CN114415246B - Underground metal object positioning method and system based on machine learning - Google Patents
Underground metal object positioning method and system based on machine learning Download PDFInfo
- Publication number
- CN114415246B CN114415246B CN202210314400.4A CN202210314400A CN114415246B CN 114415246 B CN114415246 B CN 114415246B CN 202210314400 A CN202210314400 A CN 202210314400A CN 114415246 B CN114415246 B CN 114415246B
- Authority
- CN
- China
- Prior art keywords
- metal object
- underground metal
- model
- data set
- underground
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01V—GEOPHYSICS; GRAVITATIONAL MEASUREMENTS; DETECTING MASSES OR OBJECTS; TAGS
- G01V3/00—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation
- G01V3/08—Electric or magnetic prospecting or detecting; Measuring magnetic field characteristics of the earth, e.g. declination, deviation operating with magnetic or electric fields produced or modified by objects or geological structures or by detecting devices
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01C—MEASURING DISTANCES, LEVELS OR BEARINGS; SURVEYING; NAVIGATION; GYROSCOPIC INSTRUMENTS; PHOTOGRAMMETRY OR VIDEOGRAMMETRY
- G01C21/00—Navigation; Navigational instruments not provided for in groups G01C1/00 - G01C19/00
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/086—Learning methods using evolutionary algorithms, e.g. genetic algorithms or genetic programming
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Remote Sensing (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Radar, Positioning & Navigation (AREA)
- Electromagnetism (AREA)
- Bioinformatics & Computational Biology (AREA)
- Medical Informatics (AREA)
- Environmental & Geological Engineering (AREA)
- Geology (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geophysics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Automation & Control Theory (AREA)
- Evolutionary Biology (AREA)
- Physiology (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Geophysics And Detection Of Objects (AREA)
Abstract
The invention relates to a method and a system for positioning an underground metal object based on machine learning, which belong to the field of positioning and comprise the following steps: acquiring a data set; the ith sample data in the data set comprises a plurality of electromagnetic induction signals acquired from different places of the ith area and the underground metal object position coordinates of the ith area; constructing different multiple underground metal object positioning sub-models; obtaining a plurality of trained underground metal object positioning sub-models; weighting and adding the trained underground metal object positioning sub-models to obtain an integrated underground metal object positioning model, wherein the sum of the weighted weights is 1; adjusting each weight in the integrated underground metal object positioning model by adopting a particle swarm algorithm to obtain an underground metal object real-time positioning model; and obtaining the position coordinates of the underground metal object in the range of the area to be measured by adopting the real-time positioning model of the underground metal object. The invention improves the positioning efficiency and precision.
Description
Technical Field
The invention relates to the technical field of positioning, in particular to a method and a system for positioning an underground metal object based on machine learning.
Background
With the development and progress of science and technology, the demand for positioning underground metal objects is increasing in many fields such as archaeology, prospecting, geological mapping, unexplosive bomb detection and the like. The underground metal object positioning method based on electromagnetic induction is a common underground metal object positioning method at present due to the advantages of sensitivity to metal materials, strong penetrability of emitted low-frequency electromagnetic waves and the like. The underground metal target positioning method based on electromagnetic induction utilizes electromagnetic positioning equipment to emit electromagnetic waves near an underground metal object to form a magnetic field to act on the underground metal object, the underground metal object is induced to generate eddy current, the eddy current generates an induction signal, and finally the positioning equipment receives the induction signal. However, when the sensing signal is analyzed conventionally, the position of the subsurface metal target is estimated by minimizing a mean square error function between fitted values calculated from a physical model and detected values. However, in practical application scenarios, environmental noise often exists, the method has low positioning accuracy even abnormal positioning occurs when the environmental noise exists, and the method has poor robustness, so how to analyze the received sensing signal and position the underground object with higher accuracy becomes a problem that needs to be solved urgently.
