Sandy soil movement monitoring method and sandy soil movement monitoring system based on magnetic field positioning

Abstract

The invention relates to a sandy soil motion monitoring method and a sandy soil motion monitoring system based on magnetic field localization, comprising to-be-monitored sandy soil, a plurality of marked magnetic particles implanted in the sandy soil, a sensor array arranged around or above a to-be-monitored area, the sensor array comprising at least three magnetic field sensors, and the number of the magnetic field sensors is the same as that of the marked magnetic particles. And through the magnetic field positioning technology, the limitation of the traditional technology can be overcome, and real-time high-precision monitoring is realized.

Description

Technical Field

[0001] The present invention relates to a sand movement monitoring method and a sand movement monitoring system based on magnetic field positioning, belonging to the technical field of real-time monitoring of sand movement. Background Art

[0002] Sand and soil movement monitoring has a wide range of applications in disaster prevention and mitigation, geological engineering, and infrastructure construction. Traditionally, common sand and soil monitoring technologies include visual perception, acoustic sensing, and displacement monitoring using mechanical sensors. However, these methods are only effective in specific applications and are prone to environmental interference, loss of accuracy after long-term monitoring, and difficulty adapting to complex environments.

[0003] In the existing technology, some studies have attempted to use labeled particles and optical tracking methods to monitor sand movement, but these methods are affected by line of sight obstruction, particle loss, and changes in light conditions, making it difficult to ensure accuracy in dynamic and complex environments. The so-called line of sight limitation means that visual sensors are easily affected by obstructions and changes in light conditions, resulting in reduced monitoring accuracy. Environmental interference means that acoustic and mechanical sensors are easily affected by noise and vibration in complex sand environments, making it difficult to maintain stable monitoring results. Accuracy limitations mean that existing technologies have difficulty maintaining high accuracy during long-term monitoring, and they perform poorly when faced with tiny movement changes in sand particles. Finally, in existing monitoring methods, the monitoring instruments have a large impact on the surrounding environment, making it difficult to objectively monitor the results of sand movement.

[0004] Therefore, it is necessary to study a new technical means to break through the above limitations and achieve real-time and high-precision monitoring of sand movement in complex environments. Summary of the Invention

[0005] The present invention provides a sand movement monitoring method and a sand movement monitoring system based on magnetic field positioning, which can overcome the limitations of traditional technologies and achieve real-time high-precision monitoring.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] A method for monitoring sand movement based on magnetic field positioning specifically comprises the following steps:

[0008] Step S1, selecting a number of labeled magnetic particles that meet the requirements and implanting the labeled magnetic particles into the sand in the monitoring area;

[0009] Step S2, setting a sensor array consisting of a plurality of magnetic field sensors, calibrating the sensor array, and placing the calibrated sensor array around or above the monitoring area;

[0010] Step S3, collecting real-time magnetic field data at different locations through a magnetic field sensor, and pre-processing the real-time magnetic field data by a data processing module to filter out environmental noise and interference;

[0011] Step S4: fusing the data from the magnetic field sensors processed in step S3, and using the magnetic field variation to invert the position information of the magnetic field source, the positioning algorithm module determines the specific position of each labeled magnetic particle in the monitoring area;

[0012] Step S5, tracking the specific position of the determined labeled magnetic particles in real time in the time series, recording the change of the position of the labeled magnetic particles, and generating a motion trajectory;

[0013] Step S6, analyzing the movement of sand in the monitoring area based on the movement trajectory of the labeled magnetic particles generated in step S5;

[0014] Furthermore, a plurality of magnetic field sensors are located at different heights and horizontal positions to form a sensor array;

[0015] Furthermore, in step S4, the specific steps of inverting the position information of the magnetic field source using the magnetic field change are as follows:

