Method and system for detecting filling quality of silt roadbed based on multi-source data

Abstract

The invention relates to the technical field of road engineering, in particular to a silt roadbed filling quality detection method and system based on multi-source data, by collecting multi-source data such as humidity data, compaction degree data, mechanical construction dynamic data and roadbed deformation data in real time, key characteristic variables are extracted by adopting a data fusion technology and a principal component analysis method, and the filling quality of a silt roadbed is detected. And on the basis of the characteristic variables, constructing a silt roadbed quality evaluation model to realize real-time prediction and evaluation of roadbed quality in the construction process. And if the detection result does not reach the target standard, the system dynamically adjusts the construction parameters through a self-adaptive control algorithm to optimize the construction process and ensure that the roadbed quality meets the preset requirement. In addition, the system displays real-time monitoring results and prediction data through a visual means, and the intelligent level of construction management is improved. The method can be widely applied to infrastructure construction of roads, railways and the like, and has high application prospects and popularization value.

Description

Technical Field

[0001] The present invention relates to the technical field of road engineering, and in particular to a method and system for detecting the quality of silt roadbed filling based on multi-source data. Background Art

[0002] With the continuous advancement of urbanization, silt roadbed, as a key supporting layer in the construction of highways, railways and other infrastructure, carries the stability and durability of the road. The quality of silt roadbed is directly related to the safety and service life of the road. Therefore, the filling quality monitoring of silt roadbed has become an important task in engineering construction.

[0003] At present, the existing silt roadbed quality detection technology has the following major problems: First, most traditional detection methods rely on manual sampling and single-point data collection, which cannot achieve comprehensive monitoring of the entire construction process and cannot reflect the dynamic changes in quality during the construction process in real time. Secondly, the existing technology often ignores the comprehensive use of multi-source data, resulting in the inability to accurately identify various factors affecting the roadbed quality, such as humidity, mechanical construction dynamics, roadbed deformation, etc., and cannot conduct a comprehensive quality assessment. Third, the existing quality assessment methods are mostly static assessments, lacking real-time and pertinence, and cannot timely discover potential quality problems in the construction process, resulting in slow quality control response and inability to accurately and efficiently adjust the construction. Finally, the existing technology lacks an effective dynamic optimization and real-time adjustment mechanism during the construction process. Once quality deviations occur during the construction process, manual intervention is often required, and automated and intelligent process control cannot be achieved. Summary of the invention

[0004] Based on the above purpose, the present invention provides a method and system for detecting the quality of silt roadbed filling based on multi-source data.

[0005] The silt roadbed filling quality detection method based on multi-source data includes the following steps:

[0006] S1, data collection and preprocessing: by deploying a variety of sensors and remote sensing equipment, multi-source data of silt roadbed is acquired in real time, the multi-source data includes moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data, and the collected multi-source data is preprocessed, the preprocessing includes noise filtering, data cleaning and format unification to form a preliminary data set;

[0007] S2, data fusion and feature extraction: Data fusion technology is used to integrate the pre-processed multi-source data to form a unified comprehensive data set, and the principal component analysis method is used to reduce the data dimension and extract key characteristic variables related to the roadbed filling quality. The key characteristic variables include moisture distribution, pressure distribution, deformation characteristics and construction parameters;

[0008] S3, quality assessment model establishment: Based on the extracted characteristic variables, a silt roadbed quality assessment model is constructed using a machine learning algorithm, and the silt roadbed quality assessment model is trained using historical multi-source data;

[0009] S4, quality detection and real-time prediction: input the multi-source data collected in real time into the trained silt roadbed quality assessment model to perform real-time detection and prediction of silt roadbed quality;

[0010] S5, construction process adjustment and feedback: Compare the prediction results with the target quality standards. If the prediction results do not meet the predetermined standards, use the adaptive control algorithm to adjust the construction process parameters in real time to achieve dynamic optimization and control of the roadbed quality;

[0011] S6, result visualization and decision support: Display the multi-source data collected in real time and the prediction results through visualization.