Disclosure of Invention
The invention aims to provide a method for positioning an underground metal object based on machine learning, which improves the positioning efficiency and precision.
In order to achieve the purpose, the invention provides the following scheme:
a method of machine learning based location of a subsurface metal object, comprising:
acquiring a data set; the ith sample data in the data set comprises a plurality of electromagnetic induction signals acquired from different positions of the ith area and underground metal object position coordinates of the ith area;
preprocessing electromagnetic induction signals in each sample data in the data set to obtain a preprocessed data set;
constructing a plurality of different underground metal object positioning sub-models, wherein the underground metal object positioning sub-models are machine learning models;
respectively training underground metal object positioning sub-models in various places by using a preprocessed data set and taking a plurality of preprocessed electromagnetic induction signals as input and corresponding underground metal object coordinates as output to obtain a plurality of corresponding trained underground metal object positioning sub-models;
weighting and adding the trained underground metal object positioning sub-models to obtain an integrated underground metal object positioning model, wherein the sum of the weighted weights is 1;
aiming at the minimum positioning error, adjusting each weight in the integrated underground metal object positioning model by adopting a particle swarm algorithm to obtain an underground metal object real-time positioning model;
preprocessing the collected electromagnetic induction signals of a plurality of nodes in the range of the area to be detected, inputting the preprocessed electromagnetic induction signals into the underground metal object real-time positioning model, and obtaining the position coordinates of the underground metal object in the range of the area to be detected.
Optionally, each of the subsurface metal object location sub-models comprises a support vector regression model and a neural network regression model.
Optionally, the preprocessing the electromagnetic induction signal in each sample data in the data set to obtain a preprocessed data set specifically includes:
and normalizing the electromagnetic induction signals in the sample data in the data set by adopting a linear function normalization method or a zero-mean normalization method to obtain a preprocessed data set.
Optionally, the pre-processed data set is used, a plurality of pre-processed electromagnetic induction signals are used as input, corresponding underground metal object coordinates are used as output to train underground metal object positioning submodels in various regions respectively, and a plurality of trained underground metal object positioning submodels are obtained, which specifically includes:
dividing the preprocessed data set into a training set and a verification set according to the proportion of 7: 3;
and for each underground metal object positioning sub-model, adjusting parameters in the underground metal object positioning sub-model by using a cross validation method based on the training set and the validation set to obtain the trained underground metal object positioning sub-model.
Optionally, the acquiring the data set specifically includes:
presetting n nodes at regular intervals in an area where underground metal objects are known to be buried;
electromagnetic signals are transmitted at the n nodes by the handheld mobile probe device and electromagnetic induction signals are received at the n nodes.
The invention also discloses a positioning system of the underground metal object based on machine learning, which comprises the following components:
the data set acquisition module is used for acquiring a data set; the ith sample data in the data set comprises a plurality of electromagnetic induction signals acquired from different places of the ith area and the underground metal object position coordinates of the ith area;
the data preprocessing module is used for preprocessing electromagnetic induction signals in the sample data in the data set to obtain a preprocessed data set;
the underground metal object positioning sub-model building module is used for building a plurality of different underground metal object positioning sub-models, and the underground metal object positioning sub-models are machine learning models;
the underground metal object positioning sub-model training module is used for training underground metal object positioning sub-models in various places by adopting a preprocessed data set and taking a plurality of preprocessed electromagnetic induction signals as input and corresponding underground metal object coordinates as output to obtain a plurality of trained underground metal object positioning sub-models;
the underground metal object positioning model integration module is used for weighting and adding the trained underground metal object positioning sub-models to obtain an integrated underground metal object positioning model, and the sum of the weighted weights is 1;
the underground metal object real-time positioning model determining module is used for adjusting each weight in the integrated underground metal object positioning model by adopting a particle swarm algorithm to obtain an underground metal object real-time positioning model by taking the minimum positioning error as a target;
and the underground metal object real-time positioning model application module is used for preprocessing the acquired electromagnetic induction signals of a plurality of nodes in the range of the area to be detected and then inputting the preprocessed signals into the underground metal object real-time positioning model to obtain the position coordinates of the underground metal object in the range of the area to be detected.