[0016] Step S41: construct a physical model, treat each labeled magnetic particle as a magnetic dipole, and define it as x i , the magnetic field sensor is defined as y j , then the magnetic particles are labeled x i The three-dimensional coordinate of i =(x i ,y i ,z i ), where i = 1, 2, ... n, magnetic field sensor y j The three-dimensional coordinate of j =(x j ,y j ,z j ), where j = 1, 2, ... n, n ≥ 3;

[0017] Step S42, constructing a magnetic dipole model, marking magnetic particles x i The magnetic field at a distance r is expressed as It is described as:

[0018]

[0019] In the formula, To label magnetic particles x i With magnetic field sensor j The distance between

[0020] μ0 is the vacuum permeability,

[0021] m i To label magnetic particles x i The magnetic moment vector of

[0022] Step S43: construct a measurement equation to determine the magnetic field sensor y in the sensor array. j All labeled magnetic particles x measured i The sum of the magnetic fields:

[0023]

[0024] In the formula, B j For all labeled magnetic particles x i The sum of the magnetic fields, n ≥ 3;

[0025] Step S44, based on each labeled magnetic particle x i Magnetic field and magnetic field sensory j The spatial position of the , establish the following relationship:

[0026]

[0027] In the formula, The measurement value is obtained by solving the nonlinear equation obtained by neural network training, and its measurement value is expressed as a set of unknown coordinates x i Nonlinear equations;

[0028] Step S45, by solving the nonlinear equation, obtain the labeled magnetic particles x i With magnetic field sensor j The distance between The three-dimensional coordinates of the labeled magnetic particles are determined by iterative solution through triangulation of the distance between each labeled magnetic particle and the magnetic field sensor;

[0029] Furthermore, the process of training the nonlinear equation through the neural network based on step S44 is as follows:

[0030] Step S441: Prepare training data set Among them, (B j ,m i ,y j ) is the input feature vector, is the target output, Represents all labeled magnetic particles x i The total magnetic field B j The predicted value of

[0031] Step S442: define a feedforward neural network, which includes an input layer, a hidden layer, and an output layer. The input layer inputs features (x i ,m i,y j ), after being transferred to the hidden layer, it is transformed by the nonlinear activation function, and the output of the hidden layer is further transferred to the output layer to obtain all the labeled magnetic particles x i The total magnetic field B j The predicted value of

[0032]

[0033] In the formula, f NN is a neural network model, θ is a trainable parameter in the neural network;

[0034] Step S443, calculate the predicted value and the true value B j The error between them is expressed as follows:

[0035]

[0036] In the formula, L is the loss function, N is the maximum number n of j, and N ≥ 3;

[0037] Step S444, use the back propagation algorithm to calculate the gradient of the loss function L with respect to the trainable parameter θ, that is,

[0038]

[0039] Step S445, update the weights of the neural network model by gradient descent or optimization algorithm, i.e.

[0040]

[0041] In the formula, η is the learning rate;

[0042] Step S446, repeat steps S442-S445 until the loss function converges to a preset standard minimum value or reaches a preset training round, completing the training process including the unknown coordinate x i Training of nonlinear equations;

[0043] Furthermore, in step S45, the specific process of solving the nonlinear equation is:

[0044] Step S451: Determine the three-dimensional coordinates of the labeled magnetic particles using a nonlinear equation. Here, the nonlinear equation is:

[0045]

[0046] In the formula, To label magnetic particles x i With magnetic field sensor j The distance between them is obtained according to the constructed magnetic dipole model, that is,

[0047] Step S452, using a nonlinear least squares method to solve the nonlinear equation in step S451, inputting the data of each magnetic field sensor into the solution, and obtaining the three-dimensional coordinates closest to each labeled magnetic particle;

[0048] Furthermore, based on the measurement equation constructed in step S43, the error correction method is used to correct the magnetic field sensor y j The measured labeled magnetic particles x i Adjust the magnetic field data;

[0049] Furthermore, in step S5, the specific position of each labeled magnetic particle in the monitoring area, i.e., the three-dimensional coordinates of each labeled magnetic particle, is determined according to step S4, and the change in the position of each labeled magnetic particle is recorded to generate its motion trajectory;