[0012] Optionally, the S1 includes:

[0013] S11, sensor deployment and data collection: deploy various sensor devices in the silt roadbed construction area, including humidity sensors, compaction sensors, deformation sensors, and mechanical construction dynamic data collection equipment;

[0014] The humidity sensor is used to monitor the moisture distribution of the soil in real time and obtain humidity data;

[0015] The compaction sensor is used to detect the compaction degree of the roadbed and obtain compaction data;

[0016] The deformation sensor is used to capture the roadbed deformation data during the construction process;

[0017] Mechanical construction dynamic data acquisition equipment is used to record mechanical construction dynamic data during the construction process;

[0018] S12, remote sensing equipment and satellite data acquisition: deploy remote sensing equipment (such as drones, remote sensing satellites, etc.) to monitor the silt roadbed area and obtain soil surface deformation data and spatial distribution information in real time;

[0019] S13, Mechanical construction dynamic data collection: By deploying intelligent monitoring equipment, the working status of construction machinery is monitored in real time. The collected mechanical construction dynamic data includes mechanical working frequency, operation trajectory, vibration intensity and driving speed;

[0020] S14, data preprocessing and standardization: preprocessing the collected multi-source data, including noise filtering, data cleaning and format standardization;

[0021] S15, data quality assessment: perform quality assessment on the preprocessed data set to test the validity and accuracy of the data;

[0022] S16, preliminary data set formation: Integrate the quality-assessed data to generate a preliminary data set.

[0023] Optionally, S2 includes:

[0024] S21, data fusion and integration: The pre-processed multi-source data are fused using the weighted average method to form a unified comprehensive data set;

[0025] S22, principal component analysis dimensionality reduction: principal component analysis is performed on the integrated data set formed after fusion to reduce the data dimension;

[0026] S23, feature variable extraction and screening: Based on the comprehensive data set after dimensionality reduction by principal component analysis, the Pearson correlation coefficient was used to analyze the correlation between different features and extract the key feature variables related to the roadbed filling quality.

[0027] Optionally, the S3 includes:

[0028] S31, model construction: Based on the extracted characteristic variables, a silt roadbed quality assessment model is constructed using the support vector machine algorithm;

[0029] S32, feature selection and L1 regularization: L1 regularization method is used to select feature variables.

[0030] Optionally, S3 further includes:

[0031] S33, support vector machine kernel function selection: In the constructed silt roadbed quality assessment model, the Gaussian radial basis kernel is selected as the kernel function to adapt to the nonlinear relationship in the data;

[0032] S34, kernel function parameter optimization: grid search and cross-validation methods are used to optimize the hyperparameters of the Gaussian radial basis kernel;

[0033] S35, model training and verification: After determining the optimal kernel function and parameters, the optimized silt roadbed quality assessment model is used to train and predict the silt roadbed quality.

[0034] Optionally, the S4 includes:

[0035] S41, multi-source data input and quality assessment: the collected and pre-processed multi-source data, including moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data, are input into the silt roadbed quality assessment model in real time;

[0036] S42, real-time quality prediction output: The silt roadbed quality assessment model processes and predicts based on real-time input multi-source data and outputs quality prediction results.

[0037] Optionally, S4 further includes:

[0038] S71, quality inspection result evaluation: based on the prediction result generated in S42, the predicted quality data is compared with the preset target quality standard to evaluate the quality status of the current silt roadbed. If the prediction result does not meet the preset standard, appropriate adjustment measures are taken;

[0039] S72, feedback and optimization of adjustment plans: Based on the evaluation results, if it is found that the quality does not meet the standards, the system will output adjustment suggestions to help construction personnel optimize construction parameters and adjust construction strategies to ensure that the quality of silt roadbed filling meets the standards.

[0040] Optionally, the S5 includes:

[0041] S51, quality prediction and target standard comparison: compare the quality prediction result with the predetermined target quality standard, the target quality standard includes the standard values ​​of compaction degree, moisture, and roadbed deformation quality indicators, and calculate the error value;

[0042] S52, adaptively adjusting construction parameters: if the error value E exceeds the set error threshold, indicating that the roadbed quality does not meet the target standard, the adaptive control algorithm is started to dynamically adjust the construction parameters during the construction process;

[0043] S53, construction monitoring: After adjusting the construction parameters, monitor the adjusted construction process to ensure that all parameters in the construction process always remain within the range that meets the quality standards.

[0044] Optionally, the S6 includes:

[0045] Visualization display: Through data visualization technology, the collected multi-source data and the quality prediction results of the silt roadbed quality assessment model are displayed. The visualization methods include time series graphs, spatial distribution graphs and quality warning graphs.

