CN112881812A - Full-flash real-time positioning method and device based on machine learning coding - Google Patents
Full-flash real-time positioning method and device based on machine learning coding Download PDFInfo
- Publication number
- CN112881812A CN112881812A CN202110031562.2A CN202110031562A CN112881812A CN 112881812 A CN112881812 A CN 112881812A CN 202110031562 A CN202110031562 A CN 202110031562A CN 112881812 A CN112881812 A CN 112881812A
- Authority
- CN
- China
- Prior art keywords
- low
- frequency signal
- pulse
- signal detection
- time
- 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.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R29/00—Arrangements for measuring or indicating electric quantities not covered by groups G01R19/00 - G01R27/00
- G01R29/08—Measuring electromagnetic field characteristics
- G01R29/0807—Measuring electromagnetic field characteristics characterised by the application
- G01R29/0814—Field measurements related to measuring influence on or from apparatus, components or humans, e.g. in ESD, EMI, EMC, EMP testing, measuring radiation leakage; detecting presence of micro- or radiowave emitters; dosimetry; testing shielding; measurements related to lightning
- G01R29/0842—Measurements related to lightning, e.g. measuring electric disturbances, warning systems
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R29/00—Arrangements for measuring or indicating electric quantities not covered by groups G01R19/00 - G01R27/00
- G01R29/02—Measuring characteristics of individual pulses, e.g. deviation from pulse flatness, rise time or duration
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Medical Informatics (AREA)
- Electromagnetism (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Position Fixing By Use Of Radio Waves (AREA)
Abstract
The invention provides a full-flash real-time positioning method and a full-flash real-time positioning device based on machine learning coding, wherein the method comprises the following steps: based on the discharge signals obtained by the distributed low-frequency signal detection substation, carrying out peak searching calculation on the discharge signals with preset bandwidth to obtain pulse waveforms; inputting the pulse waveform into a lightning encoder to obtain a pulse code; obtaining a pulse characteristic coding vector of each substation according to the pulse codes and the corresponding pulse peak time; selecting a substation as a master station, and respectively matching the characteristic code vectors of all other low-frequency signal detection substations with the characteristic code vector of the master station to obtain matched pulse peak time corresponding to the same discharge event; and acquiring the occurrence time and the three-dimensional position of the lightning discharge event according to the matched pulse peak time and the position of the low-frequency signal detection station (comprising the substation and the main station) corresponding to the matched pulse. The embodiment of the invention uses machine learning to encode the lightning pulse signal, thereby improving the matching efficiency, the positioning precision and the positioning efficiency.
Description
Technical Field
The invention relates to the technical field of computers and atmospheric detection, in particular to a full-flash real-time positioning method and device based on machine learning coding.
Background
Lightning is a strong atmospheric discharge phenomenon occurring in nature, and causes a great amount of casualties and property loss worldwide every year. Therefore, the lightning discharge mechanism needs to be researched to guide lightning protection and disaster reduction. Whereas the lightning discharge mechanism study relies in particular on lightning location observations. After decades of development, most regions and countries around the world form own lightning location networks, but basic parameters such as a back-strike position and occurrence time can be given, and in a real situation, cloud lightning accounts for more than 70% of all lightning.
In recent 20 years, low-frequency-based full-flash positioning systems are deployed at home and abroad and are promoting business, but the average lightning positioning capacity of the business full-flash positioning systems is dozens of or dozens of discharge events every time, and a large amount of cloud discharge information is still missed. Although the scientific research field can provide richer positioning information based on the lightning low-frequency multi-station data, the method is based on post-processing of original waveforms, and the algorithm for completing one-time refined lightning is very complex, long in positioning time and low in positioning efficiency, and is difficult to popularize in business application.
Disclosure of Invention
The invention provides a full-flash real-time positioning method and device based on machine learning coding, which are used for solving the defects of very complex algorithm, long positioning time and low positioning efficiency in the prior art and greatly improving the positioning precision and the positioning efficiency.
The invention provides a full-flash real-time positioning method based on machine learning coding, which comprises the following steps: the distributed low-frequency signal detection substation carries out peak searching calculation on the discharge signal with the preset bandwidth to obtain a pulse waveform containing peaks;
inputting the pulse waveform into a machine learning lightning encoder to obtain a pulse code, wherein the lightning encoder is obtained by taking a lightning discharge signal as a sample and taking a decoding result of a coding characteristic as an inspection object for training;
obtaining a characteristic coding vector according to the pulse code and the pulse peak time in the pulse waveform;
selecting a low-frequency signal detection substation as a master station, and respectively matching the characteristic code vectors of all the low-frequency signal detection substations with the characteristic code vectors of the master station to obtain matched pulse peak time corresponding to a discharge signal of the same discharge event;
and acquiring the lightning occurrence time and position according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched data, wherein the low-frequency signal detection station corresponding to the matched data comprises a substation and a main station.
