Underground pipeline automatic modeling method based on GIS and artificial intelligence technology

Abstract

The invention provides an underground pipeline automatic modeling method based on a GIS and artificial intelligence technology. The underground pipeline automatic modeling method comprises the steps that a lightweight sensor is used for collecting the spatial position and attribute data of an underground pipeline; a data processing flow is optimized through a machine learning algorithm, and high precision is kept while computing resource requirements are reduced; spatial analysis is carried out in combination with a geographic information system (GIS), a three-dimensional model is automatically constructed, and the accuracy of the model is verified; a model simplification technology is implemented, wrong connection or missing parts are intelligently recognized and corrected, and the processing time is shortened while the final model quality is ensured. Through the scheme of the embodiment of the invention, the problem of how to improve data processing efficiency and modeling precision under the condition of reducing hardware and software requirements can be solved.

Description

Technical Field

[0001] The present application relates to the field of geographic information system (GIS) technology, and in particular to an automatic modeling method for underground pipelines based on GIS and artificial intelligence technology. Background Art

[0002] Pipeline automatic modeling technology is a technology used in the integrated design of municipal pipelines and underground pipeline management. When conducting integrated design of municipal pipelines, designers usually receive geophysical exploration plans of existing pipelines made by the surveying and mapping department provided by Party A. These drawings will have lines representing pipelines, but since there are no attributes such as pipeline diameter, it is impossible to visually check whether there is a collision between pipelines or whether the clearance meets the specification requirements. For the horizontal and vertical clearances between pipelines, it is necessary to manually calculate or query information such as elevation for deduction.

[0003] In addition, underground pipelines are highly concealed, complex, and large in size, and traditional two-dimensional design methods are difficult to meet the needs of modern pipeline management. Therefore, improving visualization technology, making three-dimensional restoration more vivid and realistic, and completing three-dimensional visualization of pipeline veins are the key to achieving early warning and timely handling of pipeline network faults.

[0004] The automatic modeling method of underground pipelines based on GIS and artificial intelligence technology can combine geographic information system (GIS) and artificial intelligence (AI) technology to achieve a method of fast and accurate modeling of underground pipelines. By utilizing the spatial data management and analysis capabilities provided by GIS and combining technologies such as machine learning algorithms in AI, it is possible to automatically identify and extract information about underground pipelines and build a three-dimensional model. This method can significantly improve the speed and accuracy of modeling and reduce the need for manual intervention.

[0005] However, this approach also faces some challenges, especially in terms of how to improve data processing efficiency and modeling accuracy while reducing hardware and software requirements. Due to the large amount of data and complexity of underground pipelines, traditional processing methods often require high-performance computing resources to ensure the quality of modeling. Therefore, how to optimize the algorithm to adapt to a wider range of hardware configurations while ensuring that the accuracy of modeling is not affected has become an urgent problem to be solved. In addition, it is also necessary to consider how to simplify the software operation process so that non-professionals can easily use these tools for pipeline modeling. Summary of the invention

[0006] In view of this, the embodiments of the present disclosure provide an automatic modeling method for underground pipelines based on GIS and artificial intelligence technology to solve the problems existing in the prior art.

[0007] The automatic modeling method of underground pipelines based on GIS and artificial intelligence technology includes:

[0008] Use lightweight sensors to collect spatial location and attribute data of underground pipelines;

[0009] Optimize data processing through machine learning algorithms, reduce computing resource requirements while maintaining high accuracy;

[0010] Combined with Geographic Information System (GIS) for spatial analysis, automatically build three-dimensional models and verify their accuracy;

[0011] Implement model simplification techniques and intelligently identify and correct incorrect connections or missing parts to ensure the quality of the final model while reducing processing time.

[0012] According to one embodiment, the collecting of spatial position and attribute data of underground pipelines by using lightweight sensors includes:

[0013] Using lightweight sensors to emit electromagnetic wave signals of specific frequencies;

[0014] Receive the reflected signal and record the signal strength and reflection time;

[0015] Calculate the depth of underground pipelines based on reflection time;

[0016] Whether it is a valid signal is determined based on the signal strength and a preset threshold, wherein when the signal strength is greater than the preset threshold, it is determined to be a valid signal.

[0017] According to one embodiment, receiving the reflected signal and recording the signal strength and reflection time includes:

[0018] Start the timer and begin recording time;

[0019] monitoring the arrival of reflected signals;

[0020] Stop the timer and record the reflex time;

[0021] Whether it is a valid reflection is determined according to the reflection time and a preset time range, wherein when the reflection time falls within the preset time range, it is determined to be a valid reflection.

[0022] According to one embodiment, starting the timer and beginning to record time includes:

[0023] Initialize the timer;

[0024] Set the timer to start state;

[0025] Start recording time;

[0026] The timing is stopped when the recording time exceeds the preset maximum waiting time, wherein the timing is stopped when the recording time exceeds the preset maximum waiting time.

[0027] According to one embodiment, setting the timer to a start state includes:

[0028] Set the timer status flag to the start state;

[0029] Wait for the reflected signal to return;

[0030] Read the timer status after receiving the reflected signal;

[0031] If the timer status flag is still in the started state, continue to wait until the signal returns, wherein if the timer status flag is still in the started state, continue to wait until the signal returns.

