Ground penetrating radar underground anomaly risk assessment system and method embedded with industry standards

Abstract

The application provides a ground penetrating radar underground anomaly risk assessment system and method embedded with industry standards, which comprises a compliance management hub module, a data acquisition and pre-check module, a data processing and anomaly identification module, and a risk assessment and result generation module. The compliance management hub module converts industry standards into acquisition parameter compliance thresholds, anomaly identification rules, risk classification rules, and result output rules. The data acquisition and pre-check module obtains ground penetrating radar data and its acquisition parameters, and performs pre-checking by calling the acquisition parameter compliance thresholds. The data processing and anomaly identification module calls the anomaly identification rules, performs standardization processing and anomaly identification, and outputs anomaly characteristics and standard basis. The risk assessment and result generation module calls the risk classification rules and result output rules, and completes risk assessment and result output based on anomaly characteristics and engineering environment data. The system realizes automatic, standardized, and objective assessment with industry standards as the criterion.

Description

Technical Field

[0001] This invention belongs to the field of ground-penetrating radar detection and intelligent risk assessment technology, and specifically relates to a ground-penetrating radar underground anomaly risk assessment system and method with embedded industry standards. Background Technology

[0002] With the rapid increase in demand for underground space development and infrastructure operation and maintenance, ground-penetrating radar (GPR), with its advantages of being non-destructive, efficient, and high-precision, has become a core technology for detecting hidden anomalies such as underground cavities, loose bodies, and voids. Risk assessment of underground anomalies is a core basis for infrastructure safety management, emergency response, and disease control. However, existing GPR risk assessment systems still have the following problems:

[0003] The system is out of touch with industry standards, and the entire process lacks compliance control. Current national and industry standards are only used as reference texts for manual post-event verification. From setting detection operation parameters, signal processing, anomaly interpretation to risk classification, everything relies on the personal experience of operators, which easily leads to problems that do not conform to the standards. As a result, the output does not meet the requirements for project approval and cannot be directly used for governance design and emergency management, which seriously affects the efficiency of the project.

[0004] The lack of standardized risk assessment criteria results in a lack of authority. Existing systems often employ custom assessment rules, failing to strictly adhere to the evaluation indicators and judgment criteria clearly defined in industry standards. This leads to significant discrepancies in the assessment results of the same anomaly across different systems, hindering the formation of standardized control procedures and increasing the risk of safety incidents or over-management due to incorrect risk classification.

[0005] Existing technologies mostly focus on data processing algorithms or single-scenario evaluation, and none of them have achieved full-process embedded control of industry standards, nor have they formed a complete closed loop from data verification to output. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a ground-penetrating radar underground anomaly risk assessment system and method with embedded industry standards, achieving automated, standardized, and objective assessment based on industry standards.

[0007] The technical solution of the present invention is as follows:

[0008] An industry-standard ground-penetrating radar (GPR) system for assessing underground anomalies includes:

[0009] The compliance control central module is used to store and structurally decompose existing effective industry standards, and transform the standard clauses into a set of quantitative rules and standardized templates that can be called by the system. The set of quantitative rules includes at least compliance thresholds for collected parameters, anomaly detection rules, risk classification rules, and output rules.

[0010] The data acquisition and pre-verification module is used to acquire ground-penetrating radar acquisition parameters, ground-penetrating radar data and engineering environment data, and call the acquisition parameter compliance threshold in the compliance control center module to perform compliance verification on the ground-penetrating radar acquisition parameters before the ground-penetrating radar data enters the processing flow, and only outputs the ground-penetrating radar data that passes the verification to the data processing flow.

[0011] The data processing and anomaly identification module is used to call the anomaly discrimination rules in the compliance control center module, perform standardized processing and automatic anomaly identification on the verified ground penetrating radar data, and output anomaly feature information and its corresponding standard discrimination basis;

[0012] The risk assessment and result generation module is used to call the risk classification rules and result output rules in the compliance control center module, and perform risk level assessment and generate result output that meets the specifications based on the abnormal feature information and engineering environment data.

[0013] The data acquisition and pre-verification module, the data processing and anomaly identification module, and the risk assessment and result generation module are all connected to the compliance control center module, forming a closed-loop processing architecture in which the compliance control center module serves as the unified rule source, and each module automatically performs corresponding processing and assessment operations according to the rules it calls.

[0014] Furthermore, the compliance control central module includes multiple standard sub-libraries, which are divided according to engineering industry categories. Each standard sub-library stores the corresponding industry's collection parameter compliance thresholds, anomaly judgment rules, risk classification rules, and output rules in a structured manner.

[0015] The system automatically calls the corresponding standard sub-library as the unified rule source for the entire process, based on the selected target industry.

[0016] Furthermore, the engineering industry category includes at least one of municipal roads, highways, railways, airport runways, and water conservancy embankments;

[0017] The underground anomaly types defined in the anomaly discrimination rules include at least one of the following: cavities, loose bodies, voids, water-rich bodies, pipeline damage, soil fissures, and karst.

[0018] The risk classification rules divide risk levels into at least four levels.

[0019] Furthermore, the compliance verification of the data acquisition and pre-verification module includes:

[0020] Extract the acquired ground-penetrating radar (GPR) data and compare the GPR data with the corresponding compliance threshold data in the compliance control center module.

[0021] When the parameters collected by the ground penetrating radar meet the threshold requirements, the ground penetrating radar data is deemed compliant and released.

[0022] When the ground-penetrating radar (GPR) acquisition parameters do not meet the threshold requirements, the GPR data is determined to be non-compliant, the non-compliant item and the corresponding specification correction prompt information are output, and the GPR data is blocked from entering the data processing and anomaly identification module.