According to the full-flash real-time positioning method based on machine learning coding provided by the invention, the peak searching calculation is carried out on the discharge signal with the preset bandwidth to obtain the pulse waveform, and the method comprises the following steps:
calculating the noise level of the preset bandwidth discharging signal;
according to the noise level, carrying out peak searching operation on the preset bandwidth discharge signal to obtain time corresponding to a peak value;
and intercepting a pulse waveform from the preset bandwidth discharge signal according to the time corresponding to the peak value and a preset pulse width.
According to the full-flash real-time positioning method based on machine learning coding provided by the invention, the method comprises the following steps of selecting one low-frequency signal detection substation as a master station, respectively matching the feature coding vectors of all the low-frequency signal detection substations with the feature coding vector of the master station to obtain the matched pulse peak time corresponding to the discharge signal of the same discharge event, wherein the method comprises the following steps:
calculating the pulse peak value time difference between the pulse peak value time in each low-frequency signal detection substation characteristic coding vector and the pulse peak value time in the master station characteristic coding vector, calculating the distance difference between the corresponding low-frequency signal detection substation and the master station, and screening out the low-frequency signal detection substation characteristic coding vectors of which the peak value time difference does not exceed the distance difference and light velocity ratio as alternative low-frequency signal detection substation characteristic coding vectors;
calculating all feature coding vectors of each alternative low-frequency signal detection substation and feature coding vectors of the main station to perform correlation calculation, and taking the feature coding vector with the highest correlation as a matched low-frequency signal detection substation feature coding vector;
and taking the pulse peak time of the matched feature coding vectors of the low-frequency signal detection substation and the master station as the matched pulse peak time.
According to the full-flash real-time positioning method based on machine learning coding provided by the invention, the lightning occurrence time and position are obtained according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched data, and the method comprises the following steps:
according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched pulse signal, calculating an initial solution of the lightning occurrence time and position by solving a linear equation set through an arrival time method;
and solving an optimal solution by utilizing a Levenberg-Marquardt algorithm to obtain the accurate lightning occurrence time and position.
According to the full-flash real-time positioning method based on the machine learning coding, provided by the invention, the preset bandwidth discharge signal is obtained through the following steps:
and acquiring a lightning discharge signal, filtering and normalizing the lightning discharge signal, and acquiring the preset bandwidth discharge signal.
According to the full-flash real-time positioning method based on the machine learning coding, the detection standard of the lightning encoder in the training process is that the consistency of the decoded result of the output feature coding vector and the original signal is larger than a preset proportion threshold value.
The invention also provides a full-flash real-time positioning device based on machine learning coding, which comprises: not less than 4 low frequency signal detection substations and central station, low frequency signal detection substation includes: low frequency signal detection sensor, receiver and signal acquisition processor, wherein:
the low-frequency signal detection sensor is used for detecting a lightning discharge signal;
the receiver is used for integrating, amplifying and filtering the lightning discharge signal to obtain a preset bandwidth discharge signal;
the signal acquisition processor is used for carrying out peak seeking calculation on the preset bandwidth discharge signal to obtain a pulse waveform, the pulse waveform comprises a pulse of the preset bandwidth discharge signal, the pulse waveform is input into a lightning encoder to obtain a pulse code, the lightning encoder is obtained by training by taking the lightning signal as a sample and taking a decoding result of a coding characteristic as an inspection object, a characteristic coding vector is obtained according to the pulse code and the peak time of the preset bandwidth discharge signal corresponding to the pulse, and the characteristic coding vector is sent to the central station;
and selecting a low-frequency signal detection substation as a master station, wherein the central station is used for receiving the characteristic coding vectors of all the low-frequency signal detection substations, respectively matching the characteristic coding vectors of all the low-frequency signal detection substations with the characteristic coding vectors of the low-frequency signal detection master station to obtain matched pulse peak time and corresponding low-frequency signal detection station position information, and acquiring lightning occurrence time and lightning occurrence position according to the matched pulse peak time and the corresponding low-frequency signal detection station position.
The invention also provides a machine learning coding-based full-flash real-time positioning device, which comprises at least 4 low-frequency signal detection substations and a central station, wherein the low-frequency signal detection substations are used for acquiring the lightning discharge signals, and the central station is used for executing the machine learning coding-based full-flash real-time positioning method provided by the first aspect on the lightning discharge signals.
The invention also provides an electronic device, which comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor executes the program to realize the steps of the machine learning coding-based full-flash real-time positioning method.
The present invention also provides a non-transitory computer readable storage medium having stored thereon a computer program which, when being executed by a processor, implements the steps of the machine learning coding-based all-flash real-time positioning method as described in any of the above.
The invention provides a full-flash real-time positioning method and device based on machine learning coding.
Drawings
In order to more clearly illustrate the technical solutions of the present invention or the prior art, the drawings needed for the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a flow chart of a full-flash real-time positioning method based on machine learning coding according to the present invention;
FIG. 2 is a schematic structural diagram of a full-flash real-time positioning apparatus based on machine learning coding according to the present invention;
fig. 3 is a schematic physical structure diagram of an electronic device provided in the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention clearer, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are some, but not all embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
At present, the artificial intelligence technology is developing vigorously, and shows great potential in the primary application in the fields of weather early warning and forecasting and the like, and the hit rate of the early warning and forecasting is greatly improved. The embodiment of the invention applies an artificial intelligence technology to low-frequency total-flash positioning, provides a real-time fine total-flash positioning technology for waveform coding based on machine learning, and develops a corresponding system.
The embodiment of the invention provides a full-flash real-time positioning method based on machine learning coding, as shown in fig. 1, the method comprises the following steps:
110, performing peak searching calculation on the discharging signal with the preset bandwidth to obtain a pulse waveform containing a peak by using the discharging signal obtained by the distributed low-frequency signal detection substation;
firstly, the method relates to a distributed low-frequency signal detection substation which is mainly used for detecting lightning signals, in the embodiment of the invention, the number of the low-frequency signal detection substations is at least 4, and one of the 4 low-frequency signal detection substations is selected as a low-frequency signal detection main station.
The low-frequency signal detection substation detects a lightning discharge signal, the lightning discharge signal is the most initial detection signal, band-pass filtering and normalization processing are carried out on the lightning discharge signal, a preset bandwidth discharge signal meeting the positioning requirement is obtained, generally from KHZ to hundreds of KHZ, and different technicians can properly expand or reduce corresponding ranges.
The preset bandwidth discharging signal is a normalized signal after filtering processing. And carrying out peak searching calculation on the preset bandwidth discharge signal to obtain a peak in the preset bandwidth discharge signal, intercepting a discharge signal waveform with a certain width containing the obtained peak as a pulse waveform, wherein the width can be specifically selected according to actual needs.
120, inputting the pulse waveform into a machine learning lightning encoder to obtain a pulse code, wherein the lightning encoder is obtained by training by taking a lightning discharge signal as a sample and a decoding result of a coding characteristic as an inspection object;
and inputting the pulse waveform into a lightning encoder to obtain a pulse code, wherein the lightning encoder is obtained by taking a lightning discharge signal as a sample and taking a decoding result of the coding characteristic as an inspection object for training.
130, obtaining a feature coding vector according to the pulse code and the pulse peak time in the pulse waveform;
and obtaining a feature coding vector according to the pulse coding and the peak time in the pulse waveform, wherein the feature coding vector is composed of two parts of information of the pulse coding and the peak time.
The lightning discharge signals detected by the low-frequency signal detection substations are processed in the whole process, the lightning discharge signals detected by each low-frequency signal detection substation are processed in the same way, and each low-frequency signal detection substation can obtain a feature encoding vector.
140, selecting a low-frequency signal detection substation as a master station, and respectively matching the feature code vectors of all the low-frequency signal detection substations with the feature code vectors of the master station to obtain matched pulse peak time corresponding to a discharge signal of the same discharge event;
and respectively matching the characteristic coding vector of each low-frequency signal detection substation with the characteristic coding vector of the low-frequency signal detection master station to obtain the characteristic coding vector of the low-frequency signal detection substation with the highest matching degree, wherein the characteristic coding vector of the low-frequency signal detection substation with the highest matching degree and the master station correspond to the same discharge event discharge signal to obtain the matched pulse peak time of the discharge event discharge signal.
And 150, acquiring the lightning occurrence time and position according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched data, wherein the low-frequency signal detection station corresponding to the matched data comprises a substation and a main station.
And then, acquiring lightning occurrence time and position according to the matched position of the low-frequency signal detection substation, the position of the low-frequency signal detection main station and the matched pulse peak time.
The full-flash real-time positioning method based on machine learning coding innovatively uses machine learning to code lightning pulses, greatly improves positioning precision and positioning efficiency, and is an important scheme for real-time fine positioning.
On the basis of the foregoing embodiment, preferably, the performing peak searching calculation on the discharge signal with the preset bandwidth to obtain a pulse waveform includes:
calculating the noise level of the preset bandwidth discharging signal;
according to the noise level, carrying out peak searching operation on the preset bandwidth discharge signal to obtain a peak value;
and intercepting a pulse waveform from the preset bandwidth discharge signal according to the time corresponding to the peak value and a preset pulse width.
Calculating to obtain the noise level of the preset bandwidth discharge signal, performing peak searching operation on the preset bandwidth discharge signal, wherein the peak searching condition can be set, the minimum interval of the peak and the peak can be set according to the actual bandwidth of the signal, the minimum amplitude of the peak searching is greater than the bottom noise and is determined according to the actual positioning fineness requirement, and the like.
On the basis of the foregoing embodiment, preferably, the selecting one low-frequency signal detection substation as a master station, and matching feature code vectors of all the low-frequency signal detection substations with feature code vectors of the master station respectively to obtain matched pulse peak times corresponding to a discharge signal of a same discharge event includes:
calculating the pulse peak value time difference between the pulse peak value time in each low-frequency signal detection substation characteristic coding vector and the pulse peak value time in the master station characteristic coding vector, calculating the distance difference between the corresponding low-frequency signal detection substation and the master station, and screening out the low-frequency signal detection substation characteristic coding vectors of which the peak value time difference does not exceed the distance difference and light velocity ratio as alternative low-frequency signal detection substation characteristic coding vectors;
calculating all feature coding vectors of each alternative low-frequency signal detection substation and feature coding vectors of the main station to perform correlation calculation, and taking the feature coding vector with the highest correlation as a matched low-frequency signal detection substation feature coding vector;
and taking the pulse peak time of the matched feature coding vectors of the low-frequency signal detection substation and the master station as the matched pulse peak time.
And for the characteristic coding vectors of not less than four low-frequency signal detection substations, determining the master station, and matching the characteristic coding vectors of other substations with the characteristic coding vector of the master station. Matching is carried out in two steps, wherein in the first step, the peak value time difference does not exceed the ratio of the distance difference between two stations to the light speed, in the second step, under the condition that the above conditions are met, the feature code vector is subjected to correlation calculation, and the pulse with the highest correlation degree is regarded as a matching pulse. After this, a set of matched pulse peak time data is obtained.
On the basis of the foregoing embodiment, preferably, the acquiring lightning occurrence time and position according to the matched position of the low-frequency signal detection substation, the position of the low-frequency signal detection master station, and the matched pulse peak time includes:
according to the matched pulse peak time, the matched position information of the low-frequency signal detection substation and the low-frequency signal detection main station, the linear equation set is solved by using the arrival time method to calculate the three-dimensional position of the lightning and the initial solution of the time, and then the optimal solution is solved by using a Levenberg-Marquardt (Levenberg-Marquardt, LM for short) algorithm to obtain the accurate time and position (location coordinates) of the lightning.
Because the machine learning coding waveform characteristics are adopted for real-time positioning, the positioning efficiency and the refinement degree are greatly improved.
On the basis of the above embodiment, preferably, the preset bandwidth discharging signal is obtained by:
and acquiring a lightning discharge signal, filtering and normalizing the lightning discharge signal, and acquiring the preset bandwidth discharge signal.
On the basis of the above embodiment, preferably, the test criterion of the lightning encoder in the training process is that the consistency between the decoded result of the output feature coding vector and the original signal is greater than a preset proportion threshold.
Based on machine learning techniques, a lightning encoder is formed by training, wherein the input is an original lightning signal waveform and the output is a machine-learned pulse code vector. Here, the machine learning specific method may be deep learning.
In training the model, the criterion for checking whether the output feature vector is usable is that the decoded result using the vector is as consistent as possible with the original signal.
The embodiment of the invention provides a full-flash real-time positioning device based on machine learning coding, as shown in fig. 2, the device comprises: not less than 4 low frequency signal detection substations and central station, low frequency signal detection substation includes: signal detection sensor, receiver and signal acquisition processor, wherein:
the low-frequency signal detection sensor is used for detecting a lightning discharge signal;
the receiver is used for integrating, amplifying and filtering the lightning discharge signal to obtain a preset bandwidth discharge signal;
the signal acquisition processor is used for carrying out peak seeking calculation on the preset bandwidth discharge signal to obtain a pulse waveform, the pulse waveform comprises a pulse of the preset bandwidth discharge signal, the pulse waveform is input into a lightning encoder to obtain a pulse code, the lightning encoder is obtained by training by taking the lightning signal as a sample and taking a decoding result of a coding characteristic as an inspection object, a characteristic coding vector is obtained according to the pulse code and the peak time of the preset bandwidth discharge signal corresponding to the pulse, and the characteristic coding vector is sent to the central station;
and selecting a low-frequency signal detection substation as a master station, wherein the central station is used for receiving the feature coding vectors of all the low-frequency signal detection substations, matching the feature coding vectors of all the low-frequency signal detection substations with the feature coding vectors of the low-frequency signal detection master station respectively to obtain matched pulse peak time and corresponding low-frequency signal detection station position information, and acquiring lightning occurrence time and lightning occurrence position according to the matched pulse peak time and the corresponding low-frequency signal detection station position.
The device mainly comprises low-frequency signal detection substations and a central station, wherein the low-frequency signal detection substations are not less than 4 stations, each substation comprises a low-frequency signal detection sensor (antenna), a receiver, a signal acquisition processor and a power supply system, and the power supply system can be a solar cell panel.
The detection antenna may be a magnetic field signal detection antenna, an electric field change signal detection antenna (such as a fast antenna, an electric field change measuring instrument), or an electromagnetic field signal detection antenna.
Signals received by the antennas are subjected to integration, amplification, filtering and other processing through a receiver, and lightning discharge signals of a required frequency band are obtained.
The receiver outputs signals to a signal acquisition processor, and the signal acquisition processor adopts a computer or an FPGA and a real-time processing system of edge calculation.
The algorithm of the above step 110-130 is implemented in real time, and the pulse coding data is transmitted to the central station through the 4G or wired network in real time, and the steps 140 and 150 are continued to implement real-time positioning.
Another embodiment of the present invention provides a full-flash real-time positioning device based on machine learning coding, including: the lightning signal detection substation is used for acquiring the lightning discharge signals, and the central station is used for executing the full-flash real-time positioning method based on machine learning coding on the lightning discharge signals.
In the embodiment of the invention, the substation can only carry out the collection task and transmit the original waveform to the central station, and the central station carries out the positioning calculation. The sensors, receivers and signal acquisition processor are powered by a power supply system (optionally a solar panel) using steps 110 to 150.
As shown in fig. 3, an electronic device provided in an embodiment of the present invention may include: a processor (processor)310, a communication Interface (communication Interface)320, a memory (memory)330 and a communication bus 340, wherein the processor 310, the communication Interface 320 and the memory 330 communicate with each other via the communication bus 340. The processor 310 may invoke logic instructions in the memory 330 to perform a full flash real-time location method based on machine learning coding, the method comprising:
the distributed low-frequency signal detection substation carries out peak searching calculation on the discharge signal with the preset bandwidth to obtain a pulse waveform containing peaks;
inputting the pulse waveform into a machine learning lightning encoder to obtain a pulse code, wherein the lightning encoder is obtained by taking a lightning discharge signal as a sample and taking a decoding result of a coding characteristic as an inspection object for training;
obtaining a characteristic coding vector according to the pulse code and the pulse peak time in the pulse waveform;
selecting a low-frequency signal detection substation as a master station, and respectively matching the characteristic code vectors of all the low-frequency signal detection substations with the characteristic code vectors of the master station to obtain matched pulse peak time corresponding to a discharge signal of the same discharge event;
and acquiring the lightning occurrence time and position according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched data, wherein the low-frequency signal detection station corresponding to the matched data comprises a substation and a main station.
In addition, the logic instructions in the memory 330 may be implemented in the form of software functional units and stored in a computer readable storage medium when the software functional units are sold or used as independent products. 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 and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk.
In another aspect, the present invention also provides a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, when the program instructions are executed by a computer, the computer being capable of executing a full-flash real-time positioning method based on machine learning coding provided by the above methods, the method including:
the distributed low-frequency signal detection substation carries out peak searching calculation on the discharge signal with the preset bandwidth to obtain a pulse waveform containing peaks;
inputting the pulse waveform into a machine learning lightning encoder to obtain a pulse code, wherein the lightning encoder is obtained by taking a lightning discharge signal as a sample and taking a decoding result of a coding characteristic as an inspection object for training;
obtaining a characteristic coding vector according to the pulse code and the pulse peak time in the pulse waveform;
selecting a low-frequency signal detection substation as a master station, and respectively matching the characteristic code vectors of all the low-frequency signal detection substations with the characteristic code vectors of the master station to obtain matched pulse peak time corresponding to a discharge signal of the same discharge event;
and acquiring the lightning occurrence time and position according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched data, wherein the low-frequency signal detection station corresponding to the matched data comprises a substation and a main station.
In yet another aspect, the present invention further provides a non-transitory computer readable storage medium, on which a computer program is stored, the computer program being implemented by a processor to perform the method for full flash real-time positioning based on machine learning coding provided in the foregoing, the method comprising:
the distributed low-frequency signal detection substation carries out peak searching calculation on the discharge signal with the preset bandwidth to obtain a pulse waveform containing peaks;
inputting the pulse waveform into a machine learning lightning encoder to obtain a pulse code, wherein the lightning encoder is obtained by taking a lightning discharge signal as a sample and taking a decoding result of a coding characteristic as an inspection object for training;
obtaining a characteristic coding vector according to the pulse code and the pulse peak time in the pulse waveform;
selecting a low-frequency signal detection substation as a master station, and respectively matching the characteristic code vectors of all the low-frequency signal detection substations with the characteristic code vectors of the master station to obtain matched pulse peak time corresponding to a discharge signal of the same discharge event;
and acquiring the lightning occurrence time and position according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched data, wherein the low-frequency signal detection station corresponding to the matched data comprises a substation and a main station.
The above-described embodiments of the apparatus are merely illustrative, and the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
Finally, it should be noted that: the above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A full-flash real-time positioning method based on machine learning coding is characterized by comprising the following steps:
the distributed low-frequency signal detection substation carries out peak searching calculation on the discharge signal with the preset bandwidth to obtain a pulse waveform containing peaks;
inputting the pulse waveform into a machine learning lightning encoder to obtain a pulse code, wherein the lightning encoder is obtained by taking a lightning discharge signal as a sample and taking a decoding result of a coding characteristic as an inspection object for training;
obtaining a characteristic coding vector according to the pulse code and the pulse peak time in the pulse waveform;
selecting a low-frequency signal detection substation as a master station, and respectively matching the characteristic code vectors of all the low-frequency signal detection substations with the characteristic code vectors of the master station to obtain matched pulse peak time corresponding to a discharge signal of the same discharge event;
and acquiring the lightning occurrence time and position according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched data, wherein the low-frequency signal detection station corresponding to the matched data comprises a substation and a main station.
2. The machine learning coding-based full-flash real-time positioning method according to claim 1, wherein the performing peak-finding calculation on the preset bandwidth discharge signal to obtain the pulse waveform comprises:
calculating the noise level of the preset bandwidth discharging signal;
according to the noise level, carrying out peak searching operation on the preset bandwidth discharge signal to obtain time corresponding to a peak value;
and intercepting a pulse waveform from the preset bandwidth discharge signal according to the time corresponding to the peak value and a preset pulse width.
3. The machine learning coding-based all-flash real-time positioning method according to claim 1, wherein the selecting one low-frequency signal detection substation as a master station matches the feature code vectors of all the low-frequency signal detection substations with the feature code vector of the master station, respectively, to obtain matched pulse peak time corresponding to a discharge signal of the same discharge event, comprises:
calculating the pulse peak value time difference between the pulse peak value time in each low-frequency signal detection substation characteristic coding vector and the pulse peak value time in the master station characteristic coding vector, calculating the distance difference between the corresponding low-frequency signal detection substation and the master station, and screening out the low-frequency signal detection substation characteristic coding vectors of which the peak value time difference does not exceed the distance difference and light velocity ratio as alternative low-frequency signal detection substation characteristic coding vectors;
calculating all feature coding vectors of each alternative low-frequency signal detection substation and feature coding vectors of the main station to perform correlation calculation, and taking the feature coding vector with the highest correlation as a matched low-frequency signal detection substation feature coding vector;
and taking the pulse peak time of the matched feature coding vectors of the low-frequency signal detection substation and the master station as the matched pulse peak time.
4. The full-flash real-time positioning method based on machine learning coding according to claim 1, wherein the acquiring lightning occurrence time and position according to the matched pulse peak time and the position of the low frequency signal detection station corresponding to the matched data comprises:
according to the matched pulse peak time and the position of the low-frequency signal detection station corresponding to the matched pulse signal, calculating an initial solution of the lightning occurrence time and position by solving a linear equation set through an arrival time method;
and solving an optimal solution by utilizing a Levenberg-Marquardt algorithm to obtain the accurate lightning occurrence time and position.
5. The full-flash real-time positioning method based on machine learning coding according to any one of claims 1 to 4, wherein the preset bandwidth discharging signal is obtained by the following steps:
and acquiring a lightning discharge signal, filtering and normalizing the lightning discharge signal, and acquiring the preset bandwidth discharge signal.
6. The full-flash real-time location method based on machine learning coding according to any of claims 1 to 4, wherein the testing criteria of the lightning encoder in the training process is that the consistency of the decoded result of the output feature coding vector and the original signal is larger than a preset ratio threshold.
7. A full-flash real-time positioning device based on machine learning coding, which applies the full-flash real-time positioning method based on machine learning coding of any one of claims 1 to 6, characterized by not less than 4 low-frequency signal detection substations and a central station, wherein the low-frequency signal detection substations include: low frequency signal detection sensor, receiver and signal acquisition processor, wherein:
the low-frequency signal detection sensor is used for detecting a lightning discharge signal;
the receiver is used for integrating, amplifying and filtering the lightning discharge signal to obtain a preset bandwidth discharge signal;
the signal acquisition processor is used for carrying out peak seeking calculation on the preset bandwidth discharge signal to obtain a pulse waveform, the pulse waveform comprises a pulse of the preset bandwidth discharge signal, the pulse waveform is input into a lightning encoder to obtain a pulse code, the lightning encoder is obtained by training by taking the lightning signal as a sample and taking a decoding result of a coding characteristic as an inspection object, a characteristic coding vector is obtained according to the pulse code and the peak time of the preset bandwidth discharge signal corresponding to the pulse, and the characteristic coding vector is sent to the central station;
and selecting a low-frequency signal detection substation as a master station, wherein the central station is used for receiving the feature coding vectors of all the low-frequency signal detection substations, matching the feature coding vectors of all the low-frequency signal detection substations with the feature coding vectors of the low-frequency signal detection master station respectively to obtain matched pulse peak time and corresponding low-frequency signal detection station position information, and acquiring lightning occurrence time and lightning occurrence position according to the matched pulse peak time and the corresponding low-frequency signal detection station position.
8. A full-flash real-time positioning device based on machine learning coding, comprising: the lightning signal detection substation is used for acquiring the lightning discharge signal, processing the lightning discharge signal in real time to acquire a feature code, obtaining a feature code vector and sending the feature code vector to the central station, and the central station is used for executing the full-flash real-time positioning method based on the machine learning code according to any one of claims 1 to 6 on the lightning discharge signal.
9. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps of the machine learning coding based all-flash real-time localization method according to any of claims 1 to 6.
10. A non-transitory computer readable storage medium, having a computer program stored thereon, wherein the computer program, when being executed by a processor, implements the steps of the full flash real-time localization method based on machine learning coding according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110031562.2A CN112881812B (en) | 2021-01-11 | 2021-01-11 | Full-flash real-time positioning method and device based on machine learning coding |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110031562.2A CN112881812B (en) | 2021-01-11 | 2021-01-11 | Full-flash real-time positioning method and device based on machine learning coding |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112881812A true CN112881812A (en) | 2021-06-01 |
| CN112881812B CN112881812B (en) | 2022-02-22 |
Family
ID=76047802
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110031562.2A Active CN112881812B (en) | 2021-01-11 | 2021-01-11 | Full-flash real-time positioning method and device based on machine learning coding |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112881812B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113820657A (en) * | 2021-10-15 | 2021-12-21 | 南通大学 | Lightning three-dimensional positioning method |
| CN116577568A (en) * | 2023-07-13 | 2023-08-11 | 中国气象科学研究院 | Lightning positioning method and device based on hash value characteristics |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101609145A (en) * | 2009-07-29 | 2009-12-23 | 中国气象科学研究院 | The all-lightning positioning system that very low frequency (VLF) combines with very high frequency(VHF) |
| CN102298097A (en) * | 2011-07-15 | 2011-12-28 | 华中科技大学 | Method for estimating thunder impulse signal Time Difference of Arrival (TDOA) |
| CN102540145A (en) * | 2012-01-12 | 2012-07-04 | 中国人民解放军理工大学 | Single-station quasi-three-dimensional lightning positioning device and method |
| US20160119761A1 (en) * | 2014-08-12 | 2016-04-28 | Abl Ip Holding Llc | System and method for estimating the position and orientation of a mobile communications device in a beacon-based positioning system |
| CN105974207A (en) * | 2016-05-05 | 2016-09-28 | 中国科学院电工研究所 | Three dimensional full-lightning detection positioning system based on very-low frequency/low frequency |
| CN106571831A (en) * | 2016-10-28 | 2017-04-19 | 华南理工大学 | LDPC hard decision decoding method based on depth learning and decoder |
| CN107037272A (en) * | 2017-06-20 | 2017-08-11 | 云南电网有限责任公司昆明供电局 | Lightning Location Method based on thunder and lightning multivariate data peak-seeking Shicha algorithm |
| CN109188359A (en) * | 2018-09-13 | 2019-01-11 | 中国气象科学研究院 | lightning positioning method and device |