[0046] The silt roadbed filling quality detection system based on multi-source data is used to implement the silt roadbed filling quality detection method based on multi-source data, and includes the following modules:

[0047] Data acquisition module: by deploying a variety of sensors and remote sensing equipment, multi-source data related to the quality of silt roadbed filling is collected in real time, including moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data;

[0048] Data preprocessing module: preprocesses the collected multi-source data, including noise filtering, data cleaning and format unification, to generate a preliminary data set;

[0049] Data fusion and feature extraction module: Data fusion technology is used to fuse the pre-processed multi-source data to form a unified comprehensive data set, and the principal component analysis method is used to reduce the dimension of the data to extract key characteristic variables related to the roadbed quality, including moisture distribution, pressure distribution, deformation characteristics and construction parameters;

[0050] Quality assessment model building module: Based on the extracted characteristic variables, a silt roadbed quality assessment model is built using a machine learning algorithm, and the silt roadbed quality assessment model is trained using historical multi-source data;

[0051] Real-time quality detection and prediction module: inputs the multi-source data collected in real time into the trained silt roadbed quality assessment model to perform real-time detection and prediction of silt roadbed quality;

[0052] Construction adjustment and feedback module: Compare the prediction results with the target quality standards. If the prediction results do not meet the predetermined standards, the adaptive control algorithm is used to adjust the construction process parameters in real time to achieve dynamic optimization and control of the roadbed quality.

[0053] Visualization module: Display the multi-source data and prediction results collected in real time through visualization means.

[0054] Beneficial effects of the present invention:

[0055] The present invention can obtain more comprehensive and accurate roadbed quality monitoring data, including key parameters such as humidity, compaction, mechanical construction dynamics, and roadbed deformation, through real-time collection and preprocessing of multi-source data. The application of data fusion and principal component analysis technology ensures that the most representative characteristic variables are extracted from multiple data sources. These characteristic variables are more closely related to the roadbed quality, thereby greatly improving the accuracy of quality assessment. Based on this high-precision real-time data input, the constructed silt roadbed quality assessment model can provide real-time quality prediction during the construction process, discover potential problems in a timely manner, and ensure that the construction quality meets the predetermined standards.

[0056] The present invention can evaluate the construction quality in real time and give timely feedback by comparing the real-time quality prediction with the target standard. If the test result does not meet the target standard, the system adjusts the construction parameters (such as mechanical operation parameters, construction process, etc.) in real time through an adaptive control algorithm to ensure continuous optimization of the construction process. This adjustment mechanism allows the construction process to not only dynamically adapt to environmental changes, but also to minimize rework and waste of resources caused by quality problems, thereby improving construction efficiency and the stability of construction quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0058] Figure 1 A schematic diagram of a method flow of an embodiment of the present invention;

[0059] Figure 2 Schematic diagram of a system flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0060] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0061] It should be noted that the references to "one embodiment", "an embodiment", "an exemplary embodiment", "some embodiments" and the like in the specification indicate that the embodiments described may include specific features, structures or characteristics, but not every embodiment may include the specific features, structures or characteristics. In addition, when a specific feature, structure or characteristic is described in conjunction with an embodiment, it should be within the knowledge of a person skilled in the art to implement such feature, structure or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0062] In general, a term can be understood, at least in part, from its use in context. For example, depending, at least in part, on the context, the term "one or more" as used herein can be used to describe any feature, structure, or characteristic in the singular sense, or can be used to describe a combination of features, structures, or characteristics in the plural sense. Additionally, the term "based on" can be understood as not necessarily intended to convey an exclusive set of factors, but can instead, depending, at least in part, on the context, allow for the presence of other factors that are not necessarily explicitly described.

[0063] like Figure 1 As shown, the silt roadbed filling quality detection method based on multi-source data includes the following steps:

[0064] S1, data collection and preprocessing: by deploying a variety of sensors and remote sensing equipment, multi-source data of silt roadbed is acquired in real time. The multi-source data includes moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data. The collected multi-source data is preprocessed, including noise filtering, data cleaning and format unification, to form a preliminary data set;

[0065] S2, data fusion and feature extraction: Data fusion technology is used to integrate the pre-processed multi-source data to form a unified comprehensive data set, and the principal component analysis method is used to reduce the data dimension and extract key characteristic variables related to the roadbed filling quality. The key characteristic variables include moisture distribution, pressure distribution, deformation characteristics and construction parameters;

