequence phase difference data in each time window;
performing principal component analysis on the first-order differential phase difference data to obtain a first phase difference principal component and a second phase difference principal component corresponding to the short-sequence phase difference data in each time window;
and acquiring a phase difference characteristic vector according to the first phase difference main component and the second phase difference main component.
6. The method according to claim 5, wherein the step of obtaining a phase difference eigenvector from the first phase difference principal component and the second phase difference principal component comprises:
constructing a phase difference matrix corresponding to the first phase difference principal component and the second phase difference principal component;
acquiring a phase difference transpose corresponding to the phase difference matrix;
acquiring a target phase difference matrix based on the phase difference transposition and the phase difference matrix;
and acquiring a phase difference characteristic vector based on the target phase difference matrix.
7. The method according to any one of claims 1 to 6, wherein before the step of inputting the magnitude and phase difference eigenvectors to the target classifier, the method further comprises:
acquiring channel state information training data, and determining a real person number label corresponding to the channel state information training data;
acquiring a predictive population label corresponding to the channel state information training data by using an initial classifier;
calculating a loss function corresponding to the initial classifier based on the predictive population number label and the real population number label corresponding to the channel state information training data;
updating parameters of the initial classifier in a gradient decreasing manner based on the loss function;
and if the loss function reaches the convergence condition, stopping updating, and storing the target parameters corresponding to the loss function reaching the convergence condition to obtain the target classifier.
8. A CSI-based person number recognition apparatus, comprising:
the acquisition module is used for acquiring channel state information data in a scene to be detected;
the extraction module is used for extracting the characteristics of the channel state information data to obtain an amplitude characteristic vector and a phase difference characteristic vector;
the output module is used for inputting the amplitude characteristic vector and the phase difference characteristic vector into a target classifier, and outputting a corresponding number of people prediction label by the target classifier;
and the determining module is used for determining the number information under the scene to be detected based on the number prediction label.
9. A CSI-based person number recognition apparatus, comprising a processor, a memory, and a CSI-based person number recognition program stored in the memory, the CSI-based person number recognition program when executed by the processor implementing the steps of the CSI-based person number recognition method as claimed in any one of claims 1 to 7.
10. A computer storage medium having a CSI-based people recognition program stored thereon, the CSI-based people recognition program when executed by a processor implementing the steps of the CSI-based people recognition method as defined in any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010533923.9A CN111753686A (en) | 2020-06-11 | 2020-06-11 | CSI-based people number identification method, device, equipment and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010533923.9A CN111753686A (en) | 2020-06-11 | 2020-06-11 | CSI-based people number identification method, device, equipment and computer storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111753686A true CN111753686A (en) | 2020-10-09 |
Family
ID=72675128
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010533923.9A Pending CN111753686A (en) | 2020-06-11 | 2020-06-11 | CSI-based people number identification method, device, equipment and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111753686A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113810500A (en) * | 2021-09-18 | 2021-12-17 | 中国银行股份有限公司 | Crowd gathering risk identification method and related equipment |
| CN114359738A (en) * | 2022-03-16 | 2022-04-15 | 深圳大学 | Cross-scene robust indoor population wireless detection method and system |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106131958A (en) * | 2016-08-09 | 2016-11-16 | 电子科技大学 | A kind of based on channel condition information with the indoor Passive Location of support vector machine |
| CN106658590A (en) * | 2016-12-28 | 2017-05-10 | 南京航空航天大学 | Design and implementation of multi-person indoor environment state monitoring system based on WiFi channel state information |
| CN107992882A (en) * | 2017-11-20 | 2018-05-04 | 电子科技大学 | A kind of occupancy statistical method based on WiFi channel condition informations and support vector machines |
| CN108198558A (en) * | 2017-12-28 | 2018-06-22 | 电子科技大学 | A kind of audio recognition method based on CSI data |
| CN108631890A (en) * | 2018-02-08 | 2018-10-09 | 中国矿业大学 | A kind of underground coal mine based on channel state information and random forest swarms into detection method |