[0023] Furthermore, the data processing and anomaly detection module includes:

[0024] The data preprocessing unit is used to call the preprocessing process and parameter range specified in the compliance control center module, perform standardization processing on the verified ground-penetrating radar data, and output the processed data profile.

[0025] An anomaly identification unit is used to deploy a trained anomaly identification model, perform anomaly target detection and feature extraction based on the data profile and the anomaly discrimination rules in the compliance control center module, and output the anomaly type, spatial location, geometric size and corresponding standard discrimination criteria.

[0026] Furthermore, the standardization process includes DC offset removal, time zero-point correction, background noise removal, gain compensation, bandpass filtering, and offset repositioning. Each step calls the parameter range specified in the compliance control center module, and a preprocessed compliance verification report is output after the processing is completed.

[0027] The trained anomaly detection model is a deep learning model. The labeled categories of its training samples correspond one-to-one with the anomaly types defined in the compliance control center module. The output results of the anomaly detection model are automatically associated with the corresponding regulatory clauses, and false targets that do not conform to the anomaly discrimination rules are automatically eliminated.

[0028] Furthermore, the risk assessment and results generation module includes:

[0029] The risk assessment unit is used to invoke the risk classification rules in the compliance control center module, perform comprehensive risk score calculation and level classification based on abnormal feature information and engineering environment data, and match corresponding disposal measures recommendations.

[0030] The output unit is used to call the output rules in the compliance control center module to automatically generate a standardized output report containing an anomaly list, risk level and regulatory basis, as well as a visualization map of the spatial distribution of anomalies.

[0031] Furthermore, the assessment process of the risk assessment unit includes:

[0032] Determine the set of assessment dimensions, which includes two or more of the following: anomaly type, geometric dimensions, burial depth, engineering grade, geological conditions, and surrounding environmental factors;

[0033] According to the quantitative thresholds defined in the risk grading rules, each assessment dimension is quantitatively scored, and the comprehensive risk score is calculated based on the weighted algorithm defined in the rules.

[0034] Based on the scoring range defined by the rules, the comprehensive risk score is mapped to the corresponding risk level, and the handling measures and handling time limits defined by the rules are output simultaneously.

[0035] Furthermore, it also includes an iterative optimization module, which is used to receive external feedback data, adjust the data processing and anomaly identification module and / or the risk assessment and result generation module based on the feedback data, and synchronously update the rule set in the compliance control center module.

[0036] The ground-penetrating radar (GPR) subsurface anomaly risk assessment method, which incorporates industry-standard features, includes the following steps:

[0037] A quantitative rule set, converted from current industry standards, is pre-built and stored. The quantitative rule set includes at least compliance thresholds for data collection parameters, anomaly detection rules, risk classification rules, and output rules.

[0038] The system acquires ground-penetrating radar (GPR) acquisition parameters, GPR data, and engineering environment data. Before the data enters the processing flow, it calls the compliance threshold of the acquisition parameters to perform compliance verification on the acquisition parameters and only outputs GPR data that passes the verification.

[0039] The aforementioned anomaly detection rules are invoked to perform standardized processing and automatic anomaly identification on the verified ground-penetrating radar data, and to output anomaly feature information and its corresponding standardized judgment criteria.

[0040] By invoking the aforementioned risk grading rules and output rules, and based on the aforementioned anomaly characteristic information and engineering environment data, a risk level assessment is performed and outputs that meet the specifications are generated.

[0041] The rules called in each of the above steps are all derived from a pre-built set of quantization rules, and each step is executed automatically according to the rules called.

[0042] The beneficial effects of this invention are as follows:

[0043] This invention discloses a ground-penetrating radar underground anomaly risk assessment system embedded with industry standards. By structurally decomposing industry standards and embedding them into the entire system process, from data acquisition pre-verification, standardized preprocessing, standardized anomaly identification, compliance risk assessment to standardized output, each step is based on standards for control, eliminating human compliance errors. The output results meet industry standard requirements and can be directly used for project approval, remediation design, and emergency management without the need for secondary manual verification of standards, thus greatly improving project efficiency.

[0044] The system's risk assessment follows the clearly defined grading dimensions, quantitative thresholds, and level classification rules of current industry standards. The assessment results are consistent across the same engineering scenario, solving the problems of chaotic grading standards and non-recognition of results by the industry in existing technologies. The assessment results have industry authority and can provide a unified and reliable decision-making basis for infrastructure safety management and emergency response.

[0045] The system has industry-specific sub-libraries covering five core engineering scenarios: municipal, highway, railway, airport, and water conservancy. It incorporates current and valid standards for each industry and can flexibly call upon the corresponding compliance control rules according to the industry to which the project belongs. One system can meet the needs of multiple engineering applications and solve the problems of single-industry adaptability and poor universality of existing systems.

[0046] The system automates the entire process from data import, compliance verification, preprocessing, anomaly identification, risk assessment to output. It eliminates the need for manual analysis of radar profiles and manual comparison with standard clauses, significantly improving processing efficiency compared to the traditional manual mode. It also eliminates identification errors and classification biases caused by differences in human experience.

[0047] The system has set up interface for updating standards and interface for engineering verification, which can quickly synchronize with the latest version of industry standards. At the same time, it can continuously optimize the identification model and risk classification rules based on actual engineering excavation data, so as to ensure that the system’s identification accuracy and assessment matching degree continue to improve.