[0066] S3, quality assessment model establishment: Based on the extracted characteristic variables, a silt roadbed quality assessment model is constructed using a machine learning algorithm, and the silt roadbed quality assessment model is trained using historical multi-source data;

[0067] S4, quality detection and real-time prediction: input the multi-source data collected in real time into the trained silt roadbed quality assessment model to perform real-time detection and prediction of silt roadbed quality;

[0068] S5, construction process adjustment and feedback: Compare the prediction results with the target quality standards. If the prediction results do not meet the predetermined standards, use the adaptive control algorithm to adjust the construction process parameters in real time to achieve dynamic optimization and control of the roadbed quality;

[0069] S6, result visualization and decision support: Display the multi-source data collected in real time and the prediction results through visualization.

[0070] S1 includes:

[0071] S11, sensor deployment and data collection: deploy various sensor devices in the silt roadbed construction area, including humidity sensors, compaction sensors, deformation sensors, and mechanical construction dynamic data collection equipment;

[0072] The humidity sensor is used to monitor the moisture distribution of the soil in real time and obtain humidity data;

[0073] The compaction sensor is used to detect the compaction degree of the roadbed and obtain compaction data;

[0074] The deformation sensor is used to capture the roadbed deformation data during the construction process;

[0075] Mechanical construction dynamic data acquisition equipment is used to record mechanical construction dynamic data during the construction process;

[0076] S12, Remote sensing equipment and satellite data acquisition: Deploy remote sensing equipment (such as drones, remote sensing satellites, etc.) to monitor the silt roadbed area and obtain soil surface deformation data and spatial distribution information in real time. The image data obtained by remote sensing equipment can provide high-resolution changes in roadbed surface morphology, assist in evaluating quality changes during construction, especially in areas that cannot be reached. At the same time, deformation monitoring using remote sensing technology can provide spatial distribution information related to soil moisture changes, helping to accurately identify potential quality problems;

[0077] S13, Mechanical construction dynamic data collection: By deploying intelligent monitoring equipment, the working status of construction machinery is monitored in real time. The collected mechanical construction dynamic data includes mechanical working frequency, operation trajectory, vibration intensity and driving speed;

[0078] S14, data preprocessing and standardization: preprocess the collected multi-source data, including noise filtering, data cleaning and format standardization;

[0079] Use mean filling or interpolation algorithms to process missing data, and use median filtering to remove noise data. Perform unit conversion and standardization on all sensor data so that data from different sensors can be compared and analyzed under a unified standard, thus laying the foundation for subsequent multi-source data fusion and feature extraction.

[0080] S15, data quality assessment: perform quality assessment on the preprocessed data set to test the validity and accuracy of the data and ensure that the data meets the predetermined quality standards. The standards for data quality assessment include data integrity, accuracy, timeliness and consistency. By setting thresholds or using statistical methods such as skewness and kurtosis analysis, abnormal data are screened and eliminated to ensure that the final input data set is a reliable data set.

[0081] S16, preliminary data set formation: the data after quality assessment are integrated to generate a preliminary data set, which includes multi-dimensional data such as moisture, compaction, deformation, and mechanical construction dynamic data, as input for subsequent data analysis and quality assessment models;

[0082] Through the deployment of sensors, remote sensing equipment collection and dynamic data monitoring of construction machinery, multi-source data in the silt roadbed filling process was acquired in real time. After preprocessing and standardization, a high-quality preliminary data set was generated, which provided a reliable data basis for subsequent data fusion, feature extraction and quality assessment, thereby ensuring the accuracy and effectiveness of quality inspection.

[0083] S2 includes:

[0084] S21, data fusion and integration: The pre-processed multi-source data are fused using the weighted average method to form a unified comprehensive data set. The multi-source data include moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data;

[0085] The weighted average method is used to fuse these data to achieve the unification of data from different sources and information complementarity. Specifically, for each data type, different weights are assigned according to its reliability and importance, and the weighted average is calculated to generate a unified comprehensive data set. The formula of the weighted average method is as follows:

[0086]

[0087] Among them, X fused is the fused data, X i is the i-th type of data, w i is the weight of the i-th type of data, and n is the number of data types;

[0088] S22, principal component analysis dimensionality reduction: principal component analysis is performed on the integrated data set formed after fusion to reduce the data dimension. Principal component analysis calculates the data covariance matrix and decomposes its eigenvalues ​​to extract the principal components in the data, thereby reducing the data dimension, removing redundant information, and retaining the maximum amount of information. Through this step, the generated reduced dimension data set contains the most important characteristic variables, reducing the computational complexity and improving the efficiency of subsequent analysis. The specific dimensionality reduction formula is as follows:

[0089] Z = PCA(X);

[0090] Among them, X is the fused multi-source data set, Z is the data set after dimensionality reduction, each principal component of principal component analysis represents the main variability in the data, and the main information of the data is retained by selecting the first n principal components;

[0091] S23, feature variable extraction and screening: Based on the comprehensive data set after dimensionality reduction by principal component analysis, the Pearson correlation coefficient was used to analyze the correlation between different features and extract the key feature variables related to the roadbed filling quality;

[0092] Specifically, the Pearson correlation coefficient between each dimension-reduced feature and the roadbed quality standard (such as compaction, deformation, etc.) is calculated, and the feature variables with strong correlation are selected. The Pearson correlation coefficient formula is as follows:

[0093]

[0094] Among them, X i and Y i are the observed values ​​of two variables, and is the mean of the two variables, r is the Pearson correlation coefficient, which indicates the linear correlation between the two variables;

[0095] By fusing multi-source data using the weighted average method and combining data denoising with principal component analysis dimensionality reduction, characteristic variables closely related to the roadbed filling quality can be effectively extracted and screened out. This process reduces data redundancy, improves data processing efficiency, and provides more accurate and useful input for subsequent quality assessment.

[0096] S3 includes:

[0097] S31, model construction: Based on the extracted characteristic variables, a silt roadbed quality assessment model is constructed through the support vector machine algorithm. The goal of the model is to maximize the classification interval to improve the accuracy of the prediction. Its optimization goal is expressed as:

[0098]

[0099] Among them, x i is the input feature, y i is the label, θ is the parameter of the decision boundary, C is the penalty factor, b is the bias term, and m is the number of samples;

[0100] S32, feature selection and L1 regularization: The L1 regularization method is used to select feature variables. L1 regularization introduces penalty terms to make the coefficients of some features tend to zero, thereby effectively selecting features and removing features that are not related to the quality assessment of silt roadbed. This method helps to improve the sparsity of the model and reduce the risk of overfitting. The objective function of L1 regularization is:

[0101]

[0102] Among them, h θ (x (i) ) is the predicted value, y (i) is the true value, θ j is the parameter of the model, λ is the regularization parameter, m is the number of samples, and n is the number of features;

[0103] By constructing a silt roadbed quality assessment model based on the support vector machine algorithm and combining it with the L1 regularization method for feature selection, the features most relevant to the silt roadbed quality assessment can be effectively screened out, ensuring that the model has good generalization ability and prediction accuracy.

[0104] S3 also includes:

[0105] S33, support vector machine kernel function selection: In the constructed silt roadbed quality assessment model, the Gaussian radial basis kernel is selected as the kernel function to adapt to the nonlinear relationship in the data. The Gaussian radial basis kernel can map the input data to a higher-dimensional feature space, making the data easier to segment in the space. The calculation formula of the Gaussian radial basis kernel is expressed as:

[0106]

[0107] Among them, x and x ′ is the input data point, σ is the parameter of the kernel function, which controls the smoothness of the mapping to the feature space. By selecting the appropriate σ parameter, the complexity of the model and the degree of fit to the training data can be adjusted;

[0108] S34, kernel function parameter optimization: The hyperparameter σ of the Gaussian radial basis kernel is optimized by using grid search and cross validation methods. Specifically, the kernel function parameters that are most suitable for the current silt roadbed quality inspection task are found by conducting experiments within multiple predetermined parameter ranges. During the optimization process, 5-fold cross validation is used to evaluate the performance of each set of parameters, and the σ value that minimizes the model prediction error is selected;

[0109] S35, model training and verification: after determining the optimal kernel function and parameters, the optimized silt roadbed quality assessment model is used to train and predict the silt roadbed quality. The training data includes historical construction data, real-time monitoring data and other relevant quality characteristics. After model training, accurate quality prediction values ​​can be output. The trained model is verified by indicators such as mean square error and root mean square error to ensure the prediction ability and stability of the model.

[0110] By selecting the Gaussian radial basis kernel and optimizing its parameters, the nonlinear relationship of the data can be better handled and the prediction accuracy of the model for the quality of silt roadbed can be improved. In addition, by optimizing the kernel function parameters through cross-validation and grid search, the generalization ability and accuracy of the silt roadbed quality assessment model can be further improved, thereby achieving efficient assessment of the quality of silt roadbed.

[0111] S4 includes:

[0112] S41, multi-source data input and quality assessment: the collected and pre-processed multi-source data, including moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data, are input into the silt roadbed quality assessment model in real time;

[0113] S42, real-time quality prediction output: The silt roadbed quality assessment model processes and predicts based on real-time input multi-source data and outputs quality prediction results. The prediction content includes:

[0114] Evaluate the impact of moisture data on roadbed compaction;

[0115] Predict whether the compaction degree reaches the target standard;

[0116] Predict whether the roadbed deformation exceeds the set safety threshold;

[0117] Predict potential quality risks during the construction process based on mechanical construction dynamic data.

[0118] S4 also includes:

[0119] S71, quality inspection result evaluation: Based on the prediction result generated in S42, the predicted quality data is compared with the preset target quality standard to evaluate the quality status of the current silt roadbed. If the prediction result does not meet the preset standard, appropriate adjustment measures are taken. The evaluation steps include:

[0120] Evaluate the impact of multiple sources of data such as moisture, compaction and deformation on roadbed quality;

[0121] Evaluate the contribution of mechanical construction dynamic data to quality prediction results;

[0122] S72, feedback and optimization of adjustment plans: Based on the evaluation results, if it is found that the quality does not meet the standards, the system will output adjustment suggestions to help construction personnel optimize construction parameters and adjust construction strategies to ensure that the quality of silt roadbed filling meets the standards.

[0123] S5 includes:

[0124] S51, quality prediction and target standard comparison: compare the quality prediction results with the predetermined target quality standards, which include the standard values ​​of compaction, moisture, and roadbed deformation quality indicators, and calculate the error value, which is expressed as:

[0125] E=|Q pred -Q target |;

[0126] Among them, E is the error value, Q pred is the predicted quality parameter value, Q target is the target standard value. By calculating the error value E, it can be judged whether the current roadbed quality meets the predetermined requirements;

[0127] S52, adaptively adjusting construction parameters: If the error value E exceeds the set error threshold, it means that the roadbed quality does not meet the target standard, then the adaptive control algorithm is started to dynamically adjust the construction parameters during the construction process. The specific adjusted construction parameters include:

[0128] Working parameters of construction machinery (such as vibration frequency, amplitude, etc.);

[0129] Construction technology (such as compaction method, layered filling thickness, etc.);

[0130] Construction environment control (such as humidity, temperature, etc.);

[0131] The adaptive control algorithm uses a real-time feedback mechanism to adjust the construction process according to the prediction error to ensure that the roadbed quality is continuously optimized during the construction process. The adjustment process can be achieved through a machine learning model (such as reinforcement learning), which is specifically controlled by the following optimization objective function:

[0132]

[0133] Among them, E i is the error value for each prediction indicator, n is the number of selected quality parameters;

[0134] S53, construction monitoring: After adjusting the construction parameters, monitor the adjusted construction process to ensure that all parameters in the construction process always remain within the range that meets the quality standards.

[0135] S6 includes:

[0136] Visualization: Through data visualization technology, the collected multi-source data and the quality prediction results of the silt roadbed quality assessment model are displayed. The visualization methods include time series diagrams, spatial distribution diagrams and quality warning diagrams, among which:

[0137] Time series graph: displays the trend of real-time collected data such as humidity, compaction, deformation, etc. over time, and superimposes the prediction results to help construction personnel understand the evolution of quality during the construction process;

[0138] Spatial distribution map: displays the spatial distribution of quality data such as humidity, compaction, and roadbed deformation in each construction area, and identifies areas with potential quality risks in combination with quality prediction results;

[0139] Quality warning diagram: When anomalies are found in multi-source data or prediction results, quality warnings are triggered in real time, and abnormal areas are highlighted graphically to help construction personnel identify and adjust the construction process in a timely manner.

[0140] like Figure 2 As shown, the silt roadbed filling quality detection system based on multi-source data is used to implement the silt roadbed filling quality detection method based on multi-source data, and includes the following modules:

[0141] Data acquisition module: By deploying a variety of sensors and remote sensing equipment, multi-source data related to the quality of silt roadbed filling can be collected in real time. The multi-source data includes moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data;

[0142] Data preprocessing module: preprocesses the collected multi-source data, including noise filtering, data cleaning and format unification, to generate a preliminary data set;

[0143] Data fusion and feature extraction module: Data fusion technology is used to fuse the pre-processed multi-source data to form a unified comprehensive data set, and the principal component analysis method is used to reduce the dimension of the data to extract key characteristic variables related to the roadbed quality, including moisture distribution, pressure distribution, deformation characteristics and construction parameters;

[0144] Quality assessment model building module: Based on the extracted characteristic variables, a silt roadbed quality assessment model is built using a machine learning algorithm, and the silt roadbed quality assessment model is trained using historical multi-source data;

[0145] Real-time quality detection and prediction module: inputs the multi-source data collected in real time into the trained silt roadbed quality assessment model to perform real-time detection and prediction of silt roadbed quality;

[0146] Construction adjustment and feedback module: Compare the prediction results with the target quality standards. If the prediction results do not meet the predetermined standards, the adaptive control algorithm is used to adjust the construction process parameters in real time to achieve dynamic optimization and control of the roadbed quality.

[0147] Visualization module: Display the multi-source data and prediction results collected in real time through visualization means.

[0148] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0149] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A silt roadbed filling quality detection method based on multi-source data, characterized in that: The following steps are involved: S1, data collection and preprocessing: by deploying a variety of sensors and remote sensing equipment, multi-source data of silt roadbed is acquired in real time, the multi-source data includes moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data, and the collected multi-source data is preprocessed, the preprocessing includes noise filtering, data cleaning and format unification to form a preliminary data set; S2, data fusion and feature extraction: Data fusion technology is used to integrate the pre-processed multi-source data to form a unified comprehensive data set, and the principal component analysis method is used to reduce the data dimension and extract key characteristic variables related to the roadbed filling quality. The key characteristic variables include moisture distribution, pressure distribution, deformation characteristics and construction parameters; S3, quality assessment model establishment: Based on the extracted characteristic variables, a silt roadbed quality assessment model is constructed using a machine learning algorithm, and the silt roadbed quality assessment model is trained using historical multi-source data; S4, quality detection and real-time prediction: input the multi-source data collected in real time into the trained silt roadbed quality assessment model to perform real-time detection and prediction of silt roadbed quality; S5, construction process adjustment and feedback: Compare the prediction results with the target quality standards. If the prediction results do not meet the predetermined standards, use the adaptive control algorithm to adjust the construction process parameters in real time to achieve dynamic optimization and control of the roadbed quality; S6, result visualization and decision support: Display the multi-source data collected in real time and the prediction results through visualization.

2. The silt roadbed filling quality detection method based on multi-source data according to claim 1 is characterized in that: The S1 includes: S11, sensor deployment and data collection: deploy various sensor devices in the silt roadbed construction area, including humidity sensors, compaction sensors, deformation sensors, and mechanical construction dynamic data collection equipment; The humidity sensor is used to monitor the moisture distribution of the soil in real time and obtain humidity data; The compaction sensor is used to detect the compaction degree of the roadbed and obtain compaction data; The deformation sensor is used to capture the roadbed deformation data during the construction process; Mechanical construction dynamic data acquisition equipment is used to record mechanical construction dynamic data during the construction process; S12, remote sensing equipment and satellite data acquisition: deploy remote sensing equipment to monitor the silt roadbed area and obtain soil surface deformation data and spatial distribution information in real time; S13, Mechanical construction dynamic data collection: By deploying intelligent monitoring equipment, the working status of construction machinery is monitored in real time. The collected mechanical construction dynamic data includes mechanical working frequency, operation trajectory, vibration intensity and driving speed; S14, data preprocessing and standardization: preprocessing the collected multi-source data, including noise filtering, data cleaning and format standardization; S15, data quality assessment: perform quality assessment on the preprocessed data set to test the validity and accuracy of the data; S16, preliminary data set formation: Integrate the quality-assessed data to generate a preliminary data set.

3. The silt roadbed filling quality detection method based on multi-source data according to claim 2 is characterized in that: The S2 includes: S21, data fusion and integration: The pre-processed multi-source data are fused using the weighted average method to form a unified comprehensive data set; S22, principal component analysis dimensionality reduction: principal component analysis is performed on the integrated data set formed after fusion to reduce the data dimension; S23, feature variable extraction and screening: Based on the comprehensive data set after dimensionality reduction by principal component analysis, the Pearson correlation coefficient was used to analyze the correlation between different features and extract the key feature variables related to the roadbed filling quality.

4. The silt roadbed filling quality detection method based on multi-source data according to claim 3 is characterized in that: The S3 includes: S31, model construction: Based on the extracted characteristic variables, a silt roadbed quality assessment model is constructed using the support vector machine algorithm; S32, feature selection and L1 regularization: L1 regularization method is used to select feature variables.

5. The silt roadbed filling quality detection method based on multi-source data according to claim 4 is characterized in that: The S3 further includes: S33, support vector machine kernel function selection: In the constructed silt roadbed quality assessment model, the Gaussian radial basis kernel is selected as the kernel function to adapt to the nonlinear relationship in the data; S34, kernel function parameter optimization: grid search and cross-validation methods are used to optimize the hyperparameters of the Gaussian radial basis kernel; S35, model training and verification: After determining the optimal kernel function and parameters, the optimized silt roadbed quality assessment model is used to train and predict the silt roadbed quality.

6. The silt roadbed filling quality detection method based on multi-source data according to claim 5 is characterized in that: The S4 includes: S41, multi-source data input and quality assessment: input the collected and pre-processed multi-source data into the silt roadbed quality assessment model in real time; S42, real-time quality prediction output: The silt roadbed quality assessment model processes and predicts based on real-time input multi-source data and outputs quality prediction results.

7. The silt roadbed filling quality detection method based on multi-source data according to claim 6 is characterized in that: The S4 further comprises: S71, quality inspection result evaluation: based on the prediction result generated in S42, the predicted quality data is compared with the preset target quality standard to evaluate the quality status of the current silt roadbed. If the prediction result does not meet the preset standard, adjustment measures are taken; S72, feedback and optimization of adjustment plans: Based on the evaluation results, if it is found that the quality does not meet the standards, the system will output adjustment suggestions to help construction personnel optimize construction parameters and adjust construction strategies.

8. The silt roadbed filling quality detection method based on multi-source data according to claim 7 is characterized in that: The S5 includes: S51, comparing quality prediction with target standards: comparing the quality prediction result with the predetermined target quality standard and calculating the error value; S52, adaptively adjusting construction parameters: if the error value E exceeds the set error threshold, indicating that the roadbed quality does not meet the target standard, the adaptive control algorithm is started to dynamically adjust the construction parameters during the construction process; S53, construction monitoring: after adjusting the construction parameters, monitor the adjusted construction process.

9. The silt roadbed filling quality detection method based on multi-source data according to claim 8 is characterized in that: The S6 includes: Visualization display: Through data visualization technology, the collected multi-source data and the quality prediction results of the silt roadbed quality assessment model are displayed. The visualization methods include time series graphs, spatial distribution graphs and quality warning graphs.

10. A silt roadbed filling quality detection system based on multi-source data, used to implement the silt roadbed filling quality detection method based on multi-source data as claimed in any one of claims 1 to 9, characterized in that: Includes the following modules: Data acquisition module: by deploying a variety of sensors and remote sensing equipment, multi-source data related to the quality of silt roadbed filling is collected in real time, including moisture data, compaction data, mechanical construction dynamic data and roadbed deformation data; Data preprocessing module: preprocesses the collected multi-source data, including noise filtering, data cleaning and format unification, to generate a preliminary data set; Data fusion and feature extraction module: Data fusion technology is used to fuse the pre-processed multi-source data to form a unified comprehensive data set, and the principal component analysis method is used to reduce the dimension of the data to extract key characteristic variables related to the roadbed quality, including moisture distribution, pressure distribution, deformation characteristics and construction parameters; Quality assessment model building module: Based on the extracted characteristic variables, a silt roadbed quality assessment model is built using a machine learning algorithm, and the silt roadbed quality assessment model is trained using historical multi-source data; Real-time quality detection and prediction module: inputs the multi-source data collected in real time into the trained silt roadbed quality assessment model to perform real-time detection and prediction of silt roadbed quality; Construction adjustment and feedback module: Compare the prediction results with the target quality standards. If the prediction results do not meet the predetermined standards, the adaptive control algorithm is used to adjust the construction process parameters in real time to achieve dynamic optimization and control of the roadbed quality. Visualization module: Display the multi-source data and prediction results collected in real time through visualization means.

Images