A riverbed quality automatic collection control system and method based on Beidou navigation

Abstract

The present invention relates to the field of automated data collection technology, and in particular to a Beidou navigation-based automatic riverbed quality data collection control system and method. The method comprises the following steps: acquiring a Beidou satellite multidimensional data set; preprocessing the raw data of the Beidou satellite multidimensional data set to obtain a Beidou multidimensional standard data set; fusing multi-source data based on the Beidou multidimensional standard data set to obtain riverbed multi-source fused data; and generating riverbed sampling decisions using the riverbed multi-source fused data. Therefore, the present invention solves the problems of incomplete data, unreasonable sampling paths, and inefficient report generation in traditional riverbed data sampling processes through the comprehensive application of multi-source data fusion, path optimization, and automated data evaluation, thereby improving the accuracy, efficiency, and intelligence level of riverbed data collection.

Description

Technical Field

[0001] The present invention relates to the field of automated data collection technology, and in particular to a Beidou navigation-based riverbed data automatic data collection control system and method. Background Art

[0002] Traditional riverbed quality collection methods rely heavily on manual operation, which is subject to significant operational errors and low sampling efficiency. Manual sampling is not only affected by operator experience but also fails to accurately capture all types of riverbed data in real time. Sampling location and depth control accuracy is low, and sampling quality is difficult to guarantee. Furthermore, existing collection systems often lack highly integrated monitoring and control mechanisms, resulting in inaccurate fusion of multi-source riverbed information. This makes it difficult to meet the timeliness and accuracy requirements of high-precision data collection. Many traditional riverbed quality collection systems rely on fixed-position sensors and simple manual controls, which are unable to rapidly respond to dynamic environments, especially in complex or rapidly changing waters. This approach lacks flexible sampling path planning for uneven riverbed depths and complex geological layers, leading to data limitations and bias. Furthermore, while some high-precision sensors have been applied to riverbed monitoring, the lack of effective data fusion and intelligent judgment mechanisms often makes it difficult to quickly and accurately analyze and determine the geological characteristics of different riverbed layers under high real-time requirements, which in turn affects subsequent sampling decisions and operations. Furthermore, existing technologies also lack sufficient analysis of sensor data synchronization and sampling efficiency during the collection process. Many systems simply record sensor data when transmitting data back, and fail to conduct in-depth performance analysis and evaluation. This results in the inability to timely discover bottlenecks or deficiencies in system operation and the inability to accurately adjust collection strategies to cope with complex sampling environments. Summary of the Invention

[0003] Based on this, it is necessary to provide a riverbed quality automatic collection control system and method based on Beidou navigation to solve at least one of the above technical problems.

[0004] To achieve the above purpose, a method for automatically collecting and controlling riverbed quality based on Beidou navigation is provided, the method comprising the following steps:

[0005] Step S1: obtaining a Beidou satellite multidimensional dataset; performing raw data preprocessing on the Beidou satellite multidimensional dataset to obtain a Beidou multidimensional standard dataset;

[0006] Step S2: Perform multi-source data fusion based on the Beidou multi-dimensional standard data set to obtain riverbed multi-source fusion data; use the riverbed multi-source fusion data to generate riverbed sampling decisions;

[0007] Step S3: Using the riverbed sampling decision to plan the sampling path, obtain riverbed path planning data; based on the riverbed path planning data, generate a collection instruction, and use the riverbed robotic arm to perform real-time judgment and execution, and finally obtain a riverbed collection operation log;

[0008] Step S4: using the riverbed collection operation log to transmit data packets back to obtain the riverbed collection return log; performing performance index evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; constructing a riverbed quality automatic collection report based on the riverbed collection evaluation data.

[0009] The beneficial effect of the present invention is that, by acquiring a multidimensional dataset from Beidou satellites, the diversity and real-time nature of data sources are ensured, thereby providing high-precision spatial information for subsequent riverbed monitoring. During the data processing phase, standardization and preprocessing of the raw data eliminate noise and outliers in the data, thereby improving the integrity and accuracy of the data. The multi-source data fusion module further integrates Beidou satellite spatiotemporal information, remote sensing image data, and ground monitoring data, achieving multi-dimensional and multi-scale data complementarity and enhancement. Riverbed sampling decisions based on the fused data help generate more accurate sampling plans based on riverbed geomorphological characteristics and dynamic environmental changes, thereby improving the scientific nature and effectiveness of sampling. During the sampling execution phase, the riverbed manipulator performs path planning tasks, ensuring the automation and accuracy of the sampling process. At the same time, key data during the execution process is recorded in real time through the collection operation log, achieving traceability and transparency of the sampling task. After the acquisition is completed, the sampled data is promptly transmitted through the data return mechanism, and performance indicators are evaluated in combination with the return log. Quantitative analysis is conducted on multiple aspects such as data transmission delay, data integrity, and sampling success rate to ensure data reliability and accuracy. Ultimately, a riverbed quality collection report was automatically generated based on the assessment data, realizing a data-driven intelligent report generation method, significantly reducing the need for manual intervention, and improving the accuracy and timeliness of the report.

[0010] Preferably, step S1 includes the following steps:

[0011] Step S11: Set the accuracy of the BeiDou RTK positioning sensor to ±1cm and the acquisition frequency to 10Hz;

[0012] Step S12: Setting the detection depth of the high-frequency sonar sensor to 0.5-10m, the resolution to ±2cm, and the acquisition frequency to 5Hz;

[0013] Step S13: Set the pressure sensor's range to 0-50 kPa, its accuracy to ±0.5% FS, and its acquisition frequency to 20 Hz;

[0014] Step S14: using the Beidou RTK positioning sensor to collect riverbed latitude and longitude information; using the high-frequency sonar sensor to collect riverbed layer structure; using the pressure sensor to collect riverbed texture hardness data;

[0015] Step S15: Packing the riverbed latitude and longitude information, riverbed layer structure, and riverbed texture hardness data to form a Beidou satellite multidimensional data set;

[0016] Step S16: Preprocessing the raw data of the Beidou satellite multidimensional dataset to obtain a Beidou multidimensional standard dataset.

[0017] The present invention achieves precise collection of multi-dimensional riverbed data by providing Beidou RTK positioning sensors, high-frequency sonar sensors, and pressure sensors. The Beidou RTK positioning sensors provide riverbed latitude and longitude information with a high precision of ±1 cm and a collection frequency of 10 Hz, ensuring the accuracy of spatial location data and providing a reliable foundation for subsequent geographic information mapping and location association. The high-frequency sonar sensor, with a detection depth of 0.5-10 m and a resolution of ±2 cm, enables refined detection of the riverbed's layered structure. Acquiring acoustic signals at a frequency of 5 Hz helps identify the riverbed's sedimentary layers, bedrock layers, and heterogeneous structures, thereby providing accurate data for geological evolution analysis. The pressure sensor captures riverbed texture hardness data within a range of 0-50 kPa, with an accuracy of ±0.5% FS and a collection frequency of 20 Hz. This effectively characterizes the sediment density and rock strength of the riverbed, providing an important basis for riverbed stability analysis and erosion risk assessment. During the data packaging phase, the Beidou satellite multi-dimensional dataset is formed by integrating riverbed latitude and longitude information, layered structure data, and texture hardness data, achieving effective integration of multi-source data. Subsequently, the raw data preprocessing phase includes outlier detection, data smoothing, temporal synchronization, and spatial correction of the BeiDou satellite multidimensional dataset to eliminate noise and errors and ensure data consistency and integrity. The standardized BeiDou multidimensional dataset is highly accurate and compatible, facilitating subsequent data fusion, modeling analysis, and visualization applications.

[0018] Preferably, step S16 includes the following steps:

[0019] Step S161: using statistical filtering to remove suspended object noise from the Beidou satellite multidimensional dataset to obtain Beidou satellite noise reduction data, wherein the statistical filtering area radius is 0.1m;

[0020] Step S162: performing window filtering and jitter elimination processing on the Beidou satellite noise reduction data to obtain Beidou satellite jitter elimination data, wherein the window length of the window filtering and jitter elimination is 5s;

[0021] Step S163: performing data set standardization on the Beidou satellite jitter elimination data to obtain a Beidou multi-dimensional standard data set.

[0022] The present invention uses a series of refined data processing steps to remove noise, eliminate jitter, and standardize Beidou satellite multidimensional data sets, thereby significantly improving the quality and reliability of the data. In the data denoising stage, statistical filtering with a field radius of 0.1m is used to effectively identify and remove abnormal data points, especially noise caused by environmental disturbances, instrument errors, or suspended matter in riverbed water. Statistical filtering determines the deviation between data points and neighboring data, eliminating outliers while retaining the local characteristics of the data, ensuring the integrity and authenticity of spatial information. In the jitter removal stage, a window filtering method with a window length of 5s is used to smooth the jitter in the Beidou satellite noise-reduced data. Window filtering effectively reduces data jitter caused by satellite signal fluctuations or vibrations during the robotic arm sampling process by calculating the mean or median of the data within a fixed time window. Through this step, the temporal continuity and trend stability of the data are significantly improved. Finally, in the data standardization stage, the Beidou satellite jitter-removed data is normalized or mean-variance normalized to eliminate dimensional differences between different data dimensions and maintain consistency in data characteristics. The standardized BeiDou multidimensional standard dataset boasts improved numerical stability and model compatibility, providing a high-quality data foundation for subsequent multi-source data fusion, path planning, and riverbed sampling. Furthermore, the dataset's high accuracy and consistency will improve model training efficiency and prediction accuracy, further enhancing the scientific nature and reliability of riverbed monitoring and analysis.

[0023] Preferably, the multi-source data fusion in step S2 includes:

[0024] Extract Beidou RTK coordinates and sonar point cloud data based on Beidou multi-dimensional standard dataset;

[0025] Construct global free coordinates based on BeiDou RTK coordinates and BeiDou multi-dimensional standard data sets to obtain carrier pose coordinate data;

[0026] Obtain historical sedimentary layer data;

[0027] The historical sedimentary layer data and sonar point cloud data were segmented into sediment-gravel layers using the preset judgment criteria, and a layered planar surface diagram was modeled to obtain a riverbed layered planar surface diagram. The preset judgment criteria were: a density less than or equal to 1.8 g / cm³ was considered a sediment layer; a density greater than or equal to 2.2 g / cm³ was considered a gravel layer.

[0028] The carrier position coordinate data is used to mark the coordinates of the riverbed stratification surface map to generate multi-source fusion data of the riverbed.

[0029] This invention achieves precise construction and coordinate annotation of riverbed stratification maps through refined processing of the Beidou multidimensional standard dataset and multi-source data fusion, significantly improving the integrity and spatial accuracy of riverbed geological information. During the data extraction phase, Beidou RTK coordinates and sonar point cloud data are first obtained from the Beidou multidimensional standard dataset. The Beidou RTK coordinates provide centimeter-level location data, ensuring the accuracy of the riverbed's geographic location information. The sonar point cloud data, a dense point set formed by high-frequency sonar scanning, comprehensively reflects the riverbed's topography and stratification characteristics. During the global coordinate construction phase, a precise mapping of the Beidou RTK coordinates and the carrier's six-degree-of-freedom pose information is performed to generate real-time six-degree-of-freedom pose coordinate data for the carrier, accurately recording the spatial position and attitude changes of the riverbed sampling equipment in the underwater environment. This process not only provides an accurate positioning reference for subsequent riverbed data modeling but also effectively addresses the uncertainty caused by equipment drift and positioning errors during underwater operations. Subsequently, by integrating historical sedimentary layer data with sonar point cloud data, a judgment criterion set based on density thresholds is used to analyze the riverbed stratification. Data with a density less than or equal to 1.8g / cm³ is identified as a silt layer, and data with a density greater than or equal to 2.2g / cm³ is identified as a gravel layer, achieving accurate segmentation of the silt layer and the gravel layer. This density standard combines the long-term monitoring results of historical sedimentary layer data with the real-time reflection of sonar data, ensuring the scientific nature and reliability of the stratification results. The riverbed stratification surface map generated based on the stratification results intuitively displays the spatial distribution and morphological characteristics of different layers of the riverbed. Finally, the surface map is accurately annotated with the carrier pose coordinate data, further enhancing the spatial reference of the surface map and facilitating subsequent geological analysis and riverbed evolution monitoring. The annotated riverbed multi-source fusion data not only contains detailed geological information, but also retains the motion trajectory and spatial position of the sampling equipment, providing comprehensive data support for underwater geological research.

[0030] Preferably, the generation of the riverbed sampling decision in step S2 includes:

[0031] Obtain real-time riverbed acquisition data;

[0032] The real-time riverbed data is detected in real time using multi-source fusion data from the riverbed to obtain riverbed sampling decisions, which include sediment layer mode and gravel layer mode. If a sediment layer is detected, a bucket sampler is used to perform sampling based on the sediment layer mode of the riverbed sampling decision. The sampling operation has an opening diameter of 20 cm and a sampling volume of 500 ml per time to obtain sediment layer sampling data.

[0033] If a gravel layer is detected, based on the gravel layer pattern of the riverbed sampling decision, a cone sampler is used to perform sampling operations with a penetration depth of 10-50 cm to obtain gravel layer sampling data;

[0034] If other layers are detected, sampling is stopped and sampling data of unknown layers are obtained to complete the riverbed sampling decision.

[0035] This invention combines real-time riverbed data with multi-source fusion data to achieve intelligent stratified identification of riverbed quality and an adaptive sampling strategy, significantly improving the accuracy and efficiency of the sampling process. During the data acquisition phase, real-time riverbed data, including sonar point cloud data, pressure sensor data, and Beidou RTK positioning data, ensures the timeliness and spatial accuracy of riverbed geological information. Multi-source fusion data serves as a reference, providing a reliable benchmark for real-time detection through combined analysis of historical sedimentary data and real-time monitoring data. When using multi-source fusion data for real-time detection, a comprehensive assessment of multi-dimensional features, such as echo signal intensity from the sonar point cloud data, density information from pressure sensor feedback, and location coordinates, allows for precise differentiation between silt and gravel layers. For detected silt layers, sampling is performed using a bucket sampler with a 20 cm opening diameter and a single sampling volume of 500 ml. The bucket sampler's design ensures efficient sampling of loose, low-density silt layers while preserving the pristine state of the silt layer, facilitating subsequent analysis of particulate matter composition and suspended matter concentration. For the detected gravel layer, sampling is carried out using a cone sampler with a penetration depth of 10-50cm. The cone sampler has strong penetrating power and pressure resistance, and can achieve deep sampling in the hard gravel layer, effectively capturing geological characteristics at different depths, and providing basic data for analysis of gravel particle size distribution, mineral composition, and layered structure. During the decision-making execution process, the sampling equipment relies on the intelligent control system to complete the rapid switching and dynamic adjustment of the sampling strategy to ensure optimal sampling under different geological conditions. The entire process not only reduces human intervention, but also avoids unnecessary repeated sampling and sampling errors. Ultimately, through this intelligent riverbed sampling decision-making method, the obtained sediment layer and gravel layer sampling data are highly representative and accurate, and can provide a scientific basis for riverbed geological evolution analysis, water environment monitoring, and river management.

[0036] Preferably, the sampling path planning in step S3 includes:

[0037] Mark underground sampling points according to the underground sampling grid map to obtain sampling point marking data;

[0038] Generate underground sampling paths based on sampling focus mark data;

[0039] Perform shortest path optimization based on the underground sampling path to obtain the shortest sampling path optimization data;

[0040] According to the multi-source fusion data of the riverbed, the path adjustment planning is carried out on the shortest sampling path optimization data to obtain the riverbed path planning data.

[0041] This invention combines real-time riverbed data with multi-source fusion data to achieve intelligent stratified identification of riverbed quality and an adaptive sampling strategy, significantly improving the accuracy and efficiency of the sampling process. During the data acquisition phase, real-time riverbed data, including sonar point cloud data, pressure sensor data, and Beidou RTK positioning data, ensures the timeliness and spatial accuracy of riverbed geological information. Multi-source fusion data serves as a reference, providing a reliable benchmark for real-time detection through combined analysis of historical sedimentary data and real-time monitoring data. When using multi-source fusion data for real-time detection, a comprehensive assessment of multi-dimensional features, such as echo signal intensity from the sonar point cloud data, density information from pressure sensor feedback, and location coordinates, allows for precise differentiation between silt and gravel layers. For detected silt layers, sampling is performed using a bucket sampler with a 20 cm opening diameter and a single sampling volume of 500 ml. The bucket sampler's design ensures efficient sampling of loose, low-density silt layers while preserving the pristine state of the silt layer, facilitating subsequent analysis of particulate matter composition and suspended matter concentration. For the detected gravel layer, sampling is carried out using a cone sampler with a penetration depth of 10-50cm. The cone sampler has strong penetrating power and pressure resistance, and can achieve deep sampling in the hard gravel layer, effectively capturing geological characteristics at different depths, and providing basic data for analysis of gravel particle size distribution, mineral composition, and layered structure. During the decision-making execution process, the sampling equipment relies on the intelligent control system to complete the rapid switching and dynamic adjustment of the sampling strategy to ensure optimal sampling under different geological conditions. The entire process not only reduces human intervention, but also avoids unnecessary repeated sampling and sampling errors. Ultimately, through this intelligent riverbed sampling decision-making method, the obtained sediment layer and gravel layer sampling data are highly representative and accurate, and can provide a scientific basis for riverbed geological evolution analysis, water environment monitoring, and river management.

[0042] Preferably, obtaining the riverbed collection operation log in step S3 includes:

[0043] The riverbed manipulator performs a lowering operation at a speed of 0.1 m / s according to the riverbed path planning data, and uses a bucket sampler or a cone sampler for sampling. When the pressure sensor is greater than or equal to 5 kPa, it is determined to have touched the bottom, and a sampling instruction is generated;

[0044] The bucket sampler's sampling opening and closing time was set to 2s, and the sampling depth was 0.3-0.5m, and the bucket sampling log data was obtained;

[0045] The cone sampler speed was set to 10 rpm and the sampling depth was 0.5-1 m to obtain cone sampler log data;

[0046] The bucket sampling log data and the cone sampler log data are summarized into an operation log to obtain a riverbed sampling operation log.

[0047] This invention achieves intelligent lowering and sampling of the riverbed manipulator through precise execution based on riverbed path planning data and real-time judgment based on sensor feedback, significantly improving the accuracy and efficiency of the sampling process. During the manipulator's lowering phase, it descends steadily at a set speed of 0.1m / s, effectively avoiding equipment collisions or positional shifts caused by excessive speed, while ensuring the manipulator maintains good balance and motion control accuracy in the underwater environment. When the manipulator approaches the riverbed, a pressure sensor is used to detect the bottoming out in real time, and the contact status of the riverbed surface is determined using a threshold of 5kPa, avoiding misjudgments caused by visual or sonar errors. Real-time feedback from pressure data significantly improves the accuracy of bottoming out detection, providing a reliable basis for generating sampling instructions. During the sampling execution phase, differentiated sampling strategies are adopted for different types of riverbed properties. For detected sediment layers, the bucket sampler ensures stable sampling in loose sediment environments through a sampling opening and closing time of 2s and a sampling depth of 0.3-0.5m. The structural features of the bucket sampler enable it to acquire large-volume samples in a short period of time, effectively preserving the original physical state of the sediment layer while avoiding disturbance-induced damage to sample stratification. For the gravel layer detected, the cone sampler operates at a rotation speed of 10 rpm and a sampling depth of 0.5-1 m. The cone sampler penetrates the gravel layer through rotary drilling, effectively capturing deep rock particles and internal structural information of the gravel layer, while ensuring the sampler remains stable in high-hardness environments. Through the precise setting of parameters for different samplers, the sampling process takes into account both data integrity and representativeness. In addition, the riverbed acquisition operation log generated during the sampling process records in detail key parameters such as the robotic arm's position information, pressure feedback data, sampling execution time, sampling depth, and equipment status, providing basic data support for subsequent equipment operation evaluation and data traceability. The automated recording of the riverbed acquisition operation log not only improves the transparency of data collection, but also provides a decision-making basis for the optimization of sampling strategies and equipment maintenance.

[0048] Preferably, step S4 includes the following steps:

[0049] Step S41: Acquire historical riverbed collection data; use the riverbed collection operation log to return data packets to obtain a riverbed collection return log;

[0050] Step S42: Perform performance index evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data;

[0051] Step S43: Use historical riverbed collection data and riverbed collection assessment data to compare collection efficiency, and construct an automatic riverbed quality collection report.

[0052] This invention enables quantitative analysis and optimized decision-making during the acquisition process by transmitting back riverbed acquisition operation data, performing performance evaluation, and comparing efficiency, thereby improving the accuracy and efficiency of riverbed acquisition at the data level. During the data transmission phase, the riverbed acquisition return log, generated from the riverbed acquisition operation log, comprehensively records key data such as equipment operating status, location information, sampling parameters, pressure feedback, and sampling depth, ensuring the timeliness and integrity of data transmission. Automated processing of the return log reduces manual intervention and provides a comprehensive data foundation for subsequent performance evaluation. During the performance evaluation phase, key indicators such as sampling success rate, sampling depth deviation, equipment operational stability, and data integrity are extracted from the return log to generate riverbed acquisition evaluation data. Through multi-dimensional quantitative analysis, this data comprehensively reflects the equipment's sampling performance in different riverbed environments, thereby identifying potential factors affecting sampling quality. This performance evaluation data not only provides a reference for equipment maintenance and operational optimization, but also provides data support for improving sampling strategies. During the acquisition efficiency comparison phase, historical riverbed acquisition data is compared with riverbed acquisition evaluation data to analyze trends in sampling time, sampling success rate, and data integrity. By comparing efficiency differences under different sampling conditions, we can identify optimal sampling parameter configurations and equipment operation strategies, further improving sampling accuracy and reliability. Finally, based on data-driven analysis, we use the assessment data and comparison results to generate an automated riverbed quality collection report. This report not only visually demonstrates the stratification characteristics and physical properties of the riverbed quality but also provides performance indicators and improvement suggestions during the sampling process, facilitating subsequent riverbed monitoring and environmental assessments. This automated report generation process reduces human error and ensures the objectivity and timeliness of the report content.

[0053] Preferably, step S42 includes the following steps:

[0054] Step S421: Performing a synchronous evaluation of the robot arm sensor on the riverbed collection and return log to obtain riverbed sensor synchronous evaluation data;

[0055] Step S422: Analyze the sampler operation efficiency of the riverbed collection return log to obtain riverbed quality sampling efficiency data;

[0056] Step S423: Perform water flow error analysis on the riverbed collected and returned logs. If the flow rate is greater than 0.02 m / s, it is marked as riverbed environmental parameter data.

[0057] Step S424: calibrate the riverbed sensor synchronous evaluation data, riverbed quality sampling efficiency data, and riverbed environmental parameter data to obtain riverbed collection evaluation data.

[0058] This invention achieves precise assessment and quality control of the riverbed acquisition process through multi-dimensional analysis and data calibration of riverbed acquisition return logs, effectively improving data reliability and sampling efficiency. During the sensor synchronization assessment phase, riverbed sensor synchronization assessment data is generated by performing time alignment and synchronization error detection on multiple sources of information recorded in the return logs, including timestamps, location information, pressure data, and sonar echo data. This assessment method can identify timing errors caused by sensor data delays, signal loss, or inconsistent sampling frequencies, ensuring temporal consistency of data from various sensor types. Furthermore, by analyzing sensor operating status and data integrity, abnormal sensors can be promptly detected, providing a basis for subsequent data repair and equipment maintenance. During the sampling operation efficiency analysis phase, key parameters such as the manipulator lowering speed, bottoming detection time, sampling execution time, and deviation between the sample volume and the target volume are extracted from the riverbed acquisition return logs to generate riverbed quality sampling efficiency data. By comparing against preset sampling efficiency standards, the sampling equipment's execution efficiency and task completion rate can be effectively assessed, identifying issues such as motion delays, manipulator jitter, or sampler failures during the sampling process. This analysis process not only quantifies the equipment's operational performance but also provides data support for optimizing sampling paths and parameters. During the assessment data calibration phase, riverbed sensor data and riverbed quality sampling efficiency data are synchronized. Multi-dimensional data cross-validation and outlier removal further eliminate potential measurement errors and system deviations, generating riverbed collection assessment data. Calibrated assessment data exhibits increased accuracy and consistency, serving as an important basis for subsequent sampling strategy adjustments and equipment performance optimization. Furthermore, riverbed collection assessment data can be used to establish a historical sampling performance benchmark, supporting the prediction and real-time control of future sampling tasks.

[0059] In this specification, a riverbed quality automatic collection and control system based on Beidou navigation is provided, which is used to execute the above-mentioned riverbed quality automatic collection and control method based on Beidou navigation. The riverbed quality automatic collection and control system based on Beidou navigation includes:

[0060] The Beidou data processing module is used to plan sampling paths using riverbed sampling decisions and obtain riverbed path planning data. It generates collection instructions based on the riverbed path planning data and uses the riverbed robotic arm to perform real-time judgment and execution, ultimately generating a riverbed collection operation log.

[0061] The multi-source data fusion module is used to fuse multi-source data based on the Beidou multi-dimensional standard data set to obtain riverbed multi-source fusion data; and use the riverbed multi-source fusion data to generate riverbed sampling decisions;

[0062] The path planning and execution module is used to plan the sampling path using the riverbed sampling decision and obtain the riverbed path planning data; based on the riverbed path planning data, the sampling instructions are generated and executed by the riverbed robotic arm to obtain the riverbed sampling operation log;

[0063] The data return and evaluation module is used to use the riverbed collection operation log to return data packets and obtain the riverbed collection return log; perform performance indicator evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; and construct an automatic riverbed quality collection report based on the riverbed collection evaluation data. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 The present invention is a flowchart of a method for automatically collecting and controlling riverbed quality based on Beidou navigation;

[0065] Figure 2 for Figure 1 Detailed implementation steps of step S4 in FIG.

[0066] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0067] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the embodiments described are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making any creative work are within the scope of protection of the present invention.

[0068] In addition, the accompanying drawings are merely schematic illustrations of the present invention and are not necessarily drawn to scale. Identical reference numerals in the figures denote identical or similar parts, and thus repetitive descriptions thereof will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities that do not necessarily correspond to physically or logically separate entities. These functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor and / or microcontroller approaches.

[0069] It should be understood that although the terms "first," "second," and the like may be used herein to describe various elements, these elements should not be limited by these terms. These terms are used solely to distinguish one element from another. For example, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element, without departing from the scope of the exemplary embodiments. The term "and / or" as used herein includes any and all combinations of one or more of the listed associated items.

[0070] To achieve this, please refer to Figures 1 to 2 A method for automatically collecting and controlling riverbed quality based on Beidou navigation, comprising the following steps:

[0071] Step S1: obtaining a Beidou satellite multidimensional dataset; performing raw data preprocessing on the Beidou satellite multidimensional dataset to obtain a Beidou multidimensional standard dataset;

[0072] Step S2: Perform multi-source data fusion based on the Beidou multi-dimensional standard data set to obtain riverbed multi-source fusion data; use the riverbed multi-source fusion data to generate riverbed sampling decisions;

[0073] Step S3: Using the riverbed sampling decision to plan the sampling path, obtain riverbed path planning data; based on the riverbed path planning data, generate a collection instruction, and use the riverbed robotic arm to perform real-time judgment and execution, and finally obtain a riverbed collection operation log;

[0074] Step S4: using the riverbed collection operation log to transmit data packets back to obtain the riverbed collection return log; performing performance index evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; constructing a riverbed quality automatic collection report based on the riverbed collection evaluation data.

[0075] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a flow chart of a method for automatically collecting and controlling riverbed quality based on Beidou navigation according to the present invention. In this example, the method for automatically collecting and controlling riverbed quality based on Beidou navigation includes the following steps:

[0076] Step S1: obtaining a Beidou satellite multidimensional dataset; performing raw data preprocessing on the Beidou satellite multidimensional dataset to obtain a Beidou multidimensional standard dataset;

[0077] In this embodiment of the present invention, this method achieves data standardization and improved reliability by acquiring a Beidou satellite multidimensional dataset and preprocessing the raw data, providing a high-precision data foundation for subsequent riverbed analysis and sampling tasks. During the data acquisition phase, the Beidou satellite multidimensional dataset includes high-precision positioning data, satellite orbit data, signal strength data, and timestamp information. Using Beidou RTK (Real-Time Kinematic) positioning technology and a differential algorithm to correct satellite positioning errors, this method enables real-time acquisition of riverbed sampling equipment location information with centimeter-level accuracy, ensuring the reliability of the data's spatial accuracy. Furthermore, satellite constellation information and multipath effect detection data are used to perform preliminary screening for potential signal obstructions and errors, effectively reducing positioning drift. During the raw data preprocessing phase, the data is first detected and removed for outliers caused by satellite signal interference, equipment failure, or environmental changes. A statistical filtering method based on mean deviation is used to detect outliers, and local correction is performed in conjunction with the neighborhood radius method to ensure data integrity and accuracy. Subsequently, a time synchronization algorithm is used to align the multi-source data to address timing inconsistencies caused by varying sensor sampling frequencies and data transmission delays. In response to noise interference in positioning data, the Kalman filter algorithm is used for smoothing processing to further optimize data accuracy and effectively eliminate high-frequency noise caused by equipment vibration or water flow disturbance. In terms of spatial correction, the data is converted into geographic coordinates and error compensation is performed through the coordinate reference of satellite orbit data and ground base stations to ensure the spatial consistency of multi-dimensional data in a unified coordinate system. In addition, linear interpolation is used to fill small-scale gaps in the data to avoid information loss due to incomplete data. After multiple preprocessing, the generated Beidou multi-dimensional standard data set has high precision, high integrity and high spatiotemporal consistency, providing reliable data support for subsequent data fusion, path planning and sampling execution, while providing accurate spatial positioning information for riverbed stratification modeling and sediment analysis.

[0078] Step S2: Perform multi-source data fusion based on the Beidou multi-dimensional standard data set to obtain riverbed multi-source fusion data; use the riverbed multi-source fusion data to generate riverbed sampling decisions;

[0079] In this embodiment of the present invention, multi-source data fusion is performed on the Beidou multi-dimensional standard dataset, integrating and complementing data from different sensors, thereby improving the accuracy and reliability of riverbed sampling decisions. Specific technical approaches include multi-source data fusion algorithms, data registration, and joint optimization techniques. These combine sensor data from position, sonar, and pressure sensors to provide a more comprehensive understanding of the riverbed environment. During the multi-source data fusion phase, data from multiple sources, including satellite positioning data, sonar echo data, and pressure sensor data, are first extracted from the Beidou multi-dimensional standard dataset. These data sources have different spatial distributions, time scales, and accuracy requirements. To ensure that these data can be effectively analyzed within the same framework, spatial and temporal alignment of the different data sources is first required. This process is achieved through time synchronization algorithms and spatial registration techniques, ensuring consistency across the spatial and temporal dimensions of the various data types. After data alignment, the data sources are fused using fusion techniques such as weighted averaging or Kalman filtering. This process applies a data weighting strategy to optimize the fusion results by assigning different weights to each data source, taking into account the varying error characteristics of each data source (e.g., satellite positioning data affected by signal obstruction or sonar data with echo distortion). Secondly, methods such as mutual information metric and particle filtering from information theory are applied to further address uncertainty in the data, improving the reliability and accuracy of the fusion results. The fused dataset has been significantly improved in terms of spatial accuracy, temporal consistency, and information density, forming a more comprehensive multi-source fusion data of the riverbed. Based on this multi-source fusion data, the system can generate accurate riverbed sampling decisions. This decision is made by combining data from different sensors to conduct a comprehensive analysis of the riverbed's physical properties, depth levels, structural characteristics, and other aspects, thereby selecting the most appropriate sampling points, samplers, and sampling strategies. In the decision-making process, rule-based reasoning methods and machine learning algorithms are used, combined with historical sampling data and current environmental changes, to provide a scientific basis for the sampling process and maximize the representativeness and accuracy of the sampled data.

[0080] Step S3: Using the riverbed sampling decision to plan the sampling path, obtain riverbed path planning data; based on the riverbed path planning data, generate a collection instruction, and use the riverbed robotic arm to perform real-time judgment and execution, and finally obtain a riverbed collection operation log;

[0081] In embodiments of the present invention, path planning is performed based on the riverbed sampling decision results, combined with the execution mechanism of the riverbed manipulator. This optimizes the path design and operation execution of the sampling process, thereby improving sampling efficiency and data quality. Specific technical measures include path planning algorithms, kinematic model optimization, and execution feedback mechanisms. These combine sampling decision data with manipulator control data to achieve precise execution of the sampling task. The core of sampling path planning is to utilize the optimal sampling point data, terrain information, and sampler type provided by the riverbed sampling decision, applying a graph search algorithm (Dijkstra's algorithm) or a shortest path planning algorithm to calculate the optimal path from the current position to the target sampling point. These path planning algorithms can generate the shortest, safest, and most efficient sampling path based on the actual riverbed topography and water flow environment, as well as the manipulator's working range and load limits. The path planning process incorporates strategies such as obstacle avoidance, dynamic adjustment, and energy efficiency optimization to ensure that the path design maximizes the efficiency of the sampling operation while taking into account environmental changes. To further enhance the flexibility and adaptability of path planning, reinforcement learning methods can be introduced to dynamically adjust the path during the actual sampling task and optimize based on real-time environmental feedback. Next, the collection instructions generated based on the riverbed path planning data involve the kinematic control of the robotic arm. By generating the robotic arm trajectory and adjusting its speed based on the path planning results, the sampler can accurately reach the designated location and maintain stability. During this process, the robotic arm's inverse kinematics algorithm is used to calculate the angles of each joint, ensuring that it can complete the sampling movement according to the predetermined trajectory. Furthermore, the path planning data is linked to sampling task parameters such as the robotic arm's load, sampler opening and closing timing, and sampling depth to ensure accurate sampling execution. Finally, the collection instructions interact with the robotic arm's control system interface to control the precise movement of the robotic arm and record key sampling process data (such as sampling time, sampling volume, sampling location, and sampling depth) in the riverbed collection operation log. This data not only reflects the execution status of the sampling process but also provides feedback on execution efficiency and quality, supporting the optimization of subsequent sampling tasks and equipment maintenance.

[0082] Step S4: using the riverbed collection operation log to transmit data packets back to obtain the riverbed collection return log; performing performance index evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; constructing a riverbed quality automatic collection report based on the riverbed collection evaluation data.

[0083] In this embodiment of the present invention, the transmission and analysis of riverbed collection operation logs enables performance evaluation and quality control of the collection process, providing reliable data support for the generation of automatic riverbed quality collection reports. Specific technical measures include a data transmission mechanism, a performance evaluation algorithm, and a report generation model. Comprehensive data analysis is performed by combining operation logs and performance indicators. First, the transmission of riverbed collection operation logs utilizes a communication network and data transmission protocols (such as TCP / IP and MQTT) to ensure data real-time and integrity. The data packet transmission process involves two-way communication between the collection equipment and the central control system. Reliable data transmission protocols ensure that the collected log data is not lost or tampered with during transmission, and that real-time synchronization is achieved. The transmitted data includes key information such as sampling time, location, sampling depth, robotic arm operating status, and equipment health status, providing the raw data for subsequent data processing and evaluation. During the post-transmission data processing, a performance evaluation algorithm is used to analyze the collection process. First, a data preprocessing algorithm removes noise and detects outliers on the transmitted log data to ensure the accuracy of the evaluation. Then, based on key sampling task indicators (such as sampling efficiency, sampling accuracy, equipment operating time, and sampling failure rate), statistical methods and data mining techniques are used to calculate sampling task performance indicators, such as task completion time, sampling accuracy, path deviation, and equipment load. Furthermore, based on these performance indicators, an overall collection quality assessment is calculated through weighted averaging and normalization techniques. During the assessment process, machine learning algorithms can be used for regression analysis to establish a performance prediction model, automatically identifying key factors affecting collection performance, thereby enabling dynamic optimization and early warning of collection efficiency. Finally, based on the assessment results, a riverbed quality collection report is automatically generated using natural language generation technology and data visualization techniques according to a pre-set format and template. This report not only includes an overall evaluation of collection quality but also provides detailed analysis of each aspect, such as sampler efficiency, path planning rationality, and feedback on equipment health, providing a scientific basis for subsequent collection task optimization and equipment maintenance.

[0084] Preferably, step S1 includes the following steps:

[0085] Step S11: Set the accuracy of the BeiDou RTK positioning sensor to ±1cm and the acquisition frequency to 10Hz;

[0086] Step S12: Setting the detection depth of the high-frequency sonar sensor to 0.5-10m, the resolution to ±2cm, and the acquisition frequency to 5Hz;

[0087] Step S13: Set the pressure sensor's range to 0-50 kPa, its accuracy to ±0.5% FS, and its acquisition frequency to 20 Hz;

[0088] Step S14: using the Beidou RTK positioning sensor to collect riverbed latitude and longitude information; using the high-frequency sonar sensor to collect riverbed layer structure; using the pressure sensor to collect riverbed texture hardness data;

[0089] Step S15: Packing the riverbed latitude and longitude information, riverbed layer structure, and riverbed texture hardness data to form a Beidou satellite multidimensional data set;

[0090] Step S16: Preprocessing the raw data of the Beidou satellite multidimensional dataset to obtain a Beidou multidimensional standard dataset.

[0091] In this embodiment of the present invention, the Beidou RTK positioning sensor achieves centimeter-level positioning accuracy, providing precise riverbed latitude and longitude data. Technically, by employing real-time kinematic (RTK) technology, the Beidou RTK positioning system corrects satellite signal errors in real time, leveraging the differences between base stations and mobile receivers to improve positioning accuracy to 1cm. Furthermore, through 10Hz high-frequency data acquisition, it accurately tracks changes in the riverbed's position. This high-precision positioning data provides the fundamental basis for the riverbed's spatial coordinates, ensuring the accuracy of geographic location in subsequent data analysis.

[0092] The use of high-frequency sonar sensors also enables the detection of riverbed stratification. With a resolution of ±2 cm and a sampling frequency of 5 Hz, these sensors can acquire precise sonar echo data within a depth range of 0.5 to 10 meters. The time difference between sonar signal reflections is used to determine the physical characteristics of different riverbed layers. This data helps to create detailed riverbed stratification maps, providing essential information for riverbed topography and sediment analysis.

[0093] The pressure sensor measures the hardness of the riverbed with a range of 0-50 kPa, an accuracy of ±0.5% FS, and a sampling frequency of 20 Hz. By monitoring changes in riverbed hardness in real time, the physical properties of different geological layers can be accurately assessed. During the sampling process, the pressure sensor data provides a basis for selecting appropriate sampling tools and controlling sampling depth.

[0094] Through the collaborative work of these sensors, riverbed latitude and longitude information, hierarchical structure data, and texture and hardness data are packaged into a Beidou satellite multidimensional dataset. During the data fusion process, data integration methods are used to uniformly encode and format data from different sources to ensure comparability and consistency. During the data preprocessing stage, denoising, filtering, and normalization algorithms are applied to eliminate noise and bias in the sensor data, ensuring a high-quality Beidou multidimensional standard dataset. This provides reliable data support for subsequent riverbed data analysis, sampling decisions, and path planning.

[0095] Preferably, step S16 includes the following steps:

[0096] Step S161: using statistical filtering to remove suspended object noise from the Beidou satellite multidimensional dataset to obtain Beidou satellite noise reduction data, wherein the statistical filtering area radius is 0.1m;

[0097] Step S162: performing window filtering and jitter elimination processing on the Beidou satellite noise reduction data to obtain Beidou satellite jitter elimination data, wherein the window length of the window filtering and jitter elimination is 5s;

[0098] Step S163: performing data set standardization on the Beidou satellite jitter elimination data to obtain a Beidou multi-dimensional standard data set.

[0099] In an embodiment of the present invention, statistical filtering technology is applied to remove suspended matter noise from Beidou satellite multidimensional datasets. Suspended matter noise is typically random noise caused by air flow, temperature fluctuations, and other external interference within the sensor acquisition environment. Statistical filtering technology, based on the statistical characteristics of the data, utilizes information from a neighborhood with a radius of 0.1m to suppress or remove anomalous data, thereby reducing the impact of noise on subsequent data processing and analysis. This method effectively reduces the interference of high-frequency noise by evaluating the local mean or median of data points within a smaller spatial range, thereby improving data smoothness and reliability. Secondly, after data noise reduction, window filtering technology is applied to further de-jitter the data. Mechanical vibrations or external disturbances inherent in the sensor often cause small fluctuations or jitter in the collected data, affecting data smoothness. Window filtering uses a 5-second sliding window at each time point in the dataset and calculates the mean or other statistical measures of all data within this window. This effectively smooths out sudden or short-term jitter and eliminates data fluctuations caused by dynamic changes. This processing method can better preserve long-term data trends, reduce the interference of transient disturbances, and improve data accuracy and reliability. Finally, after noise reduction and jitter removal, Beidou satellite data needs to be normalized. Normalization is a crucial step in data preprocessing, aiming to eliminate the effects of dimensional differences and numerical ranges between different sensors or data sources. By normalizing each element in the Beidou satellite dataset, it is converted to standard data with zero mean and unit variance. This prevents certain features from dominating the model or results during subsequent multi-source data fusion and analysis due to significant numerical differences. This standardized Beidou multidimensional dataset is more consistent, facilitating subsequent data processing, analysis, and decision-making.

[0100] Preferably, the multi-source data fusion in step S2 includes:

[0101] Extract Beidou RTK coordinates and sonar point cloud data based on Beidou multi-dimensional standard dataset;

[0102] Construct global free coordinates based on BeiDou RTK coordinates and BeiDou multi-dimensional standard data sets to obtain carrier pose coordinate data;

[0103] Obtain historical sedimentary data;

[0104] The historical sedimentary layer data and sonar point cloud data were segmented into sediment-gravel layers using the preset judgment criteria, and a layered planar surface diagram was modeled to obtain a riverbed layered planar surface diagram. The preset judgment criteria were: a density less than or equal to 1.8 g / cm³ was considered a sediment layer; a density greater than or equal to 2.2 g / cm³ was considered a gravel layer.

[0105] The carrier position coordinate data is used to mark the coordinates of the riverbed stratification surface map to generate multi-source fusion data of the riverbed.

[0106] In this embodiment of the present invention, a high-precision coordinate acquisition method based on the RTK positioning system is employed to extract Beidou RTK coordinates and sonar point cloud data from the Beidou multi-dimensional standard dataset, ensuring centimeter-level positioning accuracy. This technical approach effectively eliminates errors in traditional GPS systems through real-time differential correction signals, providing highly accurate coordinate data. Next, using the Beidou RTK coordinates and the multi-dimensional dataset, a global free coordinate system is constructed to obtain the carrier's pose coordinate data. This step combines the inertial measurement unit (IMU) and GPS data to achieve accurate pose estimation and position calculation. This process effectively improves data accuracy and reliability by fusing Beidou RTK coordinate data with IMU data and employing techniques such as Kalman filtering, ensuring the accuracy of pose data during riverbed acquisition operations. Subsequently, the acquired historical sedimentary layer data and sonar point cloud data are combined and segmented according to pre-set criteria. Technically, using density-based segmentation techniques (density less than or equal to 1.8 g / cm³ is considered a sediment layer, and density greater than or equal to 2.2 g / cm³ is considered a gravel layer) and clustering algorithms, the sediment and sonar point cloud data are accurately segmented into sediment and gravel layers. Commonly used segmentation algorithms include K-means and DBSCAN, which effectively cluster data based on the characteristics of different data points, thereby accurately segmenting the sediment layers. After stratification, 3D modeling is used to generate a riverbed surface map based on the stratified data. This process leverages the spatial distribution of the point cloud data and uses a triangulated network (TIN) generation algorithm to reconstruct the riverbed surface, achieving accurate modeling of the riverbed quality. This modeling process can also be further integrated with Geographic Information Systems (GIS) technology to visualize the model, facilitating subsequent analysis and decision-making. Finally, the generated riverbed surface map is annotated using the carrier's pose coordinate data. This technical approach ensures the precise spatial location of each sediment layer by combining precise pose coordinates with topographic data. This process uses spatial registration technology to achieve high-precision geographic information integration by aligning and mapping data from different sources, and ultimately generates multi-source fusion data of the riverbed, providing high-quality input data for further riverbed collection and analysis.

[0107] Preferably, the generation of the riverbed sampling decision in step S2 includes:

[0108] Obtain real-time riverbed acquisition data;

[0109] The real-time riverbed data is detected in real time using multi-source fusion data from the riverbed to obtain riverbed sampling decisions, which include sediment layer mode and gravel layer mode. If a sediment layer is detected, a bucket sampler is used to perform sampling based on the sediment layer mode of the riverbed sampling decision. The sampling operation has an opening diameter of 20 cm and a sampling volume of 500 ml per time to obtain sediment layer sampling data.

[0110] If a gravel layer is detected, based on the gravel layer pattern of the riverbed sampling decision, a cone sampler is used to perform sampling operations with a penetration depth of 10-50 cm to obtain gravel layer sampling data;

[0111] If other layers are detected, sampling is stopped and sampling data of unknown layers are obtained to complete the riverbed sampling decision.

[0112] In this embodiment of the present invention, real-time riverbed data is acquired, using high-frequency sensors and equipment to continuously monitor various physical parameters of the riverbed. This data, including information such as riverbed depth, hardness, and structural stratification, is typically collected using sensors mounted on a robotic arm or other acquisition device, such as sonar sensors, pressure sensors, and displacement sensors. To ensure data accuracy and real-time performance, the acquisition frequency must be sufficiently high (e.g., multiple data points per second) to provide a real-time reflection of the riverbed's status within a short period of time. This process requires acquisition equipment with high precision and stability to meet measurement requirements in diverse environments and ensure data reliability and accuracy. Using these sensors, the system accurately captures various physical characteristics of the riverbed, generating comprehensive dynamic monitoring data. By fusion-enhancing multi-source riverbed data, real-time monitoring of the real-time riverbed data is performed. The key technical aspect of this process lies in the application of multi-source data fusion technology, typically employing weighted averaging, Kalman filtering, or deep learning algorithms to comprehensively analyze and process data from various sensors to obtain more accurate and comprehensive riverbed information. Data fusion effectively reduces noise and measurement errors, improves the signal-to-noise ratio, and further enhances the reliability and accuracy of the system. This fused data is used to monitor the physical properties of the riverbed in real time, particularly distinguishing between silt and gravel layers. During this process, a fusion algorithm models the temporal and spatial characteristics of the data to determine the current state of the riverbed in real time and distinguish different geological layers based on the data characteristics. This algorithm not only identifies the current layered structure of the riverbed but also analyzes dynamic changes in the riverbed at different points in time based on real-time data trends, further enhancing the intelligence and automation of the sampling process. When a silt layer is detected, the system automatically selects an appropriate sampler for sampling. Specifically, a bucket sampler with a 20cm opening diameter and a sampling volume of 500ml per sample is used. This sampler effectively obtains representative samples from the silt layer, ensuring that the collected data accurately reflects the characteristics of the riverbed's silt layers. When sampling the silt layer, the system adjusts the sampling depth and location in real time based on sensor data to ensure accurate and reliable sampling. Similarly, when a gravel layer is detected, the system switches to a cone sampler with a sampling depth range of 10-50cm, specifically designed for obtaining sample data from gravel layers. Due to its design, the cone sampler can effectively penetrate harder gravel layers and obtain sample data at deeper levels, ensuring that valuable representative data can be extracted from the gravel layer. By intelligently switching samplers, the system can perform sampling tasks more efficiently and accurately, minimizing manual intervention and improving the automation level of riverbed sampling and the accuracy of sampling data.

[0113] Preferably, the sampling path planning in step S3 includes:

[0114] Mark underground sampling points according to the underground sampling grid map to obtain sampling point marking data;

[0115] Generate underground sampling paths based on sampling focus mark data;

[0116] Perform shortest path optimization based on the underground sampling path to obtain the shortest sampling path optimization data;

[0117] According to the multi-source fusion data of the riverbed, the path adjustment planning is carried out on the shortest sampling path optimization data to obtain the riverbed path planning data.

[0118] In this embodiment of the present invention, gridding based on the sediment and gravel layer sampling data is accomplished using Geographic Information System (GIS) technology and spatial data analysis algorithms. By analyzing the spatial distribution of the sampling data, clustering or segmentation algorithms (such as K-means clustering or Voronoi segmentation) can be used to divide the underground area into several grids. The data within each grid represents the geological characteristics of the area (such as the distribution of the sediment or gravel layers). These grids not only provide spatial information about the underground structure but also serve as basic data for subsequent path planning. Next, the underground sampling path is generated based on the underground sampling grid map based on the real-time riverbed data. The key technical approach in this step lies in combining the real-time data with the pre-divided underground grid map, generating the sampling path through real-time updates and data fusion techniques. Real-time riverbed data, fed through sensor feedback, provides information such as riverbed depth, hardness, and stratification. This data helps the system dynamically adjust the sampling path to ensure that sampling activities cover all key areas, particularly those with high sampling value. Typically, optimization algorithms, such as the sparse matrix or Dijkstra algorithm, are used for path planning, with the shortest path, minimum energy consumption, etc. as optimization goals to ensure that the sampler can execute on the most reasonable trajectory. Path planning is performed based on the underground sampling path to obtain riverbed path planning data. Path planning technology generates an efficient path graph by considering the topographic changes of the riverbed, sensor data feedback, and equipment limitations (such as sampler depth and movement speed). In this process, the path planning algorithm needs to comprehensively consider multiple constraints, such as avoiding obstacles, ensuring that each sampling point can be effectively sampled, and maximizing sampling efficiency. Commonly used path planning methods include graph search-based algorithms (such as A* algorithm, genetic algorithm, etc.). Through these technologies, the generated path planning data will provide accurate navigation information for the sampling equipment.

[0119] Preferably, obtaining the riverbed collection operation log in step S3 includes:

[0120] The riverbed manipulator performs a lowering operation at a speed of 0.1 m / s according to the riverbed path planning data, and uses a bucket sampler or a cone sampler for sampling. When the pressure sensor is greater than or equal to 5 kPa, it is determined to have touched the bottom, and a sampling instruction is generated;

[0121] The bucket sampler's sampling opening and closing time was set to 2s, and the sampling depth was 0.3-0.5m, and the bucket sampling log data was obtained;

[0122] The cone sampler speed was set to 10 rpm and the sampling depth was 0.5-1 m to obtain cone sampler log data;

[0123] The bucket sampling log data and the cone sampler log data are summarized into an operation log to obtain a riverbed sampling operation log.

[0124] In this embodiment of the present invention, riverbed path planning data is used to lower the riverbed manipulator at a speed of 0.1 m / s. This is achieved through a path planning algorithm, primarily relying on the pre-designed lowering trajectory in the riverbed path planning data. The key to path planning is to design an optimal lowering path based on the riverbed's topographical characteristics, the sampler's size and limitations, and real-time data collection. This ensures that the manipulator can smoothly and efficiently bring the sampler to its target location while avoiding obstacles and unnecessary deviations. The manipulator's control system receives the path planning data in real time and executes the lowering operation through a precise motion control system, ensuring that the set speed requirements are met with accuracy. Pressure sensors play a crucial role in the lowering process. After the manipulator is lowered to a certain depth, the pressure sensor monitors the pressure feedback to determine whether the sampler has reached the bottom. When the pressure sensor reading reaches or exceeds 5 kPa, the system determines that the sampler has reached the bottom. At this point, the bottoming signal triggers the generation of a sampling command. Specifically, the collection command is transmitted to the manipulator through the control system, instructing it to initiate sampling. This process is based on sensor signal processing, combining real-time data flow and control algorithms to ensure accurate determination of the sampler's bottoming. In terms of sampler control, the parameter settings of the bucket sampler and cone sampler are key. The sampling opening and closing time of the bucket sampler is set to 2 seconds, and the sampling depth is 0.3 to 0.5 meters. This is achieved by adjusting the operating mechanism of the sampler through the control system. The system accurately controls the opening and closing actions of the sampler according to the data requirements to ensure that each sampling is carried out within the specified time and depth range. In addition, the rotation speed of the cone sampler is set to 10rpm, and the sampling depth is between 0.5 and 1 meter. These parameters are set through the sampler control system to achieve efficient sampling of different types of riverbed textures. The rotation speed of the cone sampler is set to 10rpm to ensure sufficient power during the sampling process and to successfully complete the task of obtaining samples from the riverbed. Finally, the real-time control and adjustment of these actions and parameters are recorded in the "Riverbed Collection Operation Log" through the execution and feedback mechanism of the robotic arm. The log contains detailed data for each sampling, including path data, pressure value, sampling time and depth information, etc.

[0125] As an example of the present invention, refer to Figure 2 As shown, in this example, step S4 includes:

[0126] Step S41: Acquire historical riverbed collection data; use the riverbed collection operation log to return data packets to obtain a riverbed collection return log;

[0127] Step S42: Perform performance index evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data;

[0128] Step S43: Use historical riverbed collection data and riverbed collection assessment data to compare collection efficiency, and construct an automatic riverbed quality collection report.

[0129] In this embodiment of the present invention, historical riverbed collection data is acquired and transmitted back through the riverbed collection operation log, generating a riverbed collection return log. This log not only records various sampling parameters (such as sampling depth, time, and sensor data), but also monitors device status and any anomalies during the sampling process. The core technology behind this data packet transmission is the real-time transmission of operating data from local devices to a central database or server via an efficient communication mechanism for subsequent analysis and processing. Specifically, this communication mechanism utilizes a high-bandwidth, low-latency network protocol and data compression algorithm to ensure data transmission accuracy while improving transmission speed and stability. Real-time transmission of this collected data enables real-time monitoring and early warning of the riverbed collection process, providing accurate data support for subsequent analysis. Performance indicators are evaluated based on the riverbed collection return log to generate riverbed collection evaluation data. The key to this process lies in meticulously analyzing the data in the return log to assess various performance indicators, such as sampling efficiency, accuracy, and device operating status. Specifically, performance evaluation relies on quantitative analysis of key data from the sampling process and scoring the collection effectiveness using specific evaluation models or algorithms (such as statistical analysis or regression analysis). These analytical results help understand whether sampling operations are proceeding smoothly according to planned objectives and identify potential performance bottlenecks or areas for improvement. For example, analyzing the relationship between sampling depth and equipment load can assess equipment durability and stability; comparing sampling success rate with sampling time can assess sampling efficiency. These analytical results not only provide a data basis for optimizing the current collection process but also support decision-making for equipment fault diagnosis and maintenance. Historical riverbed collection data is compared with riverbed collection evaluation data to generate an automated riverbed quality collection report. The comparison of evaluation data with historical data measures efficiency changes under different sampling strategies and operating modes, enabling in-depth analysis of the impact of different sampling methods or equipment configurations on sampling results. Through comparative analysis, the system can automatically assess the gap between the current collection process's effectiveness and historical best practices, thereby determining the optimal collection strategy. Through this process, the system can continuously optimize the sampling process, improve sampling efficiency and data quality, and further promote the widespread deployment and efficient operation of automated collection systems in practical applications.

[0130] Preferably, step S42 includes the following steps:

[0131] Step S421: Performing a synchronous evaluation of the robot arm sensor on the riverbed collection and return log to obtain riverbed sensor synchronous evaluation data;

[0132] Step S422: Analyze the sampler operation efficiency of the riverbed collection return log to obtain riverbed quality sampling efficiency data;

[0133] Step S423: Perform water flow error analysis on the riverbed collected and returned logs. If the flow rate is greater than 0.02 m / s, it is marked as riverbed environmental parameter data.

[0134] Step S424: calibrate the riverbed sensor synchronous evaluation data, riverbed quality sampling efficiency data, and riverbed environmental parameter data to obtain riverbed collection evaluation data.

[0135] In this embodiment of the present invention, sensor synchronization evaluation is performed on riverbed data collection and return logs to generate riverbed sensor synchronization evaluation data. The core of this process lies in analyzing the time synchronization of data from multiple sensors (such as pressure sensors, sonar sensors, and positioning sensors). Sensors typically acquire data at different frequencies, and in practical applications, sensors experience varying delays and errors. To ensure data validity and consistency, precise synchronization algorithms (such as Kalman filtering or least squares methods) are required to align the time series data from different sensors. This eliminates time deviations, improves data fusion accuracy, and ensures the reliability of subsequent analysis. Riverbed quality sampling efficiency data is generated by analyzing the sampling operation efficiency of riverbed data collection and return logs. This efficiency analysis focuses on key factors in the sampling process, including sampling speed, success rate, and coverage. Data analysis methods can include time series data analysis and statistical analysis. By comparing the actual execution of sampling tasks with expected targets, bottlenecks and efficiency issues in the sampling process can be assessed. For example, using the timestamps and sampling results in the data, the average time per sampling can be calculated to determine the efficiency of the sampling process and identify equipment or operational issues. By calibrating the riverbed sensor synchronous evaluation data and the riverbed quality sampling efficiency data, the riverbed collection evaluation data is obtained. At this stage, the data of steps S421 and S422 are combined and corrected using a data calibration method to eliminate errors or inconsistencies in the data. Typically, this calibration method can be based on regression analysis, weighted average or other statistical methods to produce a comprehensive evaluation result. This process not only helps to improve the accuracy of each data, but also provides a quantitative basis for further optimizing the sampling process, helping the system to make more precise adjustments based on the evaluation results.

[0136] It is particularly important to analyze the sampling efficiency of the riverbed collection return logs, including:

[0137] The single-point sampling time is extracted based on the riverbed collection return log to obtain the sediment-gravel layer time data;

[0138] The sampling success rate analysis is performed based on the riverbed collection return log to obtain the riverbed quality sampling success rate data, among which the sediment layer sampling success rate should be greater than 95%, and the gravel layer sampling success rate should be greater than 90%, otherwise resampling should be performed.

[0139] The efficiency data were merged based on the sediment-gravel layer time consumption data and the riverbed sampling success rate data to obtain the riverbed sampling efficiency data.

[0140] In this embodiment of the present invention, single-point sampling duration is extracted from the returned logs of riverbed sampling to obtain sediment and gravel layer duration data. Specifically, the core of sampling duration data extraction lies in analyzing the timestamps recorded in the logs to calculate the duration of each sampling operation. To accurately evaluate sampling performance in different strata (such as sediment and gravel), it is necessary to accurately extract the execution duration of each sampling operation based on the sampling start and end times recorded in the returned logs. This data, through precise time analysis (for example, using time difference calculation), provides the basis for subsequent evaluation. Next, sampling success rate analysis is performed to obtain riverbed sampling success rate data, specifically the sampling success rates for sediment and gravel layers. The sampling success rate analysis is performed by statistically analyzing the results of each sampling operation in the returned logs, including whether the sampling was successful or failed. The actual sampling data is compared against preset standards (e.g., a success rate of greater than 95% for sediment layers and greater than 90% for gravel layers) to assess whether the sampling meets the expected quality standards. This process typically involves counting successes and failures in the sampling operation and calculating the success rate. If the success rate is lower than the set standard, a resampling instruction is initiated according to the rules to ensure data quality and sampling accuracy. The technical core of this process is statistical analysis and success rate assessment to ensure that the sampling quality meets the standards. Finally, the efficiency data of the riverbed sampling efficiency is obtained by merging the time consumption data of the sediment-gravel layer and the success rate data of the riverbed sampling. In this link, the sampling time and sampling success rate of different strata are first weighted or combined in multiple dimensions to form a comprehensive sampling efficiency assessment. During the merging process, weighted averaging or multivariate regression analysis is usually used to standardize various indicators (such as time consumption, success rate, etc.) and comprehensively obtain the final sampling efficiency index. This data can provide a clear basis for subsequent optimization, so that the sampling operation can improve the performance of the entire riverbed collection process while ensuring high efficiency and success rate.

[0141] Of particular importance is the comparison of collection efficiency using historical riverbed collection data and riverbed collection assessment data, including:

[0142] Use historical riverbed collection data and riverbed collection assessment data to analyze positioning efficiency and obtain riverbed collection positioning efficiency data;

[0143] Energy efficiency analysis is performed using historical riverbed collection data and riverbed collection assessment data to obtain riverbed collection energy consumption data;

[0144] The riverbed collection positioning efficiency data and riverbed collection energy consumption data are subjected to comprehensive efficiency index weight analysis, and an automatic riverbed quality collection report is constructed.

[0145] In an embodiment of the present invention, historical riverbed collection data and riverbed collection assessment data are used to conduct positioning efficiency analysis to obtain riverbed collection positioning efficiency data. During this process, location information (such as Beidou RTK coordinates) from the historical riverbed collection data and information such as accuracy and response time from the collection assessment data are extracted and compared. Using methods such as time series analysis or spatial distribution analysis, positioning time consumption and spatial error are calculated. This data can be used to evaluate the positioning efficiency of the collection system in actual operation by calculating metrics such as positioning success rate, positioning accuracy, and positioning time. Technical methods used in this process include data mining and regression analysis to accurately measure and optimize the system's positioning capabilities. Energy efficiency analysis is conducted using historical riverbed collection data and riverbed collection assessment data to obtain riverbed collection energy consumption data. The key to this step lies in extracting and analyzing energy consumption information from the historical data, typically involving monitoring the energy consumption (e.g., electricity, fuel, etc.) of the collection equipment. By combining the collection data with the equipment's operating status, an energy consumption model is used to calculate the energy consumption of each collection activity. This process involves multivariate regression analysis or power-time-based energy consumption modeling to accurately assess the energy requirements of each collection operation. By comparing this data with factors such as collection efficiency and collection duration, specific energy efficiency indicators are derived. Finally, a comprehensive efficiency index weighting analysis is performed on the riverbed collection positioning efficiency data and riverbed collection energy consumption data to produce an automatic riverbed quality collection report. This step integrates positioning efficiency and energy consumption data by constructing a comprehensive efficiency index model. Common methods include weighted averaging, principal component analysis (PCA), or multivariate linear regression. These data are weighted to produce a comprehensive efficiency index that comprehensively reflects the balance between positioning accuracy and energy efficiency. Combining historical data with evaluation results, an automatic riverbed quality collection report is ultimately generated, providing guidance for subsequent optimization. Technical approaches at the data level include multidimensional data fusion, weighted analysis, and model building, ensuring a comprehensive assessment and optimization of the collection process efficiency.

[0146] In this specification, a riverbed quality automatic collection and control system based on Beidou navigation is provided, which is used to execute the above-mentioned riverbed quality automatic collection and control method based on Beidou navigation. The riverbed quality automatic collection and control system based on Beidou navigation includes:

[0147] The Beidou data processing module is used to plan sampling paths using riverbed sampling decisions and obtain riverbed path planning data. It generates collection instructions based on the riverbed path planning data and uses the riverbed robotic arm to perform real-time judgment and execution, ultimately generating a riverbed collection operation log.

[0148] The multi-source data fusion module is used to fuse multi-source data based on the Beidou multi-dimensional standard data set to obtain riverbed multi-source fusion data; and use the riverbed multi-source fusion data to generate riverbed sampling decisions;

[0149] The path planning and execution module is used to plan the sampling path using the riverbed sampling decision and obtain the riverbed path planning data; based on the riverbed path planning data, the sampling instructions are generated and executed by the riverbed robotic arm to obtain the riverbed sampling operation log;

[0150] The data return and evaluation module is used to use the riverbed collection operation log to return data packets and obtain the riverbed collection return log; perform performance indicator evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; and construct an automatic riverbed quality collection report based on the riverbed collection evaluation data.

[0151] The present invention is therefore intended to be illustrative and non-restrictive in all respects, with the scope of the invention being defined by the appended claims rather than the foregoing description, and all changes that come within the meaning and range of equivalents of the application documents are intended to be embraced therein.

[0152] The foregoing description is intended only to provide specific embodiments of the present invention, which will enable those skilled in the art to understand and implement the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown herein, but is to be construed in the widest possible manner consistent with the principles and novel features disclosed herein.

Claims

1. A riverbed quality automatic collection and control method based on Beidou navigation, characterized in that: The following steps are involved: Step S1: Obtain BeiDou satellite multidimensional dataset; Preprocess the raw data of the BeiDou satellite multidimensional dataset to obtain the BeiDou multidimensional standard dataset; Step S2: Perform multi-source data fusion based on the Beidou multi-dimensional standard data set to obtain riverbed multi-source fusion data; Generate riverbed sampling decisions using multi-source fusion data of the riverbed; Step S3: Using the riverbed sampling decision to plan the sampling path, obtain riverbed path planning data; based on the riverbed path planning data, generate a collection instruction, and use the riverbed robotic arm to perform real-time judgment and execution, and finally obtain a riverbed collection operation log; Step S4: using the riverbed collection operation log to return the data packet and obtain the riverbed collection return log; Perform performance index evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; An automatic riverbed quality collection report is constructed based on the riverbed collection and assessment data.

2. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Set the accuracy of the BeiDou RTK positioning sensor to ±1cm and the acquisition frequency to 10Hz; Step S12: Setting the detection depth of the high-frequency sonar sensor to 0.5-10m, the resolution to ±2cm, and the acquisition frequency to 5Hz; Step S13: Set the pressure sensor's range to 0-50 kPa, its accuracy to ±0.5% FS, and its acquisition frequency to 20 Hz; Step S14: using the Beidou RTK positioning sensor to collect riverbed latitude and longitude information; using the high-frequency sonar sensor to collect riverbed layer structure; using the pressure sensor to collect riverbed texture hardness data; Step S15: Packing the riverbed latitude and longitude information, riverbed layer structure, and riverbed texture hardness data to form a Beidou satellite multidimensional data set; Step S16: Preprocessing the raw data of the Beidou satellite multidimensional dataset to obtain a Beidou multidimensional standard dataset.

3. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 2 is characterized in that: Step S16 includes the following steps: Step S161: using statistical filtering to remove suspended object noise from the Beidou satellite multidimensional dataset to obtain Beidou satellite noise reduction data, wherein the statistical filtering area radius is 0.1m; Step S162: performing window filtering and jitter elimination processing on the Beidou satellite noise reduction data to obtain Beidou satellite jitter elimination data, wherein the window length of the window filtering and jitter elimination is 5s; Step S163: performing data set standardization on the Beidou satellite jitter elimination data to obtain a Beidou multi-dimensional standard data set.

4. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 1 is characterized in that: The multi-source data fusion described in step S2 includes: Extract Beidou RTK coordinates and sonar point cloud data based on Beidou multi-dimensional standard dataset; Construct global free coordinates based on BeiDou RTK coordinates and BeiDou multi-dimensional standard data sets to obtain carrier pose coordinate data; Obtain historical sedimentary data; The historical sedimentary layer data and sonar point cloud data were segmented into sediment-gravel layers using the preset judgment criteria, and a layered planar surface diagram was modeled to obtain a riverbed layered planar surface diagram. The preset judgment criteria were: a density less than or equal to 1.8 g / cm³ was considered a sediment layer; a density greater than or equal to 2.2 g / cm³ was considered a gravel layer. The carrier position coordinate data is used to mark the coordinates of the riverbed stratification surface map to generate multi-source fusion data of the riverbed.

5. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 1 is characterized in that: The generation of the riverbed sampling decision in step S2 includes: Obtain real-time riverbed acquisition data; The real-time riverbed collection data is detected in real time using riverbed multi-source fusion data to obtain riverbed sampling decisions, where the riverbed sampling decisions include sediment layer mode and gravel layer mode; if a sediment layer is detected, based on the sediment layer mode of the riverbed sampling decision, a bucket sampler is used to perform sampling operations with an opening diameter of 20 cm and a sampling volume of 500 ml / time to obtain sediment layer sampling data; if a gravel layer is detected, based on the gravel layer mode of the riverbed sampling decision, a cone sampler is used to perform sampling operations with a sampling depth of 10-50 cm to obtain gravel layer sampling data; if other layers are detected, sampling is stopped and unknown layer sampling data is obtained, thereby completing the riverbed sampling decision.

6. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 1 is characterized in that: The sampling path planning described in step S3 includes: Grid division is performed based on the sampling data of the sediment layer and the gravel layer to obtain an underground sampling grid map; Mark underground sampling points according to the underground sampling grid map to obtain sampling point marking data; Generate underground sampling paths based on sampling focus mark data; Perform shortest path optimization based on the underground sampling path to obtain the shortest sampling path optimization data; According to the multi-source fusion data of the riverbed, the path adjustment planning is carried out on the shortest sampling path optimization data to obtain the riverbed path planning data.

7. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 1 is characterized in that: Obtaining the riverbed collection operation log in step S3 includes: The riverbed manipulator performs a lowering operation at a speed of 0.1 m / s according to the riverbed path planning data, and uses a bucket sampler or a cone sampler for sampling. When the pressure sensor is greater than or equal to 5 kPa, it is determined to have touched the bottom, and a sampling instruction is generated; The bucket sampler's sampling opening and closing time was set to 2s, and the sampling depth was 0.3-0.5m, and the bucket sampling log data was obtained; The cone sampler speed was set to 10 rpm and the sampling depth was 0.5-1 m to obtain the cone sampler log data; The bucket sampling log data and the cone sampler log data are summarized into an operation log to obtain a riverbed sampling operation log.

8. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 1 is characterized in that: Step S4 includes the following steps: Step S41: Acquire historical riverbed collection data; use the riverbed collection operation log to return data packets to obtain a riverbed collection return log; Step S42: performing performance index evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; Step S43: Use historical riverbed collection data and riverbed collection assessment data to compare collection efficiency, and construct an automatic riverbed quality collection report.

9. The method for automatically collecting and controlling riverbed quality based on Beidou navigation according to claim 8 is characterized in that: Step S42 includes the following steps: Step S421: Performing a synchronous evaluation of the robot arm sensor on the riverbed collection and return log to obtain riverbed sensor synchronous evaluation data; Step S422: Analyze the sampler operation efficiency of the riverbed collection return log to obtain riverbed quality sampling efficiency data; Step S423: Perform water flow error analysis on the riverbed collected and returned logs. If the flow rate is greater than 0.02 m / s, it is marked as riverbed environmental parameter data. Step S424: calibrate the riverbed sensor synchronous evaluation data, riverbed quality sampling efficiency data, and riverbed environmental parameter data to obtain riverbed collection evaluation data.

10. A riverbed quality automatic collection and control system based on Beidou navigation, characterized in that: Used to execute the riverbed quality automatic collection and control method based on Beidou navigation as claimed in claim 1, the riverbed quality automatic collection and control system based on Beidou navigation comprises: Beidou data processing module, used to obtain Beidou satellite multidimensional data sets; pre-process the raw data of Beidou satellite multidimensional data sets to obtain Beidou multidimensional standard data sets; The multi-source data fusion module is used to fuse multi-source data based on the Beidou multi-dimensional standard data set to obtain riverbed multi-source fusion data; and use the riverbed multi-source fusion data to generate riverbed sampling decisions; The path planning and execution module is used to plan the sampling path using the riverbed sampling decision and obtain the riverbed path planning data; based on the riverbed path planning data, it generates the collection instructions and uses the riverbed robotic arm to perform real-time judgment and execution, ultimately obtaining the riverbed collection operation log; The data return and evaluation module is used to use the riverbed collection operation log to return data packets and obtain the riverbed collection return log; perform performance indicator evaluation based on the riverbed collection return log to obtain riverbed collection evaluation data; and construct an automatic riverbed quality collection report based on the riverbed collection evaluation data.

Images