[0048] This system achieves deep integration of ground-penetrating radar underground anomaly assessment with industry standards, forming a complete closed loop of full-process compliance management. It can be widely applied in multiple fields such as urban infrastructure operation and maintenance, transportation engineering maintenance, and water conservancy project safety management. It provides a compliant, authoritative, and efficient intelligent tool for risk management of underground hidden anomalies, which can effectively reduce the risk of safety accidents caused by underground anomalies and has high economic and social benefits. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the overall architecture of a ground-penetrating radar underground anomaly risk assessment system with embedded industry standards, according to the present invention.

[0050] Figure 2This is a schematic diagram of the architecture of a compliance control hub module of a ground-penetrating radar underground anomaly risk assessment system with embedded industry standards, according to the present invention.

[0051] Figure 3 This is a flowchart illustrating Embodiment 1 of the present invention;

[0052] Figure 4 This is a schematic diagram of the results of the anomaly identification unit and risk assessment unit of a ground-penetrating radar underground anomaly risk assessment system embedded with industry standards according to the present invention. Detailed Implementation

[0053] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. The advantages and features of the present invention will become clearer from the following description and claims. It should be noted that the drawings are all in a very simplified form and use non-precise proportions, and are only used to facilitate and clarify the illustration of the embodiments of the present invention.

[0054] like Figure 1 As shown, this embodiment provides a ground-penetrating radar (GPR) underground anomaly risk assessment system embedded with industry standards. This system includes a compliance control central module, a data acquisition and pre-verification module, a data processing and anomaly identification module, and a risk assessment and result generation module. It may also include an iterative optimization module. This intelligent GPR risk assessment system, embedded with industry standards, is suitable for GPR underground anomaly detection and risk assessment in various engineering scenarios such as municipal roads, highways, railways, airport runways, and water conservancy embankments. It supports GPR data access across the entire frequency band from 100MHz to 2GHz, and achieves embedded control of industry standards throughout the entire process, eliminating the need for manual comparison with standard clauses and thus eliminating human error in compliance.

[0055] Compliance control central module:

[0056] The compliance control central module is the core control hub of the system. It stores and structurally breaks down currently effective national and industry standards, transforming unstructured standard clauses into a set of quantifiable rules and standardized templates that the system can directly access. The quantifiable rule set includes at least compliance thresholds for collected parameters, anomaly detection rules, risk grading rules, and output rules. The standardized templates are structured carriers of the output format requirements specified in the standards, including risk assessment report templates, anomaly list templates, and output map templates. These are used by various modules to directly generate standardized outputs that meet the standard format requirements. This module provides a unified basis for compliance control throughout the entire process.

[0057] like Figure 2As shown, this module includes multiple standard sub-libraries, categorized by engineering industry: municipal road industry standard sub-libraries, highway engineering industry standard sub-libraries, railway engineering industry standard sub-libraries, airport runway industry standard sub-libraries, and water conservancy and embankment industry standard sub-libraries. The system automatically calls upon the corresponding standard sub-library as the unified rule source for the entire process based on the user-selected target industry.

[0058] Each sub-library of standards stores the corresponding industry's compliance management data in a structured manner, specifically divided into the following five core data units: compliance threshold unit for collection parameters, anomaly type and discrimination standard unit, risk classification and quantitative indicator unit, handling measures and standard basis unit, and standardized template unit for results.

[0059] The data acquisition parameter compliance threshold unit is used to store the data acquisition parameter compliance thresholds specified in the corresponding industry standards, including antenna frequency, number of sampling points, point spacing, measurement line spacing, time window length, compliance thresholds and prohibitions for data acquisition environment requirements.

[0060] The anomaly type and discrimination criteria unit is used to store the anomaly discrimination rules defined in the corresponding industry specifications, including underground anomaly types, terminology definitions and radar reflection feature discrimination criteria. Underground anomaly types include at least one of the following: cavities, loose bodies, voids, water-rich bodies, pipeline damage, soil fissures and karst.

[0061] The risk grading quantification indicator unit is used to store the risk grading rules specified in the corresponding industry standards, including risk grading dimensions, quantification thresholds and level division rules. The risk level is divided into four levels: Level 1 Major Risk, Level 2 Significant Risk, Level 3 General Risk and Level 4 Low Risk.

[0062] The "Disposal Measures and Standards" unit is used to store the corresponding standard requirements, disposal measures, disposal time limits, emergency control requirements, and the corresponding standard clause numbers and original text for different risk levels.

[0063] The standardized output template unit is used to store the output rules specified in the corresponding industry standards, including the standardized format and list of required fields for exploration reports, risk assessment reports, and output maps.

[0064] The compliance control central module incorporates currently valid national and industry standards, including but not limited to: CJJ / T210-2019 "Technical Specification for Detection and Assessment of Underground Defects in Urban Roads", CJJ 56-2012 "Specification for Municipal Engineering Survey", JTG / T 3334-2018 "Specification for Highway Subgrade Design", JTG / T 3350-03-2020 "Technical Specification for Highway Subgrade Detection", TB10012-2019 "Specification for Railway Engineering Geological Survey", MH 5010-2019 "Specification for Civil Airport Flight Area Design", and GB50026-2020 "Engineering Survey Standard". It supports a standard update interface for rapid synchronization of the latest version of the standards, providing a compliance control basis for the entire system process and ensuring that all processing procedures and output results strictly comply with the standard requirements.

[0065] This module transforms unstructured industry standard texts into quantified rule sets and standardized templates that the system can directly call, providing a unified basis for compliance control throughout the entire process. This fundamentally eliminates compliance deviations caused by inconsistent rule sources across different modules. It also supports flexible access to multiple industry standard sub-libraries, allowing a single system to meet the needs of engineering applications in various scenarios.

[0066] Data acquisition and pre-verification module:

[0067] The data acquisition and pre-verification module implements compliance control at the data source. This module acquires ground-penetrating radar (GPR) acquisition parameters, GPR data, and engineering environment data, and calls the compliance threshold of the acquisition parameters in the compliance control central module. Before the GPR data enters the processing flow, compliance verification is performed on the acquisition parameters, and only GPR data that passes the verification is output to the data processing flow.

[0068] Specifically, this module supports the acquisition and import of multi-source data, including raw data from ground-penetrating radars of different brands and frequency bands, supporting borehole geological survey data, basic engineering information (road grade, highway grade, railway grade, design service life), and surrounding environmental data (sensitive buildings, dense pipelines, traffic flow).

[0069] Once the user selects a target industry, this module automatically extracts ground-penetrating radar (GPR) acquisition parameters, including antenna frequency, point spacing, survey line spacing, number of sampling points, time window length, and acquisition time. These parameters are then compared with the corresponding compliance thresholds in the compliance control module. When the GPR acquisition parameters meet the threshold requirements, the GPR data is deemed compliant and proceeds to the data processing and anomaly identification module. Conversely, when the GPR acquisition parameters do not meet the threshold requirements, the GPR data is deemed non-compliant. The module outputs the non-compliant item, the corresponding regulatory clause, and a correction prompt, and prevents the non-compliant GPR data from entering the data processing and anomaly identification module, thus eliminating compliance risks at the source.

[0070] This module performs compliance verification on ground-penetrating radar (GPR) acquisition parameters before data enters the processing flow, allowing only data that meets the specified threshold requirements to pass through. This eliminates compliance risks caused by improper acquisition parameter settings at the source. Compared to existing technologies where data directly enters processing and specifications are only used for post-processing verification, this module moves the compliance control point forward to the data acquisition stage. This effectively avoids rework and retesting due to non-compliant parameters, significantly improving engineering efficiency and reducing labor and equipment costs.

[0071] Data processing and anomaly detection module:

[0072] The data processing and anomaly detection module automatically invokes the anomaly judgment rules in the compliance control center module to perform standardized processing and automatic anomaly detection on the verified ground-penetrating radar data, outputting anomaly characteristic information and its corresponding standard judgment basis. This module includes a data preprocessing unit and an anomaly detection unit.

[0073] The data preprocessing unit performs standardized data preprocessing under regulatory constraints. It automatically invokes the preprocessing procedures and parameter ranges specified in the compliance control central module to perform standardized processing on the verified ground-penetrating radar data and outputs the processed data profile.

[0074] Furthermore, the standardization process includes DC offset removal, time zero-point correction, background noise removal, gain compensation, bandpass filtering, and offset repositioning. Each step calls upon the parameter range specified in the compliance control module. After processing, a preprocessing compliance verification report is output to ensure that the preprocessed data profile meets the standard data quality requirements and avoids abnormal feature distortion and insufficient compliance caused by improper manual preprocessing parameter settings.

[0075] The anomaly identification unit enables intelligent identification of underground anomalies as defined by regulations. Internally, the unit deploys a trained anomaly identification model. Based on preprocessed data profiles and anomaly discrimination rules from the compliance control central module, it performs anomaly target detection and feature extraction, outputting the anomaly type, spatial location, geometric dimensions, and corresponding regulatory criteria.

[0076] Furthermore, the trained anomaly detection model is a deep learning model. The labeled categories of its training samples correspond one-to-one with the anomaly types defined in the compliance control central module. The labeled samples cover radar anomaly profiles from different industries, different geological formations, and different frequency bands. The anomaly detection model's output automatically associates with the corresponding regulatory clauses and automatically eliminates false targets that do not conform to the anomaly discrimination rules, thereby improving the accuracy and compliance of the detection.

[0077] This module deeply integrates standardized processing procedures and anomaly identification models with anomaly discrimination rules in the compliance control center module, ensuring that every processing parameter meets regulatory requirements and that every identification result is automatically associated with regulatory clauses and false targets are eliminated. This guarantees the compliance and accuracy of the entire data processing and anomaly identification process.

[0078] Risk assessment and outcome generation module:

[0079] The risk assessment and output generation module automatically invokes the risk grading rules and output rules in the compliance control center module. Based on anomaly characteristic information and engineering environment data, it performs risk level assessment and generates outputs that meet regulatory requirements. This module includes a risk assessment unit and an output unit.

[0080] The risk assessment unit implements intelligent risk assessment in accordance with industry standards. It invokes the risk grading rules in the compliance control central module to determine a set of assessment dimensions, including two or more of the following: anomaly type, geometric dimensions, burial depth, engineering grade, geological conditions, and surrounding environmental factors. For each identified anomaly, each assessment dimension is quantitatively scored according to the quantification thresholds defined in the risk grading rules. A comprehensive risk score is calculated based on a weighted algorithm defined in the rules. The comprehensive risk score is mapped to the corresponding risk level according to the scoring range defined in the rules, and the rule-defined remedial measures and timelines are output simultaneously.

[0081] Furthermore, the risk assessment unit also conducts risk causation analysis based on the assessment results and prioritizes the handling of multiple anomalies according to their risk level and the degree of impact on the project.

[0082] This risk assessment unit automatically performs risk assessments by following the clearly defined grading dimensions, quantitative thresholds, and level classification rules in industry standards. This achieves the standardization and authoritativeness of assessment standards, providing a unified and reliable basis for safety management and emergency response.

[0083] The output unit enables standardized outputs that meet industry standards. It invokes output rules from the compliance control module to automatically generate standardized reports containing anomaly lists, risk levels, and regulatory basis, as well as visual maps showing the spatial distribution of anomalies. Specifically, it includes two types of compliance deliverables: first, a formal risk assessment report that meets industry standard submission requirements, containing all required fields, appendices, and regulatory basis clauses, which can be directly used for project submissions, emergency response, and remediation design; second, lightweight visual deliverables, including radar profile anomaly annotation maps, anomaly planar distribution maps, and risk level heat maps, supporting export in common formats such as CAD, SHP, and PDF, adaptable to on-site engineering application needs.

[0084] This output unit automatically generates standardized reports and visualizations by calling the output rules in the specifications, ensuring that the output meets the requirements for submission and approval in industry standards. It can be directly used for engineering submission and remediation design without the need for secondary modifications.

[0085] Iterative optimization module:

[0086] The iterative optimization module receives external feedback data. This feedback data includes actual excavation verification data, engineering handling feedback data, and updated specification documents. Based on this feedback data, the module adjusts the anomaly identification model in the data processing and anomaly identification module and / or the risk grading rules in the risk assessment and results generation module, and simultaneously updates the rule set in the compliance control center module to ensure continuous improvement in the system's compliance and identification / assessment accuracy.

[0087] Furthermore, the module is equipped with two core interfaces: one is the specification update interface, which supports importing the latest version of industry specification documents, automatically disassembling and updating structured data units, and synchronously updating compliance control rules throughout the entire process; the other is the engineering verification interface, which supports accessing actual excavation verification data and disposal feedback data, automatically calculating the anomaly identification accuracy and risk classification matching degree, and iteratively optimizing the weight parameters of the identification model and the weighting rules of risk assessment. After optimization, the anomaly identification accuracy of the model is no less than 95%, and the risk classification matching degree with specification requirements is no less than 98%.

[0088] This iterative optimization module enables rapid synchronization of compliance control rules and continuous optimization of the identification and assessment model through standardized update interfaces and engineering verification interfaces, ensuring the long-term compliance timeliness and identification and assessment accuracy of the system.

[0089] Among them, the data acquisition and pre-verification module, the data processing and anomaly identification module, the risk assessment and result generation module, and the iterative optimization module are connected to the compliance control center module through bidirectional communication via an industrial bus, forming a closed-loop processing architecture with the compliance control center module as the unified rule source, and each module automatically performing corresponding processing and evaluation operations according to the rules called.

[0090] The aforementioned system solves the problems of existing systems being out of touch with industry standards and having inconsistent risk assessment standards by breaking down industry standards into a set of quantitative rules that the system can call. It uses a compliance control central module as a unified rule source to drive the automatic execution of the entire process of data collection and verification, anomaly identification, risk assessment, and output. This ensures that the output meets the requirements of industry standards without the need for manual comparison with the standard clauses.

[0091] This embodiment also provides a ground-penetrating radar (GPR) underground anomaly risk assessment method with embedded industry standards. This method can be executed by the GPR risk intelligent assessment system with embedded industry standards described in the above embodiment. The method includes the following steps: S1-S4, and may also include S5.

[0092] S1, pre-build and store a quantitative rule set transformed from current industry standards. The quantitative rule set includes at least compliance thresholds for collected parameters, anomaly detection rules, risk classification rules, and output rules.

[0093] Specifically, the existing national and industry standards are structurally decomposed, and unstructured standard clauses are transformed into quantitative rule sets and standardized templates that the system can directly call. The quantitative rule sets are divided into multiple standard sub-libraries according to engineering industry categories, including municipal road industry standard sub-libraries, highway engineering industry standard sub-libraries, railway engineering industry standard sub-libraries, airport runway industry standard sub-libraries, and water conservancy and embankment industry standard sub-libraries. Each standard sub-library stores the corresponding industry's collection parameter compliance thresholds, anomaly detection rules, risk classification rules, and output rules in a structured manner.

[0094] S2 acquires ground-penetrating radar (GPR) acquisition parameters, GPR data, and engineering environment data. It calls the compliance threshold of the acquisition parameters to perform compliance verification on the acquisition parameters before the data enters the processing flow, and only outputs GPR data that passes the verification.

[0095] Specifically, the process involves acquiring raw data from ground-penetrating radars (GPRs) of different brands and frequency bands, along with supporting borehole geological survey data, basic engineering information, and surrounding environmental data. It automatically extracts GPR acquisition parameters, including antenna frequency, point spacing, survey line spacing, number of sampling points, time window length, and acquisition time, and compares these parameters with corresponding compliance thresholds. When the GPR acquisition parameters meet the threshold requirements, the data is deemed compliant and allowed to proceed; when the parameters do not meet the threshold requirements, the data is deemed non-compliant, and the non-compliant item, the corresponding original regulatory clause, and correction prompts are output. The data is then blocked from further processing.

[0096] S3 invokes the anomaly detection rules to perform standardized processing and automatic anomaly identification on the verified ground-penetrating radar data, and outputs anomaly feature information and its corresponding standardized judgment criteria.

[0097] Specifically, the standardized processing includes DC offset removal, time zero-point correction, background noise removal, gain compensation, bandpass filtering, and offset repositioning. Each step calls upon the parameter range specified in the standard. After processing, a preprocessing compliance verification report is output to ensure that the processed data profile meets the standard data quality requirements.

[0098] Automatic anomaly identification is performed through a trained anomaly detection model, which is a deep learning model. The labeled categories of its training samples correspond one-to-one with the anomaly types defined in the specifications. Based on the preprocessed data profile and anomaly discrimination rules, the model performs anomaly target detection and feature extraction, outputting the anomaly type, spatial location, geometric dimensions, and corresponding specification discrimination criteria, and automatically eliminating false targets that do not conform to the anomaly discrimination rules.

[0099] S4 invokes risk grading rules and output rules, performs risk level assessment based on anomaly characteristic information and engineering environment data, and generates output that meets the specifications.

[0100] Specifically, the risk level assessment includes: determining a set of assessment dimensions, which may include two or more of the following: anomaly type, geometric dimensions, burial depth, engineering grade, geological conditions, and surrounding environmental factors; quantifying and scoring each assessment dimension according to the quantitative thresholds defined in the risk grading rules, and calculating a comprehensive risk score based on the weighted algorithm defined in the rules; mapping the comprehensive risk score to the corresponding risk level according to the scoring range defined in the rules, and simultaneously outputting the disposal measures and disposal time limits defined in the rules. Furthermore, risk causal analysis can be performed based on the assessment results, and multiple anomalies can be prioritized for disposal according to their risk level and degree of engineering impact.

[0101] The deliverables include automatically generated standardized reports containing anomaly lists, risk levels, and regulatory justifications, as well as visualizations of the spatial distribution of anomalies. Specifically, this includes a formal risk assessment report that meets industry standard submission requirements and lightweight visualizations, supporting export in common formats such as CAD, SHP, and PDF.

[0102] It may also include S5, which receives external feedback data, adjusts the model parameters for automatic anomaly identification and / or the rule weights for risk level assessment based on the feedback data, and updates the quantitative rule set synchronously.

[0103] Specifically, the latest version of industry standard documents can be imported through the standard update interface, the structured data units can be automatically disassembled and updated, and the compliance control rules of the entire process can be updated synchronously. Actual excavation verification data and disposal feedback data can be accessed through the engineering verification interface, the anomaly identification accuracy and risk classification matching degree can be automatically calculated, and the weight parameters of the identification model and the weighting rules of risk assessment can be iteratively optimized. After optimization, the anomaly identification accuracy of the model is not less than 95%, and the risk classification matching degree with the standard requirements is not less than 98%.

[0104] The rules called in each step are all derived from a pre-built set of quantitative rules. Each step is executed automatically according to the called rules, achieving unified rule-driven operation throughout the entire process.

[0105] The above method solves the problems of existing methods being out of touch with industry standards and having inconsistent evaluation standards by pre-converting industry standards into a set of quantitative rules and using this set to uniformly drive the entire process of data collection and verification, anomaly identification, risk assessment and output. It ensures that the output meets the requirements of industry standards without the need for manual comparison with the standard clauses.

[0106] Example 1:

[0107] This embodiment is applied to a project for underground anomaly detection and risk assessment on a main road. This road is an urban expressway with a design speed of 60 km / h, a width of 40 m, high traffic volume, and densely distributed municipal pipelines and residential areas, making it a core urban infrastructure. The project uses a 400MHz antenna ground-penetrating radar to deploy survey lines longitudinally along the road, with a line spacing of 1 m, a point spacing of 2 cm, and a detection depth of 0-5 m. The core objective is to identify underground cavities, loose structures, and other defects, complete a compliance risk assessment, and output an assessment report that meets industry standards, providing a basis for road safety maintenance and defect remediation. The current standards followed by the project are CJJ / T 210-2019 "Technical Specification for Detection and Assessment of Underground Defects in Urban Roads" and CJJ 56-2012 "Specification for Municipal Engineering Survey".

[0108] like Figure 3 As shown, this embodiment uses the aforementioned ground-penetrating radar underground anomaly risk assessment system with embedded industry standards. The specific implementation process is as follows:

[0109] Step 1: Industry Selection and Standard Application

[0110] After logging into the system, users select "Municipal Road Industry" as the industry to which the project belongs in the project information interface. The system automatically calls the municipal road industry standard sub-library in the compliance control center module, loads the compliance threshold, anomaly judgment rules, risk classification rules and output rules of the collection parameters corresponding to CJJ / T 210-2019 and CJJ 56-2012 in the sub-library, and completes the pre-configuration of compliance control rules.

[0111] Step Two: Data Import and Pre-Compliance Verification

[0112] Users import the ground-penetrating radar (GPR) data and its acquisition parameters, along with supporting borehole geological survey data, road infrastructure information, and surrounding environmental data, through the data acquisition and pre-verification module. This module automatically extracts the GPR acquisition parameters: antenna frequency 400MHz, point spacing 2cm, survey line spacing 1m, number of sampling points 512, and time window length 100ns. The module compares the extracted GPR acquisition parameters with the compliance thresholds for acquisition parameters in the municipal road specification sub-database. All parameters meet the acquisition parameter requirements for underground defect detection of urban expressways in CJJ / T 210-2019 and are deemed compliant data, automatically proceeding to subsequent processing. If any parameters do not meet the specification requirements, the system immediately displays a non-compliance warning, clearly indicating the non-compliant item, the corresponding specification clause number and original text, and correction suggestions, preventing the non-compliant data from entering the subsequent process.

[0113] Step 3: Standardization Preprocessing under Normative Constraints

[0114] The data preprocessing unit automatically invokes the preprocessing procedures and parameter ranges specified in the compliance control central module. It sequentially performs DC offset removal, time zero-point correction, background noise removal, AGC gain compensation, 100-800MHz bandpass filtering, and offset repositioning on compliant ground-penetrating radar data. The parameters for each step are limited to the range required by the specifications. Upon completion, a preprocessing compliance verification report is automatically generated, confirming that the preprocessed data profile meets the data quality requirements specified in CJJ / T 210-2019.

[0115] Step 4: Intelligent Identification of Underground Anomalies under Standard Definition

[0116] like Figure 4 As shown, the anomaly detection unit invokes a trained deep learning detection model. This model employs a network structure based on an improved YOLOv8, and the labeled categories of its training samples correspond one-to-one with the anomaly types defined in the municipal road specification sub-database. Based on the preprocessed data profile, and combined with the anomaly types, discrimination criteria, and radar reflection features defined in the anomaly discrimination rules, the model automatically identifies and extracts features, identifying a total of four underground anomalies:

[0117] Anomaly 1: A cavity, measuring 1.5m × 1.0m, buried at a depth of 2.3m, located beneath the fast lane of the road;

[0118] Anomaly 2: Loose body, measuring 3.0m × 1.5m, buried at a depth of 1.8m, located under the fast lane of the road;

[0119] Anomaly 3: Void, dimensions 0.8m × 0.3m, burial depth 0.6m, located below the sidewalk;

[0120] Anomaly 4: A water-rich body, measuring 2.2m × 1.0m, buried at a depth of 3.5m, located below the slow lane of the road.

[0121] Each anomaly is associated with the corresponding anomaly type definition and discrimination clause in CJJ / T 210-2019, and two false targets that do not meet the anomaly discrimination rules are automatically removed.

[0122] Step 5: Standardized Risk Intelligent Assessment

[0123] like Figure 4 As shown, the risk assessment unit calls upon the risk grading rules specified in CJJ / T 210-2019 from the municipal road specification sub-library to determine the assessment dimension set, including anomaly type, geometric dimensions, burial depth, engineering grade, geological conditions, surrounding sensitive environment, and traffic load conditions. Four anomalies are scored according to the quantitative thresholds defined in the rules, and a comprehensive risk score is calculated based on the weighted algorithm defined in the rules. The risk level is then determined based on the score range.

[0124] Anomaly 1 (Void): The overall score is 92 points, which is judged as a Level 1 major risk. The corresponding emergency control measures are to immediately take emergency control measures, close the corresponding lane, complete the emergency response within 24 hours, and complete the radical treatment within 7 days.

[0125] Anomaly 2 (Loose Body): The overall score is 78 points, which is judged as a level 2 major risk. The corresponding disposal measures are to complete the special treatment within 15 days and increase the frequency of monitoring during the treatment period.

[0126] Anomaly 3 (Void): The overall score is 55 points, which is judged as a level three general risk. The corresponding disposal measures are to include it in the daily monitoring log and monitor it once every 3 months.

[0127] Anomaly 4 (Water-rich body): The overall score is 42 points, which is judged as a level 4 low risk. The corresponding disposal measure is to include it in the annual routine testing scope.

[0128] Meanwhile, the risk assessment unit conducted risk causal analysis on the four anomalies and ranked the handling priorities according to risk level and degree of engineering impact, with anomaly 1 having the highest handling priority.

[0129] Finally, the risk classification results and disposal recommendations are verified for compliance to ensure full compliance with the requirements of the regulations. If the verification fails, the assessment process is automatically re-executed. If the verification passes, the final risk assessment result is output, thus completely resolving the problem of inconsistent existing technical risk classification standards and their disconnect from regulations.

[0130] Step Six: Output of Compliance Results

[0131] The output unit automatically calls the output rules in the municipal road specifications sub-library to generate a standardized output report that meets the requirements of CJJ / T210-2019, including an anomaly list, risk level ledger, handling suggestions, and corresponding standard provisions. It also generates a visual map of the spatial distribution of anomalies, supporting export in CAD and PDF formats.

[0132] Step 7: Closed-loop iterative optimization

[0133] After the project was completed, the construction unit excavated and verified the four anomalies. The verification results were consistent with the anomaly type, size, and burial depth identified by the system, and the risk level classification matched the specifications. The user imported the verification data into the iterative optimization module. Based on the feedback data, the system adjusted and optimized the identification model and evaluation rules, and simultaneously updated the rule set in the compliance control center module. After optimization, the anomaly identification accuracy of the model improved to 98%.

[0134] Implementation Example Verification: This implementation example uses a ground-penetrating radar underground anomaly risk assessment system with embedded industry standards. The entire process from data import to output is significantly shortened, and efficiency is significantly improved compared to the traditional manual processing and manual comparison with standards. The anomaly identification accuracy rate reaches 96%, which is improved to 98% after optimization. The risk classification matches the standard requirements 100%. The output assessment report meets the industry standard requirements and directly passed the review of the municipal management department, fully verifying the effectiveness and engineering practical value of this system.

[0135] The above description is merely a description of preferred embodiments of the present invention and is not intended to limit the scope of the present invention in any way. Any changes or modifications made by those skilled in the art based on the above disclosure shall fall within the protection scope of the claims.

Claims

1. An industry-standard ground-penetrating radar underground anomaly risk assessment system, characterized in that: include: The compliance control central module is used to store and structurally decompose existing effective industry standards, and transform the standard clauses into a set of quantitative rules and standardized templates that can be called by the system. The set of quantitative rules includes at least compliance thresholds for collected parameters, anomaly detection rules, risk classification rules, and output rules. The data acquisition and pre-verification module is used to acquire ground-penetrating radar acquisition parameters, ground-penetrating radar data and engineering environment data, and call the acquisition parameter compliance threshold in the compliance control center module to perform compliance verification on the ground-penetrating radar acquisition parameters before the ground-penetrating radar data enters the processing flow, and only outputs the ground-penetrating radar data that passes the verification to the data processing flow. The data processing and anomaly identification module is used to call the anomaly discrimination rules in the compliance control center module, perform standardized processing and automatic anomaly identification on the verified ground penetrating radar data, and output anomaly feature information and its corresponding standard discrimination basis; The risk assessment and result generation module is used to call the risk classification rules and result output rules in the compliance control center module, and perform risk level assessment and generate result output that meets the specifications based on the abnormal feature information and engineering environment data. The data acquisition and pre-verification module, the data processing and anomaly identification module, and the risk assessment and result generation module are all connected to the compliance control center module, forming a closed-loop processing architecture in which the compliance control center module serves as the unified rule source, and each module automatically performs corresponding processing and assessment operations according to the rules it calls.

2. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 1, characterized in that, The compliance control central module includes multiple standard sub-libraries, which are divided according to engineering industry categories. Each standard sub-library stores the compliance thresholds of the collection parameters, anomaly judgment rules, risk classification rules and output rules of the corresponding industry in a structured manner. The system automatically calls the corresponding standard sub-library as the unified rule source for the entire process, based on the selected target industry.

3. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 2, characterized in that, The engineering industry categories include at least one of municipal roads, highways, railways, airport runways, and water conservancy embankments; The underground anomaly types defined in the anomaly discrimination rules include at least one of the following: cavities, loose bodies, voids, water-rich bodies, pipeline damage, soil fissures, and karst. The risk classification rules divide risk levels into at least four levels.

4. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 1, characterized in that, The compliance verification of the data acquisition and pre-verification module includes: Extract the acquired ground-penetrating radar (GPR) data and compare the GPR data with the corresponding compliance threshold data in the compliance control center module. When the parameters collected by the ground penetrating radar meet the threshold requirements, the ground penetrating radar data is deemed compliant and released. When the ground-penetrating radar (GPR) acquisition parameters do not meet the threshold requirements, the GPR data is determined to be non-compliant, the non-compliant item and the corresponding specification correction prompt information are output, and the GPR data is blocked from entering the data processing and anomaly identification module.

5. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 1, characterized in that, The data processing and anomaly detection module includes: The data preprocessing unit is used to call the preprocessing process and parameter range specified in the compliance control center module, perform standardization processing on the verified ground-penetrating radar data, and output the processed data profile. An anomaly identification unit is used to deploy a trained anomaly identification model, perform anomaly target detection and feature extraction based on the data profile and the anomaly discrimination rules in the compliance control center module, and output the anomaly type, spatial location, geometric size and corresponding standard discrimination criteria.

6. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 5, characterized in that, The standardization process includes DC offset removal, time zero-point correction, background noise removal, gain compensation, bandpass filtering, and offset repositioning. Each step calls the parameter range specified in the compliance control center module. After the processing is completed, a preprocessed compliance verification report is output. The trained anomaly detection model is a deep learning model. The labeled categories of its training samples correspond one-to-one with the anomaly types defined in the compliance control center module. The output results of the anomaly detection model are automatically associated with the corresponding regulatory clauses, and false targets that do not conform to the anomaly discrimination rules are automatically eliminated.

7. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 1, characterized in that, The risk assessment and results generation module includes: The risk assessment unit is used to invoke the risk classification rules in the compliance control center module, perform comprehensive risk score calculation and level classification based on abnormal feature information and engineering environment data, and match corresponding disposal measures recommendations. The output unit is used to call the output rules in the compliance control center module to automatically generate a standardized output report containing an anomaly list, risk level and regulatory basis, as well as a visualization map of the spatial distribution of anomalies.

8. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 7, characterized in that, The assessment process of the risk assessment unit includes: Determine the set of assessment dimensions, which includes two or more of the following: anomaly type, geometric dimensions, burial depth, engineering grade, geological conditions, and surrounding environmental factors; According to the quantitative thresholds defined in the risk grading rules, each assessment dimension is quantitatively scored, and the comprehensive risk score is calculated based on the weighted algorithm defined in the rules. Based on the scoring range defined by the rules, the comprehensive risk score is mapped to the corresponding risk level, and the handling measures and handling time limits defined by the rules are output simultaneously.

9. The ground-penetrating radar underground anomaly risk assessment system with embedded industry standards as described in claim 1, characterized in that, It also includes an iterative optimization module, which receives external feedback data, adjusts the data processing and anomaly identification module and / or the risk assessment and result generation module based on the feedback data, and synchronously updates the rule set in the compliance control center module.

10. A ground-penetrating radar (GPR) method for assessing underground anomalies with embedded industry standards, characterized in that: Includes the following steps: A quantitative rule set, converted from current industry standards, is pre-built and stored. The quantitative rule set includes at least compliance thresholds for data collection parameters, anomaly detection rules, risk classification rules, and output rules. The system acquires ground-penetrating radar (GPR) acquisition parameters, GPR data, and engineering environment data. Before the data enters the processing flow, it calls the compliance threshold of the acquisition parameters to perform compliance verification on the acquisition parameters and only outputs GPR data that passes the verification. The aforementioned anomaly detection rules are invoked to perform standardized processing and automatic anomaly identification on the verified ground-penetrating radar data, and to output anomaly feature information and its corresponding standardized judgment criteria. By invoking the aforementioned risk grading rules and output rules, and based on the aforementioned anomaly characteristic information and engineering environment data, a risk level assessment is performed and outputs that meet the specifications are generated. The rules called in each of the above steps are all derived from a pre-built set of quantization rules, and each step is executed automatically according to the rules called.

Images