| CN110929811A (en) * | 2019-10-21 | 2020-03-27 | 北京理工大学 | Deep learning method for acquiring high-resolution data of full-waveform laser radar |
| CN111537853A (en) * | 2020-06-29 | 2020-08-14 | 国网山东省电力公司菏泽供电公司 | Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis |
-
2021
- 2021-01-11 CN CN202110031562.2A patent/CN112881812B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101609145A (en) * | 2009-07-29 | 2009-12-23 | 中国气象科学研究院 | The all-lightning positioning system that very low frequency (VLF) combines with very high frequency(VHF) |
| CN102298097A (en) * | 2011-07-15 | 2011-12-28 | 华中科技大学 | Method for estimating thunder impulse signal Time Difference of Arrival (TDOA) |
| CN102540145A (en) * | 2012-01-12 | 2012-07-04 | 中国人民解放军理工大学 | Single-station quasi-three-dimensional lightning positioning device and method |
| US20160119761A1 (en) * | 2014-08-12 | 2016-04-28 | Abl Ip Holding Llc | System and method for estimating the position and orientation of a mobile communications device in a beacon-based positioning system |
| CN105974207A (en) * | 2016-05-05 | 2016-09-28 | 中国科学院电工研究所 | Three dimensional full-lightning detection positioning system based on very-low frequency/low frequency |
| CN106571831A (en) * | 2016-10-28 | 2017-04-19 | 华南理工大学 | LDPC hard decision decoding method based on depth learning and decoder |
| CN107037272A (en) * | 2017-06-20 | 2017-08-11 | 云南电网有限责任公司昆明供电局 | Lightning Location Method based on thunder and lightning multivariate data peak-seeking Shicha algorithm |
| CN109188359A (en) * | 2018-09-13 | 2019-01-11 | 中国气象科学研究院 | lightning positioning method and device |
| CN110929811A (en) * | 2019-10-21 | 2020-03-27 | 北京理工大学 | Deep learning method for acquiring high-resolution data of full-waveform laser radar |
| CN111537853A (en) * | 2020-06-29 | 2020-08-14 | 国网山东省电力公司菏泽供电公司 | Intelligent detection method for partial discharge of switch cabinet based on multi-source heterogeneous data analysis |
Non-Patent Citations (2)
| Title |
|---|
| JIAQUAN WANG ET AL.: "Classification of VLF/LF Lightning Signals Using Sensors and Deep Learning Methods", 《SENSORS》 * |
| YANAN ZHU ET AL.: "A machine-learning approach to classify cloud-to-ground and intracloud lightning", 《GEOPHYSICAL RESEARCH LETTERS》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113820657A (en) * | 2021-10-15 | 2021-12-21 | 南通大学 | Lightning three-dimensional positioning method |
| CN116577568A (en) * | 2023-07-13 | 2023-08-11 | 中国气象科学研究院 | Lightning positioning method and device based on hash value characteristics |
| CN116577568B (en) * | 2023-07-13 | 2023-10-27 | 中国气象科学研究院 | Lightning positioning method and device based on hash value characteristics |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112881812B (en) | 2022-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109581546B (en) | Rainfall type identification method based on microwave link attenuation and polarization information | |
| CN112881812B (en) | Full-flash real-time positioning method and device based on machine learning coding | |
| CN104007374B (en) | Method for resisting interference signals in converter transformer partial discharge detection experiment | |
| CN102565855B (en) | Ground micro-seismic data processing method of oil field fracturing | |
| CN101881826A (en) | Scanning-mode sea clutter local multi-fractal target detector | |
| CN104331583B (en) | A kind of Multifractal Modeling method based on Observed sea clutter | |
| CN102879766A (en) | Method and apparatus for detecting and tracking faint target of high frequency ground wave radar | |
| CN106771589B (en) | A kind of implementation method of the real-time Waterfall plot of spectrum analysis | |
| CN104869630A (en) | Pseudo base station rapid positioning method and system based on offline fingerprint database | |
| CN110658494B (en) | Passive unmanned aerial vehicle positioning method and server | |
| CN112307969A (en) | Pulse signal classification identification method and device and computer equipment | |
| CN110427878A (en) | A kind of sudden and violent signal recognition method of Rapid Radio and system | |
| CN109342828A (en) | A kind of lightening pulse signal detecting method based on frequency domain constant false alarm | |
| Siyang et al. | WKNN indoor Wi-Fi localization method using k-means clustering based radio mapping | |
| CN115032693B (en) | Method and device for automatically identifying pre-earthquake of strong earthquake | |
| CN114152980A (en) | Method and device for rapidly and automatically producing seismic source mechanism solution | |
| CN109061632A (en) | A kind of unmanned plane recognition methods | |
| CN106618499B (en) | Fall detection device, fall detection method and device | |
| CN106052837A (en) | Train vibration noise identifying method used in high-speed railway earthquake warning | |
| CN113608193A (en) | Radar multi-target distance and speed estimation method based on UNet | |
| CN103592618A (en) | Ultrasonic positioning method and ultrasonic positioning system | |
| CN112964938B (en) | Lightning single-station positioning method, device and system based on artificial intelligence | |
| CN103994816A (en) | Identification method based on optical fiber multiple events | |
| CN109945075B (en) | Method and device for detecting leakage degree of water supply pipeline | |
| CN103308918A (en) | Fish identification method and system based on segmented time-domain centroid features |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |