Mine landslide mass soil moisture content measuring method based on air-coupled SFCW technology

Abstract

The invention provides a mine landslide mass soil moisture content measuring method based on an air-coupled SFCW technology, and belongs to the technical field of mine physical property parameter monitoring. The method mainly comprises the steps that an air-coupled SFCW radar is used for scanning landslide mass soil and collecting data; performing noise reduction processing on the radar data; calculating reflection coefficients of soil layers with different humidity; a precise soil moisture content measuring device is used for obtaining the corresponding moisture content, and a reflection coefficient-moisture content data set is generated; establishing a feature database, and performing fitting training on the data by using a support vector machine model; and finally predicting the moisture content by using the trained model. According to the method, the moisture content of the landslide mass soil can be obtained in real time without damage, and timely landslide early warning and prevention are facilitated. In addition, the method can also be applied to other fields where the soil moisture content needs to be monitored, such as agricultural irrigation, lake protection and the like.

Description

Technical Field

[0001] The invention belongs to the technical field of mine physical property parameter monitoring, and more specifically relates to a method for measuring soil moisture content of a mine landslide body based on an air-coupled SFCW technology. Background Art

[0002] In the field of engineering exploration, monitoring soil moisture content is an important part of ensuring the quality of geotechnical engineering and the safety of construction equipment. At present, there are many methods for measuring soil moisture content, among which the more typical ones include sampling and drying method, time domain reflectometry (TDR), frequency domain reflectometry (FDR), neutron method and capacitance method, but these measurement methods are basically contact measurement methods. Faced with the complex environment of underground mines, contact measurement methods have certain limitations. Insertion measurement devices not only damage the soil, but also have low measurement efficiency. Most of them can only be used for small-scale or single-point soil moisture measurement.

[0003] With the development of science and technology, radar, as a new, fast, and non-destructive method for mesoscale soil moisture detection, has attracted the attention of many scholars at home and abroad. The United States first proposed the use of ground penetrating radar for soil moisture detection. Subsequently, scientists in the United States and Finland conducted in-depth research on this. my country has gradually begun to study ground penetrating radar in soil moisture measurement since 1990. In recent years, most of the research used for soil moisture monitoring is based on narrow-band pulse ground penetrating radar. Studies have found that this pulse has a narrow bandwidth and requires multiple groups of radar antennas to be used in combination to perform full-range detection at different depths. The detection resolution is low, and the signal is easily affected by noise during the detection process. The SFCW signal can adjust the frequency range and step size of the signal according to specific needs. Compared with the pulse signal, it has stronger noise resistance when used in ground penetrating radar.

[0004] Based on the above research background, the present invention proposes a method for measuring soil moisture content of mine landslide body based on air-coupled SFCW. The present invention applies this method to the non-contact real-time monitoring of the moisture content of mine landslide body deposits. Timely exploration of the moisture status of the landslide body deposits at the mine exit helps the inspection personnel to make supplementary corrections to the safety management and preventive control measures of the mine slope, which is of great significance to the safe production and intelligent construction of mines. Summary of the invention

[0005] The problem to be solved by the present invention is to overcome the shortcomings of traditional contact soil moisture monitoring technology, and to propose a method for measuring the soil moisture content of a mine landslide based on air-coupled SFCW technology for the specific object of soil moisture content of a mine landslide.

[0006] In order to achieve the above object, the present invention is implemented by adopting the following technical scheme: the method comprises:

[0007] S1. Use air-coupled SFCW radar to scan the actual landslide soil and collect data;

[0008] S2, performing signal noise reduction processing on radar data;

[0009] S3, calculating the reflection coefficient R of soil layers with different humidity for the noise reduction radar signal;

[0010] S4, using a high-precision soil moisture instrument to measure the corresponding moisture content to generate a reflectance-moisture content data set;

[0011] S5, establish a feature database and use the support vector machine model SVM for data fitting. The training only relies on the support vector and is suitable for small sample data;

[0012] S6. Use the trained model to predict and invert to obtain the moisture content value.

[0013] In one solution, the specific implementation process of S1 is as follows:

[0014] According to electromagnetic wave theory, high-frequency electromagnetic waves propagate in the medium and satisfy Maxwell's equations. The differential form of Maxwell's equations is as follows:

[0015]

[0016] Among them, E is the electric field intensity; B is the magnetic induction intensity; H is the magnetic field intensity; D is the electric displacement; J is the current density; ρ is the charge density; and t is the pulse travel time.

[0017] In one embodiment, the S2 noise reduction process includes background removal, zero drift removal and bandpass filtering.

[0018] In one embodiment, the S3 reflection coefficient R process is as follows:

[0019] The relative distance between the transmitting antenna and the receiving antenna is x, the depth of the reflecting surface is h, the travel time of the reflected wave in the soil is t, and the propagation speed in the soil is v;

[0020] The propagation velocity v of the reflected wave in the soil is calculated as:

[0021]

[0022] According to the relationship between propagation velocity and dielectric constant, the relative dielectric constant ε can be calculated as:

[0023]

[0024] Assuming that the propagation direction of the electromagnetic wave is from medium 1 to medium 2, the relative dielectric constants of medium 1 and medium 2 are ε1 and ε2, the reflection coefficient R has the following relationship with the dielectric constant:

[0025]

[0026] In one embodiment, the S5 comprises the following steps:

[0027] S501, obtaining multiple sets of reflection coefficient-water content corresponding values ​​through experiments, and importing the data sets into a support vector machine model;

[0028] S502, preprocessing the data set, including removing missing values, processing outliers, and dividing the training set into a test set;

[0029] S503, extracting features related to water content from the acquired reflection coefficient data set, including reflection coefficient values, dielectric constant values, and reflection coefficient differences obtained through processing; encoding the extracted feature values ​​and constructing a feature database, and finally associating them with corresponding water content data;

[0030] S504, using the support vector machine algorithm to build a fitting model to train and test the feature database, determine whether the fitting result meets the accuracy requirement, and obtain the optimal prediction model by adjusting the weight parameters in the model;

[0031] S505: The obtained optimal model is fitted and predicted and the result is output.

[0032] In one embodiment, during the radar scanning and data collection process of S1, multi-point and multi-depth collection is performed simultaneously to obtain a more comprehensive soil moisture situation.

[0033] Beneficial effects of the present invention:

[0034] 1. The air-coupled SFCW radar technology is used to monitor the soil moisture content of mine landslides. The transmitted signal is a stepped frequency continuous wave signal. Compared with the pulse signal, the stepped frequency continuous wave signal can adjust the frequency range and step length of the signal according to specific needs to reduce the impact of interference signals. It has strong anti-interference ability and the ability to expand bandwidth.

[0035] 2. Compared with other contact measurement methods, the use of air-coupled SFCW radar technology to monitor the soil moisture content of mine landslides can achieve non-contact large-scale measurement, overcoming the limitations of current contact measurement methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 This is a flow chart of the soil moisture content measurement method of mine landslide based on air-coupled SFCW;

[0037] Figure 2 This is a schematic diagram of the specific measurement of the air-coupled SFCW radar;

[0038] Figure 3 Flowchart for support vector machine (SVM) model training. DETAILED DESCRIPTION

[0039] In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Typical embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described in the present invention. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

[0040] Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0041] Figure 1 This is a flow chart for accurately measuring the deep soil moisture content of a mine landslide using air-coupled SFCW ground-penetrating radar technology. The present invention first uses air-coupled SFCW ground-penetrating radar to conduct a deep soil moisture content test to collect radar echo signals, then performs image processing and signal denoising on the collected radar data, and then calculates the soil reflection coefficient, and finally uses a support vector machine to perform parameter fitting on the data set, thereby inverting the deep soil moisture content. In a specific embodiment, the specific implementation steps of the method for measuring the soil moisture content of a mine landslide using air-coupled SFCW radar are as follows:

[0042] S1. Use air-coupled SFCW radar to scan the actual landslide soil and collect data;

[0043] Ground penetrating radar uses high-frequency electromagnetic waves to detect the inside of the medium. According to electromagnetic wave theory, the propagation of high-frequency electromagnetic waves in the medium satisfies Maxwell's equations. The differential form of Maxwell's equations is as follows:

[0044]

[0045] Among them, E is the electric field intensity; B is the magnetic induction intensity; H is the magnetic field intensity; D is the electric displacement; J is the current density; ρ is the charge density; and t is the pulse travel time.

[0046] SFCW radar is also called stepped frequency radar. It divides the required frequency band into several narrow frequency bands within a certain period of time, and transmits and receives in the form of these narrow frequency bands. Then, the received echo signal is processed by mixing and other processes to generate synthetic pulses for research.

[0047] S2: Perform signal noise reduction processing based on the radar data collected by S1, including background removal, zero drift removal and bandpass filtering.

[0048] Background removal,For the collected radar data, the background noise needs to be removed first.,For example, this can be done by averaging the idle data and then subtracting the average value from all the data, thus removing the background interference.

[0049] Zero drift removal: For the drift influence of the environment or equipment, zero drift removal processing can be performed. For example, a location without target reflection (such as the air area above the antenna) is selected and its measured value is used as the zero point to remove zero drift interference.

[0050] Bandpass filtering is a frequency selection process. Only signals within the frequency range set by the bandpass filter can pass through. It can suppress noise that is not within this frequency range, thereby achieving the purpose of noise reduction.

[0051] S3, based on the noise reduction radar signal obtained in S2, the reflection coefficient R of soil layers with different humidity is calculated. Assume that the relative distance between the transmitting antenna and the receiving antenna is x, the depth of the reflecting surface is h, the travel time of the reflected wave in the soil is t, and the propagation speed in the soil is v. The specific measurement diagram is as follows Figure 2 As shown;

[0052] according to Figure 2 The relationship between the propagation velocity v of the reflected wave in the soil can be calculated as:

[0053]

[0054] According to the relationship between propagation velocity and dielectric constant, the relative dielectric constant ε can be calculated as:

[0055]

[0056] Assuming that the propagation direction of the electromagnetic wave is from medium 1 to medium 2, the relative dielectric constants of medium 1 and medium 2 are ε1 and ε2, the reflection coefficient R has the following relationship with the dielectric constant:

[0057]

[0058] In the embodiment of the present invention, the method for extracting the reflection coefficient R is to obtain v based on the depth and time information on the SFCW ground penetrating radar image, and then enter it into formula (6) to calculate the dielectric constant ε of the soil layer, and then calculate the reflection coefficient R on each reflection surface.

[0059] S3: Conduct experiments, process the collected radar signals based on S2 and S3, calculate the reflection coefficient R, and use a high-precision soil moisture instrument to measure the corresponding moisture content to generate a reflection coefficient-moisture content data set.

[0060] S5. Based on the reflectance-water content data set generated by S4, a feature database is then established and the support vector machine model (SVM) is used for data fitting. The support vector machine can effectively handle nonlinear problems. The model supports a variety of kernel functions and has great flexibility. Since the actual number of experiments is limited, the sample size of the data set is also limited. SVM is based on the principle of interval maximization and only relies on support vectors during training, which is very suitable for small sample data.

[0061] The S5 is based on the support vector machine model to fit the small sample reflectance-water content data set. The flow chart is as follows Figure 3 As shown, the specific steps include:

[0062] S501, obtaining multiple sets of reflection coefficient-water content corresponding values ​​through experiments, and importing the data sets into a support vector machine model;

[0063] S502: After the reflectance-water content data set is obtained based on S501, the data set needs to be preprocessed first, including removing missing values, processing outliers, and dividing the training set into a test set.

[0064] S503. Then, based on S502, features related to water content are extracted from the acquired reflection coefficient data set through observation and analysis. These features include but are not limited to reflection coefficient values, dielectric constant values, reflection coefficient differences, etc. obtained through processing. The extracted feature values ​​are feature encoded and a feature database is constructed, and finally associated with the corresponding water content data.

[0065] S504. Based on the feature database and moisture content information established in S503, a fitting model is built using the support vector machine algorithm to train and test the feature database, and whether the fitting result meets the accuracy requirements is judged. The optimal prediction model is obtained by adjusting the weight parameters in the model.

[0066] S505 , performing fitting prediction based on the optimal model obtained in S504 and outputting the result.

[0067] S6. Predict and invert the water content value based on the model trained in S505.

[0068] Firstly, the trained support vector machine model is used to input unknown samples (including a series of characteristic values ​​such as soil reflectance) into the model.

[0069] During the model calculation process, pre-trained weights and bias parameters are used, which are obtained through previous training data.

[0070] Through the calculation of the support vector machine model, the prediction result of the unknown sample, that is, the moisture content of the soil, can be obtained.

[0071] The predicted results can be analyzed and inverted to obtain the water content value. The water content value can be obtained directly, or it can be inverted into corresponding geological or environmental parameters according to actual needs.

[0072] According to actual needs, the predicted moisture content value is further analyzed and utilized, such as for early warning and prevention of mine landslides.

[0073] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0074] It should be understood that the detailed description of the technical solutions of the present invention by means of the preferred embodiments is illustrative rather than restrictive. A person skilled in the art may modify the technical solutions described in the embodiments, or replace some of the technical features by equivalents, based on reading the specification of the present invention; and these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for measuring soil moisture content of mine landslide based on air-coupled SFCW technology, characterized by: The method includes: S1. Use air-coupled SFCW radar to scan the actual landslide soil and collect data; S2, performing signal noise reduction processing on radar data; S3, calculating the reflection coefficient R of soil layers with different humidity for the noise reduction radar signal; S4, using a high-precision soil moisture instrument to measure the corresponding moisture content to generate a reflectance-moisture content data set; S5, establish a feature database and use the support vector machine model SVM for data fitting. The training only relies on the support vector and is suitable for small sample data; S6. Use the trained model to predict and invert to obtain the moisture content value.

2. The method for measuring soil moisture content of a mine landslide body based on air-coupled SFCW technology according to claim 1 is characterized in that: The specific implementation process of S1 is as follows: The propagation of high-frequency electromagnetic waves in the medium satisfies Maxwell's equations. The differential form of Maxwell's equations is as follows: Among them, E is the electric field intensity; B is the magnetic induction intensity; H is the magnetic field intensity; D is the electric displacement; J is the current density; ρ is the charge density; and t is the pulse travel time.

3. The method for measuring soil moisture content of a mine landslide body based on air-coupled SFCW technology according to claim 1 is characterized in that: The S2 noise reduction process includes background removal, zero drift removal and bandpass filtering.

4. The method for measuring soil moisture content of a mine landslide body based on air-coupled SFCW technology according to claim 1 is characterized in that: The S3 reflection coefficient R process is as follows: The relative distance between the transmitting antenna and the receiving antenna is x, the depth of the reflecting surface is h, the travel time of the reflected wave in the soil is t, and the propagation speed in the soil is v; The propagation velocity v of the reflected wave in the soil is calculated as: According to the relationship between propagation velocity and dielectric constant, the relative dielectric constant ε is calculated as: Assuming that the propagation direction of the electromagnetic wave is from medium 1 to medium 2, the relative dielectric constants of medium 1 and medium 2 are ε1 and ε2, the reflection coefficient R has the following relationship with the dielectric constant:

5. The method for measuring soil moisture content of a mine landslide body based on air-coupled SFCW technology according to claim 1 is characterized in that: The S5 comprises the following steps: S501, obtaining multiple sets of reflection coefficient-water content corresponding values ​​through experiments, and importing the data sets into a support vector machine model; S502, preprocessing the data set, including removing missing values, processing outliers, and dividing the training set into a test set; S503, extracting features related to water content from the acquired reflection coefficient data set, including reflection coefficient values, dielectric constant values, and reflection coefficient differences obtained through processing; encoding the extracted feature values ​​and constructing a feature database, and finally associating them with corresponding water content data; S504, using the support vector machine algorithm to build a fitting model to train and test the feature database, determine whether the fitting result meets the accuracy requirement, and obtain the optimal prediction model by adjusting the weight parameters in the model; S505: The obtained optimal model is fitted and predicted and the result is output.

6. The method for measuring soil moisture content of a mine landslide body based on air-coupled SFCW technology according to claim 1 is characterized in that: During the radar scanning and data collection process of S1, multi-point and multi-depth collection is performed simultaneously to obtain a more comprehensive soil moisture situation.

Images