Optionally, each of the subsurface metal object location sub-models comprises a support vector regression model and a neural network regression model.
Optionally, the data preprocessing module specifically includes:
and the data preprocessing unit is used for carrying out normalization processing on the electromagnetic induction signals in the sample data in the data set by adopting a linear function normalization method or a zero-mean normalization method to obtain a preprocessed data set.
Optionally, the underground metal object positioning sub-model training module specifically includes:
the data set dividing unit is used for dividing the preprocessed data set into a training set and a verification set according to the proportion of 7: 3;
and the underground metal object positioning sub-model training unit is used for adjusting parameters in the underground metal object positioning sub-model by using a cross validation method based on the training set and the validation set for each underground metal object positioning sub-model to obtain a trained underground metal object positioning sub-model.
Optionally, the data set obtaining module specifically includes:
the node presetting unit is used for presetting n nodes at a certain distance in an area where the underground metal object is known to be buried;
and the electromagnetic induction signal acquisition unit is used for transmitting electromagnetic signals at the n nodes through the handheld mobile detection equipment and receiving the electromagnetic induction signals at the n nodes.
According to the specific embodiment provided by the invention, the invention discloses the following technical effects:
the invention discloses a positioning method and a system of an underground metal object based on machine learning, wherein the acquired induction signal intensity data is input into an underground metal object positioning model during actual detection and positioning, so that the underground metal object can be positioned, and the positioning efficiency and precision are improved; and training by adopting a plurality of machine learning models to obtain the underground metal object positioning sub-model, and finally integrating the underground metal object positioning sub-model into an underground metal object positioning model, so that compared with the method using a single machine learning model, the generalization performance of the underground metal object positioning model is enhanced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings without inventive exercise.
FIG. 1 is a schematic flow chart of a method for locating underground metal objects based on machine learning according to the present invention;
FIG. 2 is a flow chart of the present invention for adjusting the weight of each underground metal object positioning sub-model by using a particle swarm optimization;
fig. 3 is a schematic structural diagram of a positioning system of a subsurface metal object based on machine learning according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. 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.
The invention aims to provide a method for positioning an underground metal object based on machine learning, which improves the positioning efficiency and precision.
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in further detail below.
Fig. 1 is a schematic flow chart of a method for locating an underground metal object based on machine learning according to the present invention, and as shown in fig. 1, the method for locating an underground metal object based on machine learning includes:
step 101: acquiring a data set; the ith sample data in the data set comprises a plurality of electromagnetic induction signals acquired from different positions of the ith area and underground metal object position coordinates of the ith area.
And acquiring sample data in the data set by an offline acquisition submodule, wherein the offline acquisition submodule comprises a transmitting module, a receiving module and a data preprocessing module.
The offline acquisition submodule presets n nodes at certain intervals in an area where underground metal objects at different known positions are buried, and the manual handheld mobile detection equipment is arranged in the n nodes through the transmitting moduleThe nodes transmit electromagnetic signals and the receiving module receives induction signal strength data (electromagnetic induction signals) [ s ] at the n nodes1,s2,…,sn]Finally, the induction signal intensity received by the n nodes is preprocessed by the data preprocessing module [ s ]1’,s2’,…,sn’]Coordinates [ x, y, z ] with underground metal objects]Sample data [ s ] is constructed1’,s2’,…,sn’,x,y,z]And sent to the underground metal object positioning sub-model building module.
Wherein, step 101 specifically includes:
in a region where an underground metal object is known to be buried, n nodes are set in advance at regular intervals.
Electromagnetic signals are transmitted at the n nodes by the handheld mobile probe device and electromagnetic induction signals are received at the n nodes.
Step 102: and preprocessing the electromagnetic induction signals in the sample data in the data set to obtain a preprocessed data set.
Wherein, step 102 specifically comprises:
and normalizing the electromagnetic induction signals in the sample data in the data set by adopting a linear function Normalization (Min-Max Scaling) method or a zero mean Normalization (Z-Score Normalization) method to obtain a preprocessed data set.
For electromagnetic induction signal a = [ s ]1,s2,…,sn]The method used for preprocessing can be linear function normalization:
wherein, anormRepresents a normalized electromagnetic induction signal vector,a minis the minimum value in the original electromagnetic induction signal a,a maxis the maximum value in the original electromagnetic induction signal a.
Step 103: and constructing a plurality of different underground metal object positioning sub-models, wherein the underground metal object positioning sub-models are machine learning models.
The sub-models of the underground metal object positioning comprise a support vector regression model and a neural network regression model.
Step 104: and respectively training underground metal object positioning sub-models in various regions by adopting the preprocessed data set and taking the preprocessed electromagnetic induction signals as input and corresponding underground metal object coordinates as output to obtain a plurality of corresponding trained underground metal object positioning sub-models.
Wherein, step 104 specifically includes:
and dividing the preprocessed data set into a training set and a verification set according to a ratio of 7: 3.
And for each underground metal object positioning sub-model, adjusting parameters in the underground metal object positioning sub-model by using a cross validation method based on a training set and a validation set, training by using the training set to obtain the underground metal object positioning sub-model, and evaluating on the validation set until obtaining the underground metal object positioning sub-model with the minimum positioning mean square error loss.
Step 105: and weighting and adding the trained underground metal object positioning sub-models to obtain an integrated underground metal object positioning model, wherein the sum of the weighted weights is 1.
Wherein, step 105 specifically comprises: setting different weights for m underground metal object positioning submodels, wherein each underground metal object positioning submodelf j(a) Is represented by a weight ofw jThe cumulative sum of all weights is 1, i.e.w 1+w 2+…+w m =1，j∈[1,m]Then, multiplying the weight by the underground metal object positioning sub-model respectively and accumulating to obtain a final underground metal object positioning model:
f(a)=w 1 f 1(a)+w 2 f 2(a)+…+w m f m (a)。
step 106: and aiming at the minimum positioning error, adjusting each weight in the integrated underground metal object positioning model by adopting a particle swarm algorithm to obtain the underground metal object real-time positioning model.
Wherein, step 106 specifically includes: and the weight is adaptively adjusted by utilizing the data set and adopting a particle swarm algorithm, so that the positioning error of the underground metal object positioning model on the sample data set is minimum after the weight is adjusted.
The process of adjusting the weight by using the particle swarm optimization is shown in fig. 2, and a particle swarm is initialized, including the number of particles and the positions and the velocities thereof, wherein the number of particles is set to be 30, and the positions and the velocities of the particles are set to be m-dimensional vectors. Wherein the position of the particle corresponds to the weight vector of the subsurface metal object localization sub-modelw 1,w 2,…,w m]The speed range is limited to-3, 3]Within range, fitness functionFSetting the value of the underground metal object positioning model and the negative value of the square error loss function of the position coordinates of the underground metal object as follows:
F=-（w 1 f 1(a)+w 2 f 2(a)+…+w m f m (a)-[x,y,z])2；
and evaluating the fitness of the particles by using a fitness function, and if the current fitness value is higher, updating the historical optimal position of the individual by using the current position. And simultaneously comparing the current adaptive value of each particle with the global optimum position, if the current adaptive value is higher, updating the global optimum position by using the current position, then updating the speed and the position of each particle according to the speed and the position updating formula of the particle, then judging whether the ending condition is met or not, namely the maximum iteration times is reached, if the ending condition is met, ending, otherwise, continuously calculating the fitness corresponding to the particle.
Wherein the particlesiTo (1) adThe velocity and position update formula of the dimension is:
in the formula (I), the compound is shown in the specification,
represents the d-dimensional component of the flight velocity vector of the k-th iteration particle i,
represents the d-dimension component of the flight velocity vector of the (k-1) -th iteration particle i,
the d-th dimension component of the position vector of the particle i at the k-th iteration,
the d-dimensional component of the location vector of the particle i at the (k-1) th iteration,c 1andc 2the acceleration constant is used to adjust the learning maximum step size,r 1andr 2is two random functions with the value range of [0,1 ]]And is used for increasing the randomness of the random,wwhich is an inertial weight, is a non-negative number used to adjust the search range for the solution space.
Step 107: preprocessing the acquired electromagnetic induction signals of a plurality of nodes in the range of the area to be detected, and inputting the preprocessed signals into the real-time positioning model of the underground metal object to obtain the position coordinates of the underground metal object in the range of the area to be detected.
And electromagnetic induction signals of n nodes in the range of the area to be detected are acquired by the online acquisition sub-module. The online acquisition submodule comprises a transmitting module, a receiving module and a data preprocessing module.
The on-line sub-acquisition module is mainly used for transmitting electromagnetic signals by the transmitting module at n set nodes and receiving induction signal intensity data [ s ] by the receiving module when detecting underground metal objects at unknown positions in practical application1,s2,…,sn]Then the induction signal intensity data is preprocessed by the data preprocessing block [ s ]1’,s2’,…,sn’]And sent to the underground metal object real-time positioning module.
The invention discloses a positioning method of an underground metal object based on machine learning, which is characterized in that an underground metal object positioning sub-model is constructed by adopting various machine learning methods based on sample data, then a plurality of positioning sub-models are combined in a weighting mode, the setting of the weight value of each positioning sub-model is determined based on the sample data by adopting a particle algorithm, according to the finally obtained underground metal object positioning model with excellent generalization performance, the obtained induction signal intensity data is input into the underground metal object positioning model during actual detection and positioning, so that the positioning of the underground metal object can be realized, and the efficiency and the precision of the positioning are further improved by the method.
Compared with the traditional method for estimating the position of the underground metal target by minimizing the mean square error function between the fitting value and the detection value calculated by the physical model, the method disclosed by the invention can train the model by using sample data in an off-line stage under the condition of not solving the physical model, estimate the coordinates of the target position by using the trained model during real-time positioning, has low requirement on the calculation capability of positioning equipment, has higher positioning speed, has higher positioning precision in the presence of environmental noise and has higher robustness. In addition, because the method adopts a plurality of machine learning models to train and obtain the underground metal object positioning sub-model and finally integrates the underground metal object positioning sub-model into the underground metal object positioning model, compared with the method using a single machine learning model, the generalization performance of the underground metal object positioning model is enhanced.
Fig. 3 is a schematic structural diagram of a positioning system of a subsurface metal object based on machine learning according to the present invention, and as shown in fig. 3, the positioning system of a subsurface metal object based on machine learning includes:
a data set obtaining module 201, configured to obtain a data set; the ith sample data in the data set comprises a plurality of electromagnetic induction signals acquired from different positions of the ith area and underground metal object position coordinates of the ith area.
The data preprocessing module 202 is configured to preprocess the electromagnetic induction signal in each sample data in the data set, so as to obtain a preprocessed data set.
The underground metal object positioning sub-model building module 203 is used for building a plurality of different underground metal object positioning sub-models, and the underground metal object positioning sub-models are machine learning models.
And the underground metal object positioning sub-model training module 204 is used for training underground metal object positioning sub-models in various regions by using the preprocessed data set and using a plurality of preprocessed electromagnetic induction signals as input and corresponding underground metal object coordinates as output to obtain a plurality of trained underground metal object positioning sub-models.
And the underground metal object positioning model integrating module 205 is configured to weight and add the trained underground metal object positioning sub-models to obtain an integrated underground metal object positioning model, where the sum of the weighted weights is 1.
And the real-time positioning model determining module 206 for the underground metal object is configured to adjust each weight in the integrated underground metal object positioning model by using a particle swarm algorithm with the minimum positioning error as a target to obtain the real-time positioning model of the underground metal object.
The underground metal object real-time positioning model application module 207 is configured to input the acquired electromagnetic induction signals of the plurality of nodes in the region to be detected into the underground metal object real-time positioning model after preprocessing, and obtain the position coordinates of the underground metal object in the region to be detected.
The sub-models of the underground metal object positioning comprise a support vector regression model and a neural network regression model.
The data preprocessing module 202 specifically includes:
and the data preprocessing unit is used for carrying out normalization processing on the electromagnetic induction signals in the sample data in the data set by adopting a linear function normalization method or a zero-mean normalization method to obtain a preprocessed data set.
The underground metal object locator model training module 204 specifically includes:
and the data set dividing unit is used for dividing the preprocessed data set into a training set and a verification set according to the proportion of 7: 3.
And the underground metal object positioning sub-model training unit is used for adjusting parameters in the underground metal object positioning sub-model by using a cross validation method based on the training set and the validation set for each underground metal object positioning sub-model to obtain a trained underground metal object positioning sub-model.
The data set obtaining module 201 specifically includes:
and the node presetting unit is used for presetting n nodes at regular intervals in an area where the underground metal object is known to be buried.
And the electromagnetic induction signal acquisition unit is used for transmitting electromagnetic signals at the n nodes through the handheld mobile detection equipment and receiving the electromagnetic induction signals at the n nodes.
The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the description of the method part.
The principles and embodiments of the present invention have been described herein using specific examples, which are provided only to help understand the method and the core concept of the present invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, the specific embodiments and the application range may be changed. In view of the above, the present disclosure should not be construed as limiting the invention.
Claims (8)
1. A method for locating a subsurface metal object based on machine learning, comprising:
acquiring a data set; the ith sample data in the data set comprises a plurality of electromagnetic induction signals acquired from different places of the ith area and the underground metal object position coordinates of the ith area;
preprocessing electromagnetic induction signals in each sample data in the data set to obtain a preprocessed data set;
constructing a plurality of different underground metal object positioning sub-models, wherein the underground metal object positioning sub-models are machine learning models;
respectively training underground metal object positioning sub-models in various places by using a preprocessed data set and taking a plurality of preprocessed electromagnetic induction signals as input and corresponding underground metal object coordinates as output to obtain a plurality of corresponding trained underground metal object positioning sub-models;
weighting and adding the trained underground metal object positioning sub-models to obtain an integrated underground metal object positioning model, wherein the sum of the weighted weights is 1;
aiming at the minimum positioning error, adjusting each weight in the integrated underground metal object positioning model by adopting a particle swarm algorithm to obtain an underground metal object real-time positioning model;
preprocessing the collected electromagnetic induction signals of a plurality of nodes in the range of the area to be detected, and inputting the preprocessed electromagnetic induction signals into the underground metal object real-time positioning model to obtain the position coordinates of the underground metal object in the range of the area to be detected;
the acquiring the data set specifically includes:
presetting n nodes at regular intervals in an area where an underground metal object is known to be buried;
electromagnetic signals are transmitted at the n nodes by the handheld mobile probe device and electromagnetic induction signals are received at the n nodes.
2. A method for machine learning-based location of subsurface metal objects as claimed in claim 1 wherein each of said subsurface metal object location sub-models comprises a support vector regression model and a neural network regression model.
3. The machine-learning-based underground metal object positioning method according to claim 1, wherein the preprocessing of the electromagnetic induction signals in each sample data in the data set to obtain a preprocessed data set specifically comprises:
and normalizing the electromagnetic induction signals in the sample data in the data set by adopting a linear function normalization method or a zero-mean normalization method to obtain a preprocessed data set.
4. The machine learning-based underground metal object positioning method according to claim 1, wherein the using of the preprocessed data set, taking a plurality of preprocessed electromagnetic induction signals as input, and taking corresponding underground metal object coordinates as output, respectively trains underground metal object positioning sub-models in various regions to obtain a plurality of corresponding trained underground metal object positioning sub-models, specifically comprises:
dividing the preprocessed data set into a training set and a verification set according to the proportion of 7: 3;
and for each underground metal object positioning sub-model, adjusting parameters in the underground metal object positioning sub-model by using a cross validation method based on the training set and the validation set to obtain the trained underground metal object positioning sub-model.
5. A machine learning based locating system for underground metallic objects, comprising:
the data set acquisition module is used for acquiring a data set; the ith sample data in the data set comprises a plurality of electromagnetic induction signals acquired from different places of the ith area and the underground metal object position coordinates of the ith area;
the data preprocessing module is used for preprocessing electromagnetic induction signals in the sample data in the data set to obtain a preprocessed data set;
the underground metal object positioning sub-model building module is used for building a plurality of different underground metal object positioning sub-models, and the underground metal object positioning sub-models are machine learning models;
the underground metal object positioning sub-model training module is used for training underground metal object positioning sub-models in various places by adopting a preprocessed data set and taking a plurality of preprocessed electromagnetic induction signals as input and corresponding underground metal object coordinates as output to obtain a plurality of trained underground metal object positioning sub-models;
the underground metal object positioning model integration module is used for weighting and adding the trained underground metal object positioning sub-models to obtain an integrated underground metal object positioning model, and the sum of the weighted weights is 1;
the underground metal object real-time positioning model determining module is used for adjusting each weight in the integrated underground metal object positioning model by adopting a particle swarm algorithm to obtain an underground metal object real-time positioning model by taking the minimum positioning error as a target;
the underground metal object real-time positioning model application module is used for preprocessing the collected electromagnetic induction signals of a plurality of nodes in the range of the area to be detected and then inputting the preprocessed signals into the underground metal object real-time positioning model to obtain the position coordinates of the underground metal object in the range of the area to be detected;
the data set acquisition module specifically includes:
the node presetting unit is used for presetting n nodes at a certain distance in an area where the underground metal object is known to be buried;
and the electromagnetic induction signal acquisition unit is used for transmitting electromagnetic signals at the n nodes through the handheld mobile detection equipment and receiving the electromagnetic induction signals at the n nodes.
6. A machine learning based localization system for subsurface metal objects as claimed in claim 5, wherein each of said subsurface metal object localization submodels comprises a support vector regression model and a neural network regression model.
7. A machine learning based localization system for a subsurface metal object as claimed in claim 5, wherein the data preprocessing module specifically comprises:
and the data preprocessing unit is used for carrying out normalization processing on the electromagnetic induction signals in the sample data in the data set by adopting a linear function normalization method or a zero-mean normalization method to obtain a preprocessed data set.
8. The machine learning-based underground metal object positioning system of claim 5, wherein the underground metal object positioning sub-model training module specifically comprises:
the data set dividing unit is used for dividing the preprocessed data set into a training set and a verification set according to the proportion of 7: 3;
and the underground metal object positioning sub-model training unit is used for adjusting parameters in the underground metal object positioning sub-model by using a cross validation method based on the training set and the validation set to obtain the trained underground metal object positioning sub-model for each underground metal object positioning sub-model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210314400.4A CN114415246B (en) | 2022-03-29 | 2022-03-29 | Underground metal object positioning method and system based on machine learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210314400.4A CN114415246B (en) | 2022-03-29 | 2022-03-29 | Underground metal object positioning method and system based on machine learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114415246A CN114415246A (en) | 2022-04-29 |
| CN114415246B true CN114415246B (en) | 2022-06-24 |
Family
ID=81263993
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210314400.4A Active CN114415246B (en) | 2022-03-29 | 2022-03-29 | Underground metal object positioning method and system based on machine learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114415246B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115795353B (en) * | 2023-02-06 | 2023-04-21 | 北京科技大学 | Underground metal target classification method and system based on unbalanced data set |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109782360A (en) * | 2019-02-19 | 2019-05-21 | 核工业航测遥感中心 | A kind of low-resistance area of coverage deep sand body detection method |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107703480B (en) * | 2017-08-28 | 2021-03-23 | 南京邮电大学 | Mixed kernel function indoor positioning method based on machine learning |
| CN107876773A (en) * | 2017-12-27 | 2018-04-06 | 科大天工智能装备技术（天津）有限公司 | A kind of power spreading device of powder bed increase and decrease material combined-machining equipment |
| CN109188533B (en) * | 2018-07-20 | 2020-08-14 | 中国科学院信息工程研究所 | Detection method, device and equipment of electronic equipment |
| US11199643B2 (en) * | 2019-09-23 | 2021-12-14 | Halliburton Energy Services, Inc. | Machine learning approach for identifying mud and formation parameters based on measurements made by an electromagnetic imager tool |
| CN112735097A (en) * | 2020-12-29 | 2021-04-30 | 中国地质环境监测院 | Regional landslide early warning method and system |
| CN113702719B (en) * | 2021-08-03 | 2022-11-29 | 北京科技大学 | Broadband near-field electromagnetic positioning method and device based on neural network |
-
2022
- 2022-03-29 CN CN202210314400.4A patent/CN114415246B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109782360A (en) * | 2019-02-19 | 2019-05-21 | 核工业航测遥感中心 | A kind of low-resistance area of coverage deep sand body detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114415246A (en) | 2022-04-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110686677B (en) | Global positioning method based on geometric information | |
| Lin et al. | An improved transformed unscented FastSLAM with adaptive genetic resampling | |
| CN109937341A (en) | The estimation of the position of oneself | |
| CN114415246B (en) | Underground metal object positioning method and system based on machine learning | |
| CN108828501B (en) | Method for real-time tracking and positioning of mobile sound source in indoor sound field environment | |
| CN105975967A (en) | Target positioning method and system | |
| CN110766067A (en) | Millimeter wave radar material identification method based on machine learning | |
| US8194501B2 (en) | Apparatus and method of sniper localization | |
| CN111695461B (en) | Radar clutter intelligent classification method based on image frequency characteristics | |
| CN111901749A (en) | High-precision three-dimensional indoor positioning method based on multi-source fusion | |
| CN112560981B (en) | Training method, device, apparatus, program, and storage medium for generating countermeasure model | |
| Khassanov et al. | Finer-level sequential wifi-based indoor localization | |
| Souli et al. | Relative positioning of autonomous systems using signals of opportunity | |
| Siyang et al. | WKNN indoor Wi-Fi localization method using k-means clustering based radio mapping | |
| Labinghisa et al. | A deep learning based scene recognition algorithm for indoor localization | |
| Lee et al. | An adaptive sensor fusion framework for pedestrian indoor navigation in dynamic environments | |
| CN113923776B (en) | System and method for confirming shielding target position and self-adaptive interference of unmanned aerial vehicle | |
| Magrin et al. | Hierarchical sensor fusion method based on fingerprint knn and fuzzy features weighting for indoor localization of a mobile robot platform | |
| CN109803234A (en) | Unsupervised fusion and positioning method based on the constraint of weight different degree | |
| Guo et al. | Accurate indoor localization with crowd sensing | |
| KR20200043146A (en) | Apparatus for adaptive distace measurement base on artificial neural networks | |
| Kao et al. | Feature extraction from a broadband sonar sensor for mapping structured environments efficiently | |
| JP6819797B2 (en) | Position estimation device, position estimation method and program, and position estimation system | |
| Kjellson et al. | Accurate indoor positioning based on learned absolute and relative models | |
| CN113613188A (en) | Fingerprint library updating method and device, computer equipment and storage medium |
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 |