[0050] Furthermore, in step S6, based on the movement trajectory of the marked magnetic particles, the overall movement characteristics of the sand are analyzed, including the flow velocity, the change in scouring depth, and the stress condition of the soil;

[0051] A sand movement monitoring system adopts the sand movement monitoring method based on magnetic field positioning; the system includes sand to be monitored, a number of labeled magnetic particles are implanted in the sand, and a sensor array is arranged around or above the area to be monitored, the sensor array includes at least three magnetic field sensors, and the number of magnetic field sensors is the same as the number of labeled magnetic particles.

[0052] Through the above technical solution, compared with the existing technology, the present invention has the following beneficial effects:

[0053] 1. The sand movement monitoring method based on magnetic field positioning provided by the present invention is based on magnetic field detection and positioning algorithms of a multi-sensor array. It can work in irregular and complex sand environments and has strong anti-interference capabilities.

[0054] 2. The sand movement monitoring method based on magnetic field positioning provided by the present invention combines the data obtained from the sensor array with the processing method and algorithm for generating sand movement trajectories, which can accurately detect tiny displacement changes and achieve higher-precision sand movement monitoring;

[0055] 3. The sand movement monitoring system provided by the present invention is based on magnetic field positioning technology. Micro-marker particles with specific magnetic field characteristics are implanted into the sand for monitoring. It can maintain stable high-precision performance during long-term continuous monitoring. At the same time, it is not dependent on external light conditions and is suitable for long-term dynamic monitoring tasks. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The present invention will be further described below with reference to the accompanying drawings and examples.

[0057] Figure 1 is a flow chart of a monitoring method according to a preferred embodiment of the present invention;

[0058] Figure 2 is a system schematic diagram of a preferred embodiment provided by the present invention;

[0059] Figure 3 This is a result interface display diagram of a preferred embodiment provided by the present invention. DETAILED DESCRIPTION

[0060] The present invention will now be described in further detail with reference to the accompanying drawings. In the description of this application, it should be understood that the terms "left side", "right side", "upper", "lower", etc., indicating orientations or positional relationships, are based on the orientations or positional relationships shown in the accompanying drawings, and are intended only to facilitate the description of the present invention and simplify the description, and do not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. "First", "second", etc. do not indicate the importance of the components and therefore should not be understood as limiting the present invention. The specific dimensions used in this embodiment are only for illustrative purposes only and do not limit the scope of protection of the present invention.

[0061] As explained in the background technology, existing technologies, whether based on visual perception, acoustic wave perception, or mechanical sensor displacement monitoring, have also attempted to use labeled particles and optical tracking methods to monitor sand movement. However, these methods are susceptible to line of sight obstruction, particle loss, and changes in illumination, and cannot guarantee monitoring accuracy in dynamic and complex environments. Subsequent developments have also considered positioning technologies based on acoustic waves or methods that use optical imaging combined with deep learning for motion analysis. However, acoustic waves are also susceptible to environmental noise, and the monitoring accuracy of optical imaging will be significantly reduced when sand obstructs and ambient lighting conditions change.

[0062] Therefore, this application uses magnetic field positioning technology to provide a more accurate, stable, and suitable method for monitoring sand movement in complex environments. The entire monitoring method is to implant marker particles with micro-magnetic field characteristics in the sand and use magnetic field detection equipment to track the movement trajectory of the particles in real time. At the same time, the system is composed of multiple magnetic field sensor arrays, combined with an efficient positioning algorithm to calculate the three-dimensional position changes of the marker particles in the sand. Through the precise measurement of the magnetic field, it is possible to overcome interference from ambient light, noise, etc., and achieve accurate tracking of sand movement.