| CN108988968A (en) * | 2018-07-27 | 2018-12-11 | 河北工程大学 | Human behavior detection method, device and terminal device |
| CN109255874A (en) * | 2018-09-19 | 2019-01-22 | 电子科技大学 | A kind of passage and number detection method based on general commercial WiFi equipment |
| CN109587645A (en) * | 2018-11-12 | 2019-04-05 | 南京邮电大学 | Personnel's recognition methods under indoor environment based on channel state information |
| EP3492945A1 (en) * | 2017-12-01 | 2019-06-05 | Origin Wireless, Inc. | Method, apparatus, and system for periodic motion detection and monitoring |
| CN109961100A (en) * | 2019-03-26 | 2019-07-02 | 北京邮电大学 | WIFI indoor occupant intrusion detection method based on Principal Component Analysis |
-
2020
- 2020-06-11 CN CN202010533923.9A patent/CN111753686A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106131958A (en) * | 2016-08-09 | 2016-11-16 | 电子科技大学 | A kind of based on channel condition information with the indoor Passive Location of support vector machine |
| CN106658590A (en) * | 2016-12-28 | 2017-05-10 | 南京航空航天大学 | Design and implementation of multi-person indoor environment state monitoring system based on WiFi channel state information |
| CN107992882A (en) * | 2017-11-20 | 2018-05-04 | 电子科技大学 | A kind of occupancy statistical method based on WiFi channel condition informations and support vector machines |
| EP3492945A1 (en) * | 2017-12-01 | 2019-06-05 | Origin Wireless, Inc. | Method, apparatus, and system for periodic motion detection and monitoring |
| CN108198558A (en) * | 2017-12-28 | 2018-06-22 | 电子科技大学 | A kind of audio recognition method based on CSI data |
| CN108631890A (en) * | 2018-02-08 | 2018-10-09 | 中国矿业大学 | A kind of underground coal mine based on channel state information and random forest swarms into detection method |
| CN108988968A (en) * | 2018-07-27 | 2018-12-11 | 河北工程大学 | Human behavior detection method, device and terminal device |
| CN109255874A (en) * | 2018-09-19 | 2019-01-22 | 电子科技大学 | A kind of passage and number detection method based on general commercial WiFi equipment |
| CN109587645A (en) * | 2018-11-12 | 2019-04-05 | 南京邮电大学 | Personnel's recognition methods under indoor environment based on channel state information |
| CN109961100A (en) * | 2019-03-26 | 2019-07-02 | 北京邮电大学 | WIFI indoor occupant intrusion detection method based on Principal Component Analysis |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113810500A (en) * | 2021-09-18 | 2021-12-17 | 中国银行股份有限公司 | Crowd gathering risk identification method and related equipment |
| CN113810500B (en) * | 2021-09-18 | 2023-01-24 | 中国银行股份有限公司 | Crowd gathering risk identification method and related equipment |
| CN114359738A (en) * | 2022-03-16 | 2022-04-15 | 深圳大学 | Cross-scene robust indoor population wireless detection method and system |
| CN114359738B (en) * | 2022-03-16 | 2022-06-14 | 深圳大学 | Cross-scene robust indoor people number wireless detection method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111741471B (en) | Intrusion detection method and device based on CSI and computer storage medium | |
| Arshad et al. | Leveraging transfer learning in multiple human activity recognition using WiFi signal | |
| CN104683984B (en) | The real-time monitoring process method of wireless communication signals and system | |
| Shahzad et al. | Expecting the unexpected: Fast and reliable detection of missing RFID tags in the wild | |
| CN102883360B (en) | A kind of method and system of the passive user's detection of indoor wireless omnidirectional | |
| KR20160095856A (en) | System and method for detecting intrusion intelligently based on automatic detection of new attack type and update of attack type | |
| CN111753686A (en) | CSI-based people number identification method, device, equipment and computer storage medium | |
| EP3198547A1 (en) | Multisensory change detection for internet of things domain | |
| CN109889512B (en) | Charging pile CAN message abnormity detection method and device | |
| CN108197581B (en) | Unmanned aerial vehicle signal identification detection method based on improved AC-WGANs | |
| US10459066B2 (en) | Self-adaptive system and method for robust Wi-Fi indoor localization in large public site | |
| CN109802973A (en) | Method and apparatus for detection flows | |
| Siachalou et al. | RSSI-based maximum likelihood localization of passive RFID tags using a mobile cart | |
| CN110856178B (en) | Behavior identification method based on wireless network physical layer IQ signal | |
| CN112949487A (en) | WiFi-based ship cab personnel number detection method and system | |
| CN107666398B (en) | Group notification method, system and storage medium based on user behavior | |
| CN110830504A (en) | Network intrusion behavior detection method and system | |
| CN111476059A (en) | Target detection method and device, computer equipment and storage medium | |
| CN106446791A (en) | Smart city public monitoring system | |
| WO2021218409A1 (en) | System and method for estimating number of rfid tags, and processor-readable medium | |
| CN108199757A (en) | A kind of method for carrying out real-time intrusion detection to civil small-scale unmanned plane using channel state information | |
| CN107643544B (en) | Method for automatically detecting metal | |
| CN106254818A (en) | Method for monitoring area | |
| CN108494807A (en) | Next-generation key message infrastructure network intruding detection system based on cloud computing | |
| Song et al. | UAS detection and negation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |