Intelligent metering method of sand excavation amount monitoring system

Abstract

The invention relates to the technical field of sand excavation metering, in particular to an intelligent metering method of a sand excavation amount monitoring system, which comprises the following steps of: S1, scanning and recording an operation area of a sand dredger by adopting a depth finder in cooperation with RTK GPS, and establishing a three-dimensional map before operation; s2, after sand excavation operation is finished, the operation area of the sand dredger is scanned again, the three-dimensional map after operation is obtained and compared with the three-dimensional map before operation, and topographic changes before and after sand excavation are calculated. When the method is used, high-precision gravel data can be stably obtained under different water flow conditions, sand excavation dispatching optimization and overload avoiding are facilitated through high-precision volume calculation, and the working efficiency of the sand dredger is improved. The method is beneficial to reducing sensor noise, improving measurement precision, realizing real-time monitoring of the volume of the sandstone in the cabin, accurately calculating the volume of the sandstone, improving the precision of supervision data, avoiding oversampling, improving the quality calculation precision, avoiding errors caused by fixed density calculation, automatically calculating in the whole process, avoiding manual intervention, and improving the measurement efficiency and stability.

Description

Technical Field

[0001] The invention relates to the technical field of sand mining metering, and in particular to an intelligent metering method for a sand mining quantity monitoring system. Background Art

[0002] The sand mining volume monitoring system is an intelligent supervision tool based on multi-sensor fusion, AI algorithm correction and blockchain evidence storage. It aims to accurately measure the operating volume of sand mining ships in real time, automatically identify violations, and build an enforcement evidence chain through data solidification technology to solve the pain points of traditional manual supervision such as low efficiency, large errors and difficulty in obtaining evidence.

[0003] The patent publication number is CN114235082A, and it states in the specification that "the present invention discloses an intelligent metering method based on a sand mining volume monitoring system, the system includes a sand mining volume monitoring terminal and a sand mining volume metering management device; the method utilizes the monitoring information fusion processing of three sensors, namely, a photoelectric sensor, a vibration sensor and an ultrasonic sensor, comprehensively determines the working state of the sand mining ship according to the vibration sensor and the photoelectric sensor, fuses the information of the photoelectric sensor and the ultrasonic sensor, calculates the sand mining volume by using the information processing method, and sends the sand mining volume information to the sand mining volume management device in real time; the sand mining volume management device receives the sand mining volume information, and uses a deep neural network algorithm in the sand mining volume metering management device to calculate the sand mining volume information. The sand mining volume is corrected, and then the over-mining warning is displayed on the website and stored in the database server. The present invention realizes real-time dynamic and accurate measurement of sand mining by sand mining ships, and transmits sand mining volume measurement information in real time, and can realize positioning and image information evidence collection and early warning for over-mining behavior. "Although the above technology realizes high-precision real-time measurement of sand mining volume through multi-sensor fusion and deep learning correction, it has the function of solving the problem of large errors (>30%) in traditional manual estimation, but ultrasonic ranging is affected by temperature. In extreme humidity (>95%RH) or low temperature (<-20℃) environments, the measurement error is expanded to 5% (2025 Heilongjiang Songhua River winter measurement data), and the sand and gravel density needs to be preset manually, which is not conducive to supervision.

[0004] In summary, developing an intelligent metering method for sand mining volume monitoring system is still a key issue that needs to be urgently solved in the field of sand mining metering technology. Summary of the invention

[0005] The purpose of the present invention is to solve the problem in the prior art that although the above-mentioned technology realizes high-precision real-time measurement of sand mining volume through multi-sensor fusion and deep learning correction, it has solved the problem of large error (>30%) of traditional manual estimation, but ultrasonic ranging is affected by temperature. In extreme humidity (>95%RH) or low temperature (<-20℃) environment, the measurement error is expanded to 5% (2025 Heilongjiang Songhua River winter measurement data), and the density of sand and gravel needs to be manually preset, which is not conducive to supervision.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The present invention provides an intelligent metering method for a sand mining volume monitoring system, including the following steps: S1. Use a depth sounder in cooperation with RTK GPS to scan and record the operation area of the sand dredger, and establish a three-dimensional map before operation;

[0008] S2. After the sand mining operation is completed, scan the operation area of the sand dredger again, compare the three-dimensional map after operation with the three-dimensional map before operation, calculate the topographic changes before and after sand mining, and obtain the excavation volume data;

[0009] S3. During the sand mining operation, obtain the three-dimensional point cloud data of the sand and gravel inside the cabin through the data of lidar, depth camera and inertial measurement unit;

[0010] S4. Use the SLAM algorithm to analyze the accumulation form of the sand and gravel according to the three-dimensional point cloud data, calculate the volume change, and obtain the volume data of the sand and gravel in the cabin;

[0011] S5. Use the density model of the sand and gravel to convert the excavation volume data and the volume data of the sand and gravel in the cabin into the excavation sand and gravel mass data and the cabin sand and gravel mass data respectively.

[0012] Further, in step S1, the method of using a depth sounder in cooperation with RTK GPS to scan and record the operation area of the sand dredger and establish a three-dimensional map before operation is as follows:

[0013] The depth sounder provides accurate data of the underwater terrain, and the terrain information of the scanned area by the depth sounder can be expressed as: where z(x, y, t) represents the water depth at any position and time, q is the propagation speed of sound waves in water (which can be dynamically corrected according to water temperature, salinity and depth changes), Δt is the time difference between the emission and reception of sound waves, and RTKGPS ensures the accuracy of geographical coordinates through its high-precision positioning system. The specific coordinates of the ship in the operation area can be expressed as: Q(x, y, z, t) = {Q x (t), Q y (t), Q z (t)}, where Q(x, y, z, t) represents the position of the ship in three-dimensional space, and {Q x (t), Q y (t), Q z (t)} represent the positions of the ship on the x, y, and z coordinate axes respectively, t is the real-time measurement time point, and the influence of water flow on the measurement error is corrected through the water flow error correction model. The influence formula of the water flow speed on the measurement value: where r′ is the corrected underwater depth, r is the original depth sounder data, Indicates the cumulative calculation of the impact of water flow from the start of measurement (t = 0) to the current time (t). is the water flow velocity at coordinates x, y and time t. is the slope change rate of the underwater terrain. (x, y, t) are three-dimensional coordinates and time variables, dt′ is a small time increment. By fusing RTKGPS positioning data and water depth data obtained by a sounding instrument, the position information of each data point (x, y, z) is modeled in three-dimensional space. The coordinates of the three-dimensional point cloud can be reconstructed by the following formula: where Y(x, y, z) is the total volume data of the three-dimensional map reconstruction, Δx i (t), Δy i (t), Δz i (t) is the dynamic change amount of each scan point in the direction. is the sum of the coordinate data of N measurement points. N is the number of scan points of the sounding instrument used, and T is the total duration of the scanning time.

[0014] Furthermore, in step S2, after the sand mining operation is completed, the operation area of the sand mining ship is scanned again. By comparing the three-dimensional map after the operation with the three-dimensional map before the operation, the method for calculating the terrain change before and after sand mining to obtain the excavation volume data is as follows:

[0015] Use the sounding instrument to obtain the underwater depth data again: U post (x, y) = f ES (x, y, t), where U post (x, y) is the underwater height data after the operation, f ES (·) represents the measurement function of the sounding instrument, (x, y) are the coordinates of the sampling points, and t is the time. The volume change is obtained by discrete integration: where I excavated is the excavation volume, O cell is the area of each sampling grid unit, N′, M are the number of rows and columns of the terrain grid, ΔU(x i , y j ) is the height change amount. is a double summation operation. An error correction method is introduced, and artificial intelligence, deep learning and error correction algorithms are used to correct the volume calculation deviation. The main sources of errors are: measurement error, sedimentation error and equipment drift error. Let the actual excavation volume be I actual , the measured volume be I measured , then the error model is: ∈ I = I measured - I actual , if ∈ I > 0, it means that the measurement overestimates the excavation amount. If ∈ I < 0, it means that the measurement underestimates the excavation amount.

[0016] Further, in step S2, after the sand mining operation is completed, the operation area of the sand dredger is scanned again, and the three-dimensional map after the operation is compared with the three-dimensional map before the operation to calculate the topographic changes before and after sand mining. The method for obtaining the excavation volume data is as follows:

[0017] Use the deep learning neural network β PA (·) for error prediction:

[0018] Where is the water flow velocity, χ is the sediment settlement time, is the predicted error correction value, and finally the corrected excavation volume data I final is obtained, and the error range is provided: Where δ I is the standard deviation of the correction error, and the measurement result is optimized by Kalman filtering: Where D t is the Kalman gain to control the weight distribution between the current measurement and the historical measurement, is the final volume value after filtering, is the corrected volume value at the previous moment, so as to automatically correct the changes caused by sediment movement and water flow and ensure the accuracy of the volume calculation result.

[0019] Further, in step S3, during the sand mining operation, the method for obtaining the three-dimensional point cloud data of the sand and gravel inside the cabin through the data of the lidar, depth camera and inertial measurement unit is as follows:

[0020] Synchronize the data collected by the lidar, depth camera and IMU sensors in time, and their measured values are F Lidar (t), F Camera (t), F IMU (t), and the state model is estimated: Where G t is the state vector of the state model, H t is the position vector, J t is the velocity vector, K t is the acceleration vector, and the state at the next moment is predicted: Where L t is the state transition matrix, Z t is the control matrix, c t is the control input vector, v′ t is the process noise, and the state of the state model is updated according to the sensor measurement value F t :

[0021] Further, in step S3, during the sand mining operation, the method for obtaining the three-dimensional point cloud data of the sand and gravel inside the cabin through the lidar, depth camera, and inertial measurement unit data is as follows:

[0022] Where Q' t is the Kalman gain representing the weighted ratio between prediction and observation, and the calculation formula is: Where is the predicted error covariance matrix, W t is the measurement matrix, E t is the observation noise covariance matrix. Based on the Kalman filter, the extended Kalman filter is used to process the nonlinear system, and the state prediction formula: Q' t = f(Q' t-1 , r t '), where f(·) is the nonlinear state transition function, and the linearization formula: The state transition function is linearized through Taylor expansion, and the update step formula: Where Y'(·) is the measurement function.

[0023] Further, in step S4, the method for calculating the volume change and obtaining the volume data of the sand and gravel in the cabin by analyzing the accumulation shape of the sand and gravel according to the three-dimensional point cloud data using the SLAM algorithm is as follows:

[0024] Introduce a real-time error correction mechanism to optimize the pose estimation by combining the visual inertial odometer with the camera and IMU data. The IMU prediction formula is: Where U' t is the spatial position at the current moment, O' t-1 is the velocity at the previous moment, p' t-1 is the acceleration at the previous moment, Δt is the time interval, and the camera update formula: T″ t = T″ t-1 + ΔT″ t , where T″ t is the pose at the current moment, T″ t-1 represents the pose at the previous moment, and ΔT″ t is the pose increment calculated through the visual odometer.

[0025] Further, in step S4, the method for calculating the volume change and obtaining the volume data of the sand and gravel in the cabin by analyzing the accumulation shape of the sand and gravel according to the three-dimensional point cloud data using the SLAM algorithm is as follows:

[0026] Optimize the position and map in the SLAM through the graph optimization algorithm: Where s i ′ j (a i ′, a' j) is the error function between nodes, where i and j represent indices, ||·|| represents the Euclidean norm, and g′ ij is the observed data, and Σ ij is the covariance of the observations. Based on the optimized 3D map and point cloud data, the gravel volume formula is obtained by calculating the 3D grid volume: H′ = ∫ H′ f(x, y, z)·d′H′, where H′ is the region for which the volume is sought, and f(x, y, z) represents the density function of the point cloud in 3D space.

[0027] Furthermore, in step S5, the method of using the density model of gravel to convert the excavation volume data and the gravel volume data in the cabin into the excavation gravel mass data and the gravel mass data in the cabin respectively is as follows:

[0028] Let the total gravel volume in the excavation area be The gravel volume inside the cabin is The density of gravel is affected by various factors, including but not limited to humidity, temperature, and the composition of sediment particles. A dynamic density model: ε(x, t) is introduced, where x represents the position and t represents the time, to simulate the change of density under different environmental conditions. The dynamic density model can be expressed as: where ε0 is the reference density, φ is the sensitivity coefficient of temperature to density change, is the sensitivity coefficient of humidity to density change, is the temperature at the current position x and time t, is the humidity at the current position x and time t, is the standard temperature and humidity.

[0029] Furthermore, in step S5, the method of using the density model of gravel to convert the excavation volume data and the gravel volume data in the cabin into the excavation gravel mass data and the gravel mass data in the cabin respectively is as follows:

[0030] Obtain the actual measurement data through sensors and Substitute them into the dynamic density model to obtain the density change situation in different regions, and then adjust the density value. Finally, the calculation formulas for the excavation gravel mass and the gravel mass in the cabin can be expressed as: where ε0 is the reference density, φ is the influence factor of temperature on density change, is the influence factor of humidity on density change,

[0031] is the temperature at position x and time t, is the humidity at position x and time t, is the standard temperature and humidity, respectively represent the volume of sand and gravel in the excavation area and the cabin interior, respectively represent the mass of sand and gravel in the excavation area and the cabin interior. In the calculation formula of the mass of excavated sand and gravel and the mass of cabin sand and gravel in the calculation formula, the volume data is calculated by integration, and the density is adjusted according to the location and environmental conditions.

[0032] Advantageous effects

[0033] Adopting the technical solution provided by the present invention, compared with the known public technologies, it has the following

[0034] advantageous effects:

[0035] When the present invention is in use, it is beneficial to stably obtain high-precision sand and gravel data under different water flow conditions. Through high-precision volume calculation, it is beneficial to optimize sand mining scheduling, avoid overloading, reduce sensor noise, improve measurement accuracy, realize real-time monitoring of the volume of cabin sand and gravel, accurately calculate the volume of sand and gravel, improve the accuracy of supervision data, avoid over-exploitation, improve the accuracy of quality calculation, avoid errors caused by fixed density calculation, and perform full-process automatic calculation without manual intervention, improving measurement efficiency and stability. Description of the drawings

[0036] Figure 1 is a flowchart of an intelligent metering method for a sand mining volume monitoring system of the present invention. Detailed implementation manners

[0037] In order to enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0038] It should be noted that the terms "first", "second", etc. in the description, claims and above-mentioned drawings of the present invention are used to distinguish similar objects, and do not necessarily have to be used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present invention described here can be implemented in an order other than those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but includes other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0039] The present invention will be further described in detail below with reference to the accompanying drawings:

[0040] Embodiment:

[0041] As Figure 1 shown, the present invention provides an intelligent metering method for a sand mining volume monitoring system, including the following steps: S1. Using a depth sounder in cooperation with RTK GPS to scan and record the operation area of the sand mining vessel, and establishing a three-dimensional map before operation;

[0042] Further, in step S1, the method of using a depth sounder in cooperation with RTK GPS to scan and record the operation area of the sand mining vessel and establishing a three-dimensional map before operation is as follows:

[0043] The depth sounder provides accurate data on the underwater terrain. The terrain information of the area scanned by the depth sounder can be expressed as: where z(x, y, t) represents the water depth at any position and time, q is the propagation speed of sound waves in water (which can be dynamically corrected according to changes in water temperature, salinity and depth), Δt is the time difference between the emission and reception of sound waves, and RTK GPS ensures the accuracy of geographical coordinates through its high-precision positioning system. The specific coordinates of the vessel in the operation area can be expressed as: Q(x, y, z, t) = {Q x (t), Q y (t), Q z (t)}, where Q(x, y, z, t) represents the position of the vessel in three-dimensional space, and {Q x (t), Q y (t), Q z (t)} respectively represent the positions of the vessel on the x, y, and z coordinate axes, t is the time point of real-time measurement, and the influence of water flow on the measurement error is corrected through a water flow error correction model. The influence formula of the water flow velocity on the measurement value: where r′ is the corrected underwater depth and r is the original depth sounder data, It represents the cumulative calculation of the influence of water flow from the start of measurement (t = 0) to the current moment (t). is the water flow velocity at coordinates x, y and time t, is the slope change rate of the underwater terrain, (x, y, t) are three-dimensional coordinates and time variables, dt′ is a small time increment. Through the fusion of RTKGPS positioning data and water depth data obtained by a depth sounder, the position information of each data point (x, y, z) is modeled in three-dimensional space. The coordinates of the three-dimensional point cloud can be reconstructed by the following formula: where Y(x, y, z) is the total volume data of the three-dimensional map reconstruction, Δx i (t), Δy i (t), Δz i (t) is the dynamic change amount of each scanning point in the direction, is the summation of the coordinate data of N measurement points. N is the number of scanning points of the depth sounder used, and T is the total duration of the scanning time.

[0044] In this embodiment, a depth sounder and RTKGPS are installed on a sand dredging ship, a measurement route is set, the sand dredging ship sails slowly along the established route to ensure complete data coverage, water depth data is collected, the sound velocity is corrected in real time according to the changes in water temperature and salinity, RTKGPS data is collected, the water depth data and geographical coordinates are fused, a water flow error correction model is adopted to eliminate the interference of water flow on the sounding data, and a three-dimensional point cloud model before operation is generated to provide reference data for subsequent sand dredging volume calculation. Combining the depth sounder and RTKGPS is conducive to achieving centimeter-level accuracy in underwater terrain measurement, and through water flow error correction, measurement deviation is eliminated. By adopting a dynamic sound velocity correction model and a water flow compensation algorithm, the system can provide stable and reliable measurement results in different hydrological environments.

[0045] S2. After the sand dredging operation is completed, scan the operation area of the sand dredging ship again, compare the three-dimensional map after operation with the three-dimensional map before operation, calculate the terrain change before and after sand dredging, and obtain the excavation volume data;

[0046] Furthermore, in step S2, after the sand dredging operation is completed, scan the operation area of the sand dredging ship again, compare the three-dimensional map after operation with the three-dimensional map before operation, calculate the terrain change before and after sand dredging, and the method for obtaining the excavation volume data is as follows:

[0047] Use the depth sounder to obtain the underwater depth data again: U post (x, y) = f ES (x, y, t), where U post (x, y) is the underwater height data after operation, f ES (·) represents the measurement function of the depth sounder, (x, y) are the coordinates of the sampling points, t is the time, and the volume change is obtained by discrete integration: where I excavated is the excavation volume, O cell is the area of each sampling grid cell, N′, M are the number of rows and columns of the terrain grid, ΔU(x i , y j ) is the height change amount, is a double summation operation. An error correction method is introduced, and artificial intelligence, deep learning, and error correction algorithms are used to correct the volume calculation deviation. The main sources of error are: measurement error, sedimentation error, and equipment drift error. Let the actual excavation volume be I actual , and the measured volume be I measured . Then the error model is: ∈ I = I measured - I actual . If ∈ I > 0, it means the measurement overestimates the excavation volume. If ∈ I < 0, it means the measurement underestimates the excavation volume.

[0048] Furthermore, in step S2, after the sand mining operation is completed, the operation area of the sand mining ship is scanned again. By comparing the three-dimensional map after the operation with the three-dimensional map before the operation, the method for calculating the terrain change before and after sand mining to obtain the excavation volume data is as follows:

[0049] Use the deep learning neural network β PA (·) to perform error prediction:

[0050] where is the water flow velocity, χ is the sediment settlement time, is the predicted error correction value, and finally the corrected excavation volume data I final is obtained, and the error range is provided: where δ I is the standard deviation of the correction error, and the Kalman filter is used to optimize the measurement result: where D t is the Kalman gain to control the weight distribution between the current measurement and the historical measurement, is the final volume value after filtering, is the corrected volume value at the previous moment, thereby automatically correcting the changes caused by sediment movement and water flow to ensure the accuracy of the volume calculation result.

[0051] In this embodiment, a bathymeter and RTK GPS are used to collect the underwater terrain data after the operation, calculate the terrain changes before and after sand mining, calculate the volume difference using the discrete integral method, introduce an error correction model, use deep learning and Kalman filtering to eliminate the water flow deposition error, finally output the corrected excavation volume data, and give the error range, which is beneficial to reducing the volume measurement error, improving the accuracy, automatically adjusting the error correction parameters in combination with artificial intelligence, reducing manual intervention, and improving the calculation efficiency.

[0052] S3. During the sand mining operation, three-dimensional point cloud data of the sand and gravel inside the cabin is obtained through lidar, depth camera, and inertial measurement unit data;

[0053] Further, in step S3, the method for obtaining the three-dimensional point cloud data of the sand and gravel inside the cabin through lidar, depth camera, and inertial measurement unit data during the sand mining operation is as follows:

[0054] Synchronize the data collected by the lidar, depth camera, and IMU sensors in time, and their measured values are F Lidar (t), F Camera (t), F IMU (t). State model estimation: where G t is the state vector of the state model, H t is the position vector, J t is the velocity vector, K t is the acceleration vector. Predict the state at the next moment: where L t is the state transition matrix, Z t is the control matrix, c t is the control input vector, v′ t is the process noise. Update the state of the state model according to the sensor measurement value F t :

[0055] Further, in step S3, the method for obtaining the three-dimensional point cloud data of the sand and gravel inside the cabin through lidar, depth camera, and inertial measurement unit data during the sand mining operation is as follows:

[0056] where Q′ t is the Kalman gain representing the weighted ratio between prediction and observation, and the calculation formula is: where is the predicted error covariance matrix, W t is the measurement matrix, E t is the observation noise covariance matrix. On the basis of Kalman filtering, extended Kalman filtering is used to process the nonlinear system, and the state prediction formula: Q′ t = f(Q′t-1 , r t ′), where f(·) is a non - linear state transition function, and the linearization formula: Linearize the state transition function through Taylor expansion, and the update step formula: where Y′(·) is the measurement function.

[0057] In this embodiment, a lidar, a depth camera, and an IMU sensor are used for data acquisition to obtain the point cloud data inside the cabin. Data time synchronization is performed to ensure the consistency of data fusion. Kalman filtering is used to fuse sensor data to improve the robustness of the data. Extended Kalman filtering is utilized for non - linear error correction to enhance the measurement accuracy, generating a high - precision three - dimensional point cloud map of the sand and gravel in the cabin, calculating the volume of the sand and gravel, which is beneficial to stably obtain high - precision sand and gravel data under different water flow conditions. Through high - precision volume calculation, it is beneficial to optimize the sand mining scheduling and avoid overloading or under - loading.

[0058] S4. Adopt the SLAM algorithm to analyze the accumulation form of the sand and gravel based on the three - dimensional point cloud data, calculate the volume change, and obtain the volume data of the sand and gravel in the cabin;

[0059] Furthermore, in step S4, the method of adopting the SLAM algorithm to analyze the accumulation form of the sand and gravel based on the three - dimensional point cloud data, calculate the volume change, and obtain the volume data of the sand and gravel in the cabin is as follows:

[0060] Introduce a real - time error correction mechanism to optimize the pose estimation by combining the visual inertial odometer with camera and IMU data. The IMU prediction formula is: where U′ t is the spatial position at the current moment, O′ t-1 is the velocity at the previous moment, p′ t-1 is the acceleration at the previous moment, Δt is the time interval, and the camera update formula: T″ t = T″ t-1 + ΔT″ t , where T″ t is the pose at the current moment, T″ t-1 represents the pose at the previous moment, and ΔT″ t is the pose increment calculated through the visual odometer.

[0061] Furthermore, in step S4, the method of adopting the SLAM algorithm to analyze the accumulation form of the sand and gravel based on the three - dimensional point cloud data, calculate the volume change, and obtain the volume data of the sand and gravel in the cabin is as follows:

[0062] Optimize the position and map in SLAM through the graph optimization algorithm: where s′ ij (a′ i , a′j ) represents the error function between nodes, where i and j represent indices, ||·|| represents the Euclidean norm, and g′ ij is the observed data, and ∑ ij is the covariance of the observations. Based on the optimized 3D map and point cloud data, the gravel volume formula is obtained by calculating the 3D grid volume: H′ = ∫ H′ f(x, y, z)·d′H′, where H′ is the region for volume calculation, and f(x, y, z) represents the density function of the point cloud in 3D space.

[0063] In this embodiment, a lidar and a depth camera are installed in the cabin to obtain the 3D point cloud data of the gravel. The IMU sensor is used to provide the hull attitude information to correct the measurement error. The visual inertial odometer is used for pose estimation to improve the SLAM accuracy. The graph optimization algorithm is combined to reduce the error and optimize the 3D map. Through point cloud gridding and volume integration calculation, the volume of the gravel in the cabin is accurately calculated. The Kalman filter is used to correct the volume data in real time to reduce the influence of noise. The IMU+camera SLAM fusion algorithm is used, which is beneficial to reducing sensor noise and improving the measurement accuracy, realizing the real-time monitoring of the gravel volume in the cabin, accurately calculating the gravel volume, and avoiding over-excavation.

[0064] S5. Using the density model of the gravel, convert the excavation volume data and the gravel volume data in the cabin into the excavation gravel mass data and the gravel mass data in the cabin respectively;

[0065] Further, in step S5, the method of using the density model of the gravel to convert the excavation volume data and the gravel volume data in the cabin into the excavation gravel mass data and the gravel mass data in the cabin respectively is as follows:

[0066] Let the total gravel volume in the excavation area be The gravel volume inside the cabin is The density of the gravel is affected by various factors, including but not limited to humidity, temperature, and the composition of sediment particles. A dynamic density model: ε(x, t) is introduced, where x represents the position and t represents the time, to simulate the change of density under different environmental conditions. The dynamic density model can be expressed as: where ε0 is the reference density, φ is the sensitivity coefficient of temperature to density change, is the sensitivity coefficient of humidity to density change, is the temperature at the current position x and time t, is the humidity at the current position x and time t, is the standard temperature and humidity.

[0067] Further, in step S5, the method of using the density model of sand and gravel to convert the excavation volume data and the sand and gravel volume data in the cabin into the excavation sand and gravel mass data and the sand and gravel mass data in the cabin is as follows:

[0068] Obtain actual measurement data through sensors and Substitute into the dynamic density model to obtain the density change situation in different regions, and then adjust the density value. Finally, the calculation formulas for the excavation sand and gravel mass and the sand and gravel mass in the cabin can be expressed as: where ε0 is the reference density, φ is the influence factor of temperature on density change, is the influence factor of humidity on density change, is the temperature at position x and time t, is the humidity at position x and time t, is the standard temperature and humidity, respectively represent the sand and gravel volumes in the excavation area and the cabin interior, respectively represent the sand and gravel masses in the excavation area and the cabin interior. In the calculation formulas for the excavation sand and gravel mass and the sand and gravel mass in the cabin , the volume data is calculated by integration, and the density is adjusted according to the position and environmental conditions.

[0069] In this embodiment, RTK-GPS, temperature and humidity sensors, and depth sounders are arranged inside the cabin and in the sand excavation area for environmental data collection and volume calculation. The temperature and humidity coefficients are calibrated using historical data, the sand and gravel density is calculated in real time, the calculation formula is adjusted, the sand and gravel volume is obtained through SLAM point cloud mapping, the mass data is calculated using the dynamic density model and updated in real time. Based on the dynamic density model, the density value is adjusted in combination with the temperature and humidity data, which is beneficial to improving the mass calculation accuracy and avoiding the errors caused by fixed density calculation. The whole process is automatically calculated without manual intervention, improving the measurement efficiency and stability.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements will not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent metering method for a sand mining monitoring system, characterized in that: The following steps are involved: S1. Use a depth sounder in conjunction with RTKGPS to scan and record the sand mining ship's operating area, and create a three-dimensional map before the operation; S2. After the sand mining operation is completed, the sand mining ship operation area is scanned again, and the three-dimensional map after the operation is obtained is compared with the three-dimensional map before the operation, and the terrain changes before and after the sand mining are calculated to obtain the excavation volume data; S3. During the sand mining operation, the three-dimensional point cloud data of the sand and gravel inside the cabin is obtained through the laser radar, depth camera and inertial measurement unit data; S4, using SLAM algorithm to analyze the accumulation form of sand and gravel according to the three-dimensional point cloud data, calculate the volume change, and obtain the volume data of sand and gravel in the cabin; S5. Using the density model of sand and gravel, the excavation volume data and the cabin sand and gravel volume data are converted into excavation sand and gravel mass data and cabin sand and gravel mass data respectively.

2. The intelligent metering method of a sand mining monitoring system according to claim 1 is characterized in that: In step S1, the method of using a depth sounder in conjunction with RTK GPS to scan and record the sand mining ship operation area and establish a three-dimensional map before the operation is as follows: The depth sounder provides accurate data of the underwater topography. The topographic information of the depth sounder scanning area can be expressed as: Where z(x, y, t) represents the water depth at any position and time, q is the propagation speed of sound waves in water (which can be dynamically corrected according to changes in water temperature, salinity and depth), Δt is the time difference from the emission to the reception of the sound wave, and RTKGPS ensures the accuracy of geographic coordinates through its high-precision positioning system. The specific coordinates of the ship in the operating area can be expressed as: Q(x, y, z, t) = {Q x (t), Q y (t), Q z (t)}, where Q(x, y, z, t) represents the position of the ship in three-dimensional space, {Q x (t), Q y (t), Q z (t)} represent the position of the ship on the x, y, and z axes respectively, t is the time point of real-time measurement, and the influence of water flow on the measurement error is corrected by the water flow error correction model. The influence formula on the measured value is: Where r′ is the corrected bottom depth, r is the original bathymetric data, Indicates the cumulative calculation of the impact on water flow from the start of measurement (t=0) to the current time (t), is the velocity of the water at coordinates x, y and time t, is the slope change rate of the underwater terrain, (x, y, t) is the three-dimensional coordinate and time variable, dt′ is the small time increment, and the position information of each data point (x, y, z) is fused with the water depth data obtained by the echo sounder to perform three-dimensional spatial modeling. The coordinates of the three-dimensional point cloud can be reconstructed by the following formula: Where Y(x, y, z) is the total volume data of the 3D map reconstruction, Δx i (t), Δy i (t), Δz i (t) is the dynamic change of each scanning point in direction, The coordinate data of N measurement points are summed, where N is the number of echo sounder scanning points used and T is the total scanning time.

3. The intelligent metering method of a sand mining amount monitoring system according to claim 2 is characterized in that: In step S2, after the sand mining operation is completed, the sand mining ship operation area is scanned again to obtain a three-dimensional map after the operation and compare it with the three-dimensional map before the operation, and the terrain changes before and after the sand mining are calculated. The method for obtaining the excavation volume data is as follows: Use the depth sounder to obtain underwater depth data again: post (x, y) = f ES (x, y, t), where U post (x, y) is the water bottom height data after operation, f ES (·) represents the measurement function of the depth sounder, (x, y) is the coordinate of the sampling point, t is the time, and the volume change is calculated by discrete integration: Among them I excavated is the excavation volume, O cell is the area of ​​each sampling grid unit, N′, M is the number of rows and columns of the terrain grid, ΔU(x i ,y j ) is the height change, It is a double summation operation. The error correction method is introduced. Artificial intelligence, deep learning and error correction algorithm are used to correct the volume calculation deviation. The main sources of error are: measurement error, deposition error and equipment drift error. Assume that the actual excavation volume I actual , the measured volume I measured , then the error model is: ∈ I =I measured -I actual , if ∈ I > 0 means that the measurement overestimates the amount of mining. I <0 means the measurement underestimates the excavation volume.

4. The intelligent metering method of a sand mining monitoring system according to claim 3 is characterized in that: In step S2, after the sand mining operation is completed, the sand mining ship operation area is scanned again to obtain a three-dimensional map after the operation and compare it with the three-dimensional map before the operation, and the terrain changes before and after the sand mining are calculated. The method for obtaining the excavation volume data is as follows: Using deep learning neural network β PA (·) Make error prediction: in is the water velocity, χ is the sediment settling time, is the predicted error correction value, and finally the corrected excavation volume data I is obtained final , and provide a margin of error: where δ I is the standard deviation of the correction error, and the Kalman filter is used to optimize the measurement results: Where D t The Kalman gain controls the weight distribution between current measurement and historical measurement. is the final volume value after filtering, It is the corrected volume value of the previous moment, thereby automatically correcting the changes caused by sediment movement and water flow to ensure the accuracy of the volume calculation results.

5. The intelligent metering method of the sand mining monitoring system according to claim 4 is characterized in that: In step S3, during the sand mining operation, the method for obtaining the three-dimensional point cloud data of the sand and gravel inside the cabin through the laser radar, depth camera and inertial measurement unit data is as follows: The data collected by the laser radar, depth camera, and IMU sensor are synchronized in time, and their measurement values ​​are F Lidar (t), F Camera (t), F IMU (t), the state model estimates: Among them G t is the state vector of the state model, H t is the position vector, J t is the velocity vector, K t is the acceleration vector, predicting the state at the next moment: Where L t is the state transfer matrix, Z t is the control matrix, c t is the control input vector, v′ t is the process noise, based on the sensor measurement value F t Update the state model state:

6. The intelligent metering method of the sand mining monitoring system according to claim 5 is characterized in that: In step S3, during the sand mining operation, the method for obtaining the three-dimensional point cloud data of the sand and gravel inside the cabin through the laser radar, depth camera and inertial measurement unit data is as follows: where Q′ t The Kalman gain represents the weighted ratio between prediction and observation, and is calculated as: in is the predicted error covariance matrix, W t is the measurement matrix, E t is the observation noise covariance matrix. Based on Kalman filtering, the extended Kalman filter is used to process nonlinear systems. The state prediction formula is: Q′ t =f(Q′ t-1 ,r t ′), where f(·) is the nonlinear state transfer function, and the linearization formula is: The state transfer function is linearized by Taylor expansion, and the update step formula is: where Y′(·) is the measurement function.

7. The intelligent metering method of the sand mining monitoring system according to claim 6 is characterized in that: In step S4, the method of using the SLAM algorithm to analyze the accumulation form of sand and gravel according to the three-dimensional point cloud data and calculate the volume change to obtain the volume data of sand and gravel in the cabin is: A real-time error correction mechanism is introduced to optimize the pose estimation by combining the visual inertial odometer with the camera and IMU data. The IMU prediction formula is: where U′ t is the spatial position at the current moment, O′ t-1 is the velocity at the previous moment, p′ t-1 is the acceleration at the previous moment, Δt is the time interval, and the camera update formula is: T″ t =T″ t-1 +ΔT″ t , where T″ t is the current position, T″ t-1 Indicates the posture at the previous moment, ΔT″ t is the pose increment calculated by the visual odometry.

8. The intelligent metering method of the sand mining monitoring system according to claim 7 is characterized in that: In step S4, the method of using the SLAM algorithm to analyze the accumulation form of sand and gravel according to the three-dimensional point cloud data and calculate the volume change to obtain the volume data of sand and gravel in the cabin is: Optimize the position and map in SLAM through graph optimization algorithm: where s i ' j (a i ′,a′ j ) represents the error function between nodes, i, j represent indexes, ||·|| represents the Euclidean norm, and g′ ij is the observed data, ∑ ij is the observed covariance. Based on the optimized 3D map and point cloud data, the sand and gravel volume formula is obtained by calculating the 3D grid volume: H′=∫ H′ f(x, y, z)·d′H′, where H′ is the area whose volume is to be found, and f(x, y, z) represents the density function of the point cloud in three-dimensional space.

9. The intelligent metering method of the sand mining monitoring system according to claim 8 is characterized in that: In step S5, the method of converting the excavation volume data and the cabin sand and gravel volume data into the excavation sand and gravel mass data and the cabin sand and gravel mass data respectively by using the density model of the sand and gravel is as follows: Assume the total gravel volume in the excavation area is The volume of sand and gravel inside the cabin is The density of sand and gravel is affected by many factors, including but not limited to humidity, temperature, and the composition of sediment particles. The dynamic density model is introduced: ε(x, t), where x represents the position and t represents the time, to simulate the change of density under different environmental conditions. The dynamic density model can be expressed as: Where ε0 is the reference density, φ is the sensitivity coefficient of temperature to density change, is the sensitivity coefficient of humidity to density change, is the temperature at the current position x and time t, is the humidity at the current position x and time t, It is standard temperature and humidity.

10. The intelligent metering method of the sand mining monitoring system according to claim 8 is characterized in that: In step S5, the method of converting the excavation volume data and the cabin sand and gravel volume data into the excavation sand and gravel mass data and the cabin sand and gravel mass data respectively by using the density model of the sand and gravel is as follows: Get actual measurement data through sensors and Substitute it into the dynamic density model to obtain the density changes in different areas, and then adjust the density value. Finally, the quality of excavated sand and gravel and the quality of sand and gravel in the cabin The calculation formula can be expressed as: Where ε0 is the reference density, φ is the influence factor of temperature on density change, is the factor affecting density change due to humidity, is the temperature at position x and time t, is the humidity at location x and time t, It is the standard temperature and humidity. Respectively represent the volume of sand and gravel in the excavation area and inside the cabin, Respectively represent the quality of sand and gravel in the excavation area and inside the cabin. and the quality of sand and gravel in the cabin In the calculation formula, volume data is calculated by integration and density is adjusted according to location and environmental conditions.

Images