[0063] First, the present application provides a method for monitoring sand movement based on magnetic field positioning, which includes the following steps:

[0064] Step S1, selecting a number of labeled magnetic particles that meet the requirements and implanting the labeled magnetic particles into the sand in the monitoring area;

[0065] The criteria for selecting the marker magnetic particles here are that they should have strong magnetic field properties and be small enough to ensure they can move freely in the sand without affecting the fluidity of the sand. They are usually specially prepared magnets or tiny particles containing magnetic materials. These tiny particles have the following advantages: when analyzing sand samples in the laboratory, the magnetic particles can be quickly aggregated by applying a magnetic field, greatly improving the efficiency of sample processing. In complex sand environments, this marking method can accurately identify and locate sand in specific areas, helping to more accurately understand the dynamic behavior of sand.

[0066] The magnetic properties of magnetic particles make them extremely sensitive to changes in their surroundings. Even small changes in stress, strain, or displacement in the sand can cause a change in the particles' magnetism. This allows for the detection of subtle dynamic changes in the sand, improving monitoring accuracy. Magnetic particles exhibit excellent stability and durability, allowing them to persist in the sand for extended periods while maintaining their magnetic properties. This enables long-term dynamic monitoring of the sand, which is of great significance for studying the long-term dynamics of sand and geological evolution.

[0067] Once these labeled magnetic particles are implanted in the sandy soil of the area being monitored, they move with the movement of the sand. The unique properties of the magnetic particles allow them to accurately reflect the dynamics of the sandy soil even in complex environments, effectively eliminating environmental interference.

[0068] It should be noted that during the experiment, the present applicant found that at least three labeled magnetic particles need to be implanted for the detected data to be of reference significance.

[0069] Step S2: setting a sensor array consisting of a plurality of magnetic field sensors, calibrating the sensor array, and placing the calibrated sensor array around or above the monitoring area.

[0070] The number of magnetic field sensors deployed matches the number of labeled magnetic particles. Advanced magnetic monitoring equipment and technology enable remote monitoring of magnetic particles without the need for direct contact with sand samples. This has important applications in difficult-to-reach or dangerous areas, such as remote geological disaster monitoring sites and deep-sea sediment monitoring. The magnetic field sensors are arranged in an array around or above the sand monitoring area because, ideally, the sensor array should cover the entire three-dimensional space of the sand, accurately detecting changes in the position of the labeled magnetic particles.

[0071] It should be noted that to improve positioning accuracy, magnetic field sensors should be arranged at different heights and horizontal positions. The distance between magnetic field sensors should be adjusted according to the size of the sand monitoring area and the magnetic field strength of the labeled magnetic particles.

[0072] Once the magnetic field sensors and labeled magnetic particles are in place, each magnetic field sensor can measure the magnetic field strength around it in real time. As the labeled magnetic particles move through the sand, the distribution of the magnetic field changes, and the magnetic field sensor determines the position of the labeled magnetic particles by detecting these changes. By fusing data from multiple magnetic field sensors, the changes in the magnetic field can be used to infer the three-dimensional position of the labeled magnetic particles in the sand. The sensor array can capture the magnetic field strength at different locations and, by inverting the position information of the magnetic field source, determine the specific position of the labeled magnetic particles within the monitoring area, completing the three-dimensional positioning operation.

[0073] Of course, before actual monitoring, the sensor array needs to be calibrated to ensure the measurement accuracy of each magnetic field sensor. Calibration can be performed using labeled magnetic particles at known positions.

[0074] Step S3: Real-time magnetic field data at different locations is collected by magnetic field sensors. The data processing module pre-processes the real-time magnetic field data to filter out environmental noise and interference. Common signal processing techniques include filtering, denoising, and signal enhancement.

[0075] Next, based on the magnetic field strength detected by the magnetic field sensor, the three-dimensional position of the labeled magnetic particles is calculated through an inversion algorithm.

[0076] Specifically, step S4 comprises fusing the data of the magnetic field sensors processed in step S3, and the positioning algorithm module inverts the position information of the magnetic field source using the magnetic field change to determine the specific position of each labeled magnetic particle in the monitoring area;

[0077] The specific steps for inverting the location information of the magnetic field source using the magnetic field changes are:

[0078] Step S41, constructing a physical model, treating each labeled magnetic particle as a small magnetic dipole and defining it as x i , the magnetic field sensor is defined as y j , field sensor y j The magnetic field strength generated by the dipole can be measured, so the magnetic particle x is marked i The three-dimensional coordinate of i =(x i ,y i ,z i ), where i = 1, 2, ... n, magnetic field sensor y j The three-dimensional coordinate of j=(x j ,y j ,z j ), where j = 1, 2, ... n, n ≥ 3;

[0079] Step S42: constructing a magnetic dipole model. This step can establish an accurate spatial expression method and solve the problem of the difficulty in describing accurate displacement in the prior art. i The magnetic field at a distance r is expressed as It is described as:

[0080]

[0081] In the formula, To label magnetic particles x i With magnetic field sensor j The distance between

[0082] μ0 is the vacuum permeability,

[0083] m i To label magnetic particles x i The magnetic moment vector of

[0084] Step S43: construct a measurement equation to determine the magnetic field sensor y in the sensor array. j All labeled magnetic particles x measured i The sum of the magnetic fields:

[0085]

[0086] In the formula, B j For all labeled magnetic particles x i The sum of the magnetic fields, n ≥ 3;

[0087] This part is about the construction of the measurement equation. Based on the measurement equation, the error correction method can be used to correct the magnetic field sensor y j The measured labeled magnetic particles x i The magnetic field data is adjusted. Since the sensor may be interfered by the environmental magnetic field, it gets rid of the problem that the existing monitoring method is easily restricted by line of sight in principle, and improves the positioning accuracy.

[0088] Step S44, based on each labeled magnetic particle x i The magnetic field generated, and the magnetic field sensor y is also known j The spatial position of , the following relationship can be established:

[0089]

[0090] In the formula, The measurement value is obtained by solving the nonlinear equation obtained by neural network training, and its measurement value is expressed as a set of unknown coordinates x i Nonlinear equations;

[0091] It should be explained here that, under normal circumstances, if we want to solve the above nonlinear equations, i ,but m i 、y j The three quantities should be known. In reality, m i 、y j is a known constant. When using intelligent methods to solve environmental interference problems, the neural network method should be used to directly solve x i However, since the working conditions of sand in this application are relatively complex, too many interference possibilities will lead to low numerical accuracy of the solution, so we first solve way to improve the solution accuracy.

[0092] About Select Solver The applicant gave an explanation, combining the above nonlinear equation with the constructed measurement equation, the corresponding That is to say, for several labeled magnetic particles and several magnetic field sensors, the training data will contain different combinations of positions, magnetic moments and magnetic field sensors. j Contains the position information of each labeled magnetic particle, but in this formula x i It is the variable to be solved and cannot be directly obtained.

[0093] In solving When , the nonlinear equation is first trained through the neural network, and the training process is:

[0094] Step S441: Prepare training data set Among them, (B j ,m i ,y j ) is the input feature vector, is the target output, Represents all labeled magnetic particles x i The total magnetic field B j The predicted value of

[0095] For several labeled magnetic particles and several magnetic field sensors, the training data will contain different combinations of positions, magnetic moments, and magnetic field sensors. That is, it contains the position information of each labeled magnetic particle.

[0096] Step S442, define a feedforward neural network, set it to include an input layer, a hidden layer and an output layer, the input layer inputs features, the hidden layer can be one or more layers, and the output layer outputs the predicted value. Since the magnetic field is a three-dimensional vector, the size of the output layer is 3.

[0097] Forward propagation, input layer input features (x i ,m i ,y j ), and then transmitted to the hidden layer and transformed by a nonlinear activation function (preferably ReLU), and the output of the hidden layer is further transmitted to the output layer to obtain all the labeled magnetic particles x i The total magnetic field B j The predicted value of

[0098]

[0099] In the formula, f NN is a neural network model, θ is a trainable parameter in the neural network;

[0100] Step S443, error calculation, calculate the predicted value and the true value B j The error between them is expressed as follows:

[0101]

[0102] In the formula, L is the loss function, N is the maximum number n of j, and N ≥ 3;

[0103] Step S444, back propagation, uses the back propagation algorithm to calculate the gradient of the loss function L with respect to the trainable parameter θ, i.e.

[0104]

[0105] Step S445: weight update, update the weight of the neural network model by gradient descent or optimization algorithm, that is,

[0106]

[0107] In the formula, η is the learning rate;

[0108] Step S446, repeat the training of step S442-step S445 until the loss function converges to a preset standard minimum value or reaches a preset training round, completing the training process including the unknown coordinate x i Training of nonlinear equations;

[0109] Step S45, by solving the nonlinear equation, obtain the labeled magnetic particles x i With magnetic field sensorj The distance between The three-dimensional coordinates of the labeled magnetic particles are determined by iterative solution through triangulation of the distance between each labeled magnetic particle and the magnetic field sensor;

[0110] Step S451: for each magnetic field sensor y j In actual operation, the magnetic field strength vector B of several magnetic field sensors can be measured j , in order to calculate the value of each labeled magnetic particle x i The spatial coordinates of the magnetic particles need to be solved to determine the three-dimensional coordinates of the labeled magnetic particles. The nonlinear equation is:

[0111]

[0112] In the formula, To label magnetic particles x i With magnetic field sensor j The distance between them is obtained according to the constructed magnetic dipole model, that is,

[0113] In step S452, since this is a nonlinear equation system, the nonlinear least square method is used to solve the nonlinear equations in step S451, and the data of each magnetic field sensor is input into it, which can gradually approximate the precise three-dimensional coordinates of each labeled magnetic particle.

[0114] Finally, the data from the magnetic field sensor provides enough information to solve the problem of x for each labeled magnetic particle. i The spatial coordinates of .

[0115] Step S5, according to step S4, determines the specific position of each labeled magnetic particle in the monitoring area through multiple iterative calculations, that is, the three-dimensional coordinates of each labeled magnetic particle. Once the position of the labeled magnetic particle is determined, the system can track its movement in real time in the time series, record the changes in the position of each labeled magnetic particle, generate its three-dimensional motion trajectory, and finally output it.

[0116] Based on the movement trajectories of the labeled magnetic particles, it is possible to analyze the overall movement characteristics of the sand, such as flow velocity, scour depth variations, and soil stress. The software then visualizes the particle movement trajectories, allowing the operator to intuitively observe the dynamic changes within the sand. This long-term sand movement data is recorded and stored for further analysis, such as sand scour prediction and movement pattern recognition.

[0117] To complete the above monitoring, this application provides a sand movement monitoring system, including sand to be monitored, a number of labeled magnetic particles implanted in the sand, and a sensor array arranged around or above the area to be monitored, wherein the sensor array includes at least three magnetic field sensors, and the number of magnetic field sensors is the same as the number of labeled magnetic particles. The entire system, if displayed as a module, includes four modules: a magnetic field detection module, which arranges multiple magnetic field sensors to detect the movement of labeled magnetic particles in the sand. A positioning algorithm module, which uses a data fusion algorithm of multiple magnetic field sensors to calculate the motion trajectory of each labeled magnetic particle. A data processing module, which performs data preprocessing, filtering, and visualization of the motion trajectory. A sand movement analysis module, which analyzes long-term monitoring data and provides trends and dynamic characteristics of sand movement.

[0118] Preferably, Figure 2 As shown, this application uses three labeled magnetic particles, matched, such as Figure 1 As shown in FIG, a flow chart of the monitoring method is given using three labeled magnetic particles as an example. Figure 3 As shown, it is reflected in real time on the integrated interface of the smart sand movement monitoring system.

[0119] In summary, the sand movement monitoring method and sand movement monitoring system based on magnetic field positioning provided by this application are intended to solve: the limitations of traditional methods in sand movement monitoring, especially the problem of difficulty in accurately locating the movement trajectory of sand particles in complex environments; the existing technology is difficult to accurately capture real-time information on sand movement in dynamic and irregular sand environments, and cannot effectively monitor sand changes for a long time under high-precision conditions. It has the following advantages: not affected by light → magnetic field positioning technology does not rely on external light conditions and is suitable for all-weather monitoring; higher accuracy → magnetic field sensors can accurately detect tiny displacement changes, achieving higher-precision sand movement monitoring; adaptability to complex environments → the system can work in irregular and complex sand environments and has strong anti-interference capabilities; long-term stable monitoring → the system can maintain stable high-precision performance during long-term continuous monitoring, and is suitable for long-term dynamic monitoring tasks.

[0120] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by those skilled in the art to which this application belongs. It should also be understood that terms such as those defined in common dictionaries should be understood to have meanings consistent with their meanings in the context of the prior art and, unless defined as such herein, will not be interpreted in an idealized or overly formal sense.

[0121] The meaning of "and / or" in this application means that both situations where each exists alone or both exist at the same time are included.

[0122] The term “connection” as used in this application may mean a direct connection between components or an indirect connection between components via other components.

[0123] With the above-described preferred embodiments of the present invention as a guide, and with reference to the above description, relevant personnel are fully capable of making various changes and modifications without departing from the technical scope of this invention. The technical scope of this invention is not limited to the contents of the specification and must be determined according to the scope of the claims.

Claims

1. A method for monitoring sand movement based on magnetic field positioning, characterized by: The specific steps include: Step S1, selecting a number of labeled magnetic particles that meet the requirements and implanting the labeled magnetic particles into the sand in the monitoring area; Step S2, setting a sensor array consisting of a plurality of magnetic field sensors, calibrating the sensor array, and placing the calibrated sensor array around or above the monitoring area; Step S3, collecting real-time magnetic field data at different locations through a magnetic field sensor, and pre-processing the real-time magnetic field data by a data processing module to filter out environmental noise and interference; Step S4: fusing the data from the magnetic field sensors processed in step S3, and using the magnetic field variation to invert the position information of the magnetic field source, the positioning algorithm module determines the specific position of each labeled magnetic particle in the monitoring area; Step S5, tracking the specific position of the determined labeled magnetic particles in real time in the time series, recording the change of the position of the labeled magnetic particles, and generating a motion trajectory; Step S6: Analyze the movement of sand in the monitoring area based on the movement trajectory of the labeled magnetic particles generated in step S5.

2. The method for monitoring sand movement based on magnetic field positioning according to claim 1, characterized in that: Several magnetic field sensors are located at different heights and horizontal positions to form a sensor array.

3. The method for monitoring sand movement based on magnetic field positioning according to claim 1, characterized in that: In step S4, the specific steps of inverting and deriving the position information of the magnetic field source by using the magnetic field change are as follows: Step S41: construct a physical model, treat each labeled magnetic particle as a magnetic dipole, and define it as x i , the magnetic field sensor is defined as y j , then the magnetic particles are labeled x i The three-dimensional coordinate of i =(x i ,y i ,z i ), where i = 1, 2, ... n, magnetic field sensor y j The three-dimensional coordinate of j =(x j ,y j ,z j ), where j = 1, 2, ... n, n ≥ 3; Step S42, constructing a magnetic dipole model, marking magnetic particles x i The magnetic field at a distance r is expressed as It is described as: In the formula, To label magnetic particles x i With magnetic field sensor j The distance between μ0 is the vacuum permeability, m i To label magnetic particles x i The magnetic moment vector of Step S43: construct a measurement equation to determine the magnetic field sensor y in the sensor array. j All labeled magnetic particles x measured i The sum of the magnetic fields: In the formula, B j For all labeled magnetic particles x i The sum of the magnetic fields, n ≥ 3; Step S44, based on each labeled magnetic particle x i Magnetic field and magnetic field sensory j The spatial position of the , establish the following relationship: In the formula, The measurement value is obtained by solving the nonlinear equation obtained by neural network training, and its measurement value is expressed as a set of unknown coordinates x i Nonlinear equations; Step S45, by solving the nonlinear equation, obtain the labeled magnetic particles x i With magnetic field sensor j The distance between The three-dimensional coordinates of the labeled magnetic particles are determined by iterative solution through triangulation of the distance between each labeled magnetic particle and the magnetic field sensor.

4. The method for monitoring sand movement based on magnetic field positioning according to claim 3, characterized in that: The process of training the nonlinear equation by the neural network based on step S44 is as follows: Step S441: Prepare training data set Among them, (B j ,m i ,y j ) is the input feature vector, is the target output, Represents all labeled magnetic particles x i The total magnetic field B j The predicted value of Step S442: define a feedforward neural network, which includes an input layer, a hidden layer, and an output layer. The input layer inputs features (x i ,m i ,y j ), after being transferred to the hidden layer, it is transformed by the nonlinear activation function, and the output of the hidden layer is further transferred to the output layer to obtain all the labeled magnetic particles x i The total magnetic field B j The predicted value of In the formula, f NN is a neural network model, θ is a trainable parameter in the neural network; Step S443, calculate the predicted value and the true value B j The error between them is expressed as follows: In the formula, L is the loss function, N is the maximum number n of j, and N ≥ 3; Step S444, use the back propagation algorithm to calculate the gradient of the loss function L with respect to the trainable parameter θ, that is, Step S445, update the weights of the neural network model by gradient descent or optimization algorithm, i.e. In the formula, η is the learning rate; Step S446, repeat steps S442-S445 until the loss function converges to a preset standard minimum value or reaches a preset training round, completing the training process including the unknown coordinate x i Training of nonlinear equations.

5. The method for monitoring sand movement based on magnetic field positioning according to claim 4, characterized in that: In step S45, the specific process of solving the nonlinear equation is: Step S451: Determine the three-dimensional coordinates of the labeled magnetic particles using a nonlinear equation. Here, the nonlinear equation is: In the formula, To label magnetic particles x i With magnetic field sensor j The distance between them is obtained according to the constructed magnetic dipole model, that is, In step S452 , the nonlinear least square method is used to solve the nonlinear equation in step S451 , and the data of each magnetic field sensor is input into the solution to obtain the three-dimensional coordinates closest to each labeled magnetic particle.

6. The method for monitoring sand movement based on magnetic field positioning according to claim 3, characterized in that: Based on the measurement equation constructed in step S43, the error correction method is used to correct the magnetic field sensor y j The measured labeled magnetic particles x i The magnetic field data is adjusted.

7. The method for monitoring sand movement based on magnetic field positioning according to claim 1, characterized in that: In step S5, the specific position of each labeled magnetic particle in the monitoring area, ie, the three-dimensional coordinates of each labeled magnetic particle, is determined according to step S4, and the position change of each labeled magnetic particle is recorded to generate its motion trajectory.

8. The method for monitoring sand movement based on magnetic field positioning according to claim 1, characterized in that: In step S6, based on the movement trajectory of the marked magnetic particles, the overall movement characteristics of the sand are analyzed, including flow velocity, changes in scouring depth, and soil stress conditions.

9. A sand movement monitoring system, characterized by: A sand movement monitoring method based on magnetic field positioning as described in any one of claims 1-8 is adopted; the system includes sand to be monitored, a number of labeled magnetic particles are implanted in the sand, and a sensor array is arranged around or above the area to be monitored, wherein the sensor array includes at least three magnetic field sensors, and the number of magnetic field sensors is the same as the number of labeled magnetic particles.

Images