[0032] According to one embodiment, setting the status flag of the timer to the start state includes:

[0033] Clear the timer status flag;

[0034] Set the timer status flag to the start state;

[0035] Monitor the changes of the timer status flag;

[0036] If the timer status flag bit does not change, the start state is maintained, wherein if the timer status flag bit does not change, the start state is maintained.

[0037] According to one embodiment, the method of optimizing the data processing flow by a machine learning algorithm to reduce computing resource requirements while maintaining high accuracy includes:

[0038] Use the preprocessing module to perform preliminary cleaning on the collected data;

[0039] Use feature selection algorithm to select the feature set that has the greatest impact on model construction;

[0040] Apply lightweight model training technology to reduce computing resource consumption;

[0041] The model parameters are adjusted based on the feature importance score FIS, which is output by the machine learning model.

[0042] According to one embodiment, the feature selection algorithm selects the feature set that has the greatest impact on model building, including:

[0043] Calculate the correlation coefficient between each feature and the target variable;

[0044] Set a correlation coefficient threshold T, where the feature is retained when the correlation coefficient is greater than T;

[0045] Feature screening is performed based on the correlation coefficient threshold T, where the value range of T is between 5 and 9.

[0046] According to one embodiment, setting the correlation coefficient threshold T includes:

[0047] Determine a candidate threshold set C;

[0048] Traverse each threshold in the candidate threshold set C;

[0049] Calculate the performance index of the model at different thresholds based on the candidate threshold set C;

[0050] The threshold that makes the performance index optimal is selected as T.

[0051] According to one embodiment,

[0052] For each candidate threshold, the model performance was evaluated using a cross-validation method;

[0053] Calculate the average accuracy ACC based on the cross-validation results;

[0054] If the average accuracy ACC is higher than the predetermined minimum accuracy requirement, the current threshold is recorded;

[0055] The threshold value that gives the highest average accuracy rate ACC among the recorded threshold values ​​is selected as T.

[0056] According to one embodiment, if the average accuracy rate ACC is higher than a predetermined minimum accuracy rate requirement, the method includes:

[0057] Set the minimum accuracy requirement LAR;

[0058] Based on the minimum accuracy requirement, LAR determines whether the average accuracy under the current threshold meets the conditions;

[0059] If the average accuracy ACC is greater than or equal to LAR, the current threshold is recorded;

[0060] Otherwise, continue to traverse the next threshold.

[0061] The present invention provides an automatic modeling method for underground pipelines based on GIS and artificial intelligence technology, including: using lightweight sensors to collect spatial location and attribute data of underground pipelines; optimizing data processing flow through machine learning algorithms to reduce computing resource requirements while maintaining high accuracy; combining with geographic information system GIS for spatial analysis, automatically building a three-dimensional model and verifying its accuracy; implementing model simplification technology and intelligently identifying and correcting wrong connections or missing parts to ensure the quality of the final model while reducing processing time. Through the solution of the embodiment of the present disclosure, it is possible to solve how to improve data processing efficiency and modeling accuracy while reducing hardware and software requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the multiple drawings represent the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments disclosed in the present application and should not be regarded as limiting the scope of the present application.

[0063] Figure 1 It is a flow chart of the automatic modeling method of underground pipelines based on GIS and artificial intelligence technology;

[0064] Figure 2 It is a flow chart for collecting spatial location and attribute data of underground pipelines using lightweight sensors;

[0065] Figure 3 It is a flow chart for receiving the reflected signal and recording the signal strength and reflection time;

[0066] Figure 4 It is a flowchart to start the timer and start recording time;

[0067] Figure 5 It is a flowchart for setting the timer to the start state;

[0068] Figure 6 It is a flow chart for setting the status flag of the timer to the start state;

[0069] Figure 7 It is a flowchart that optimizes the data processing process through machine learning algorithms, reduces the demand for computing resources while maintaining high accuracy;

[0070] Figure 8 It is a flowchart of the feature selection algorithm to select the feature set that has the greatest impact on model construction;

[0071] Fig. 9 is a flow chart for setting the correlation coefficient threshold T;

[0072] Fig.10 It is a flowchart for calculating the performance indicators of the model at different thresholds;

[0073] Fig.11 It is a flow chart if the average accuracy ACC is higher than the predetermined minimum accuracy requirement. DETAILED DESCRIPTION

[0074] The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

[0075] The following describes the embodiments of the present disclosure through specific examples, and those skilled in the art can easily understand other advantages and effects of the present disclosure from the contents disclosed in this specification. Obviously, the described embodiments are only a part of the embodiments of the present disclosure, rather than all of the embodiments. The present disclosure can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present disclosure. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present disclosure.

[0076] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein may be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present disclosure, it should be understood by those skilled in the art that an aspect described herein may be implemented independently of any other aspect, and two or more of these aspects may be combined in various ways. For example, any number of aspects described herein may be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein may be used to implement this device and / or practice this method.

[0077] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present disclosure. The drawings only show components related to the present disclosure rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0078] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects described may be practiced without these specific details.

[0079] Next, refer to Figure 1 Description of this underground pipeline automatic modeling method based on GIS and artificial intelligence technology:

[0080] S101: Use lightweight sensors to collect spatial location and attribute data of underground pipelines. This process involves the use of advanced sensor technology to obtain detailed information about underground pipelines. The choice of lightweight sensors is to reduce the cost and complexity of the overall system while ensuring the quality of data collection. For example, in practical applications, non-invasive sensors such as ground penetrating radar GPR and magnetometers can be used to detect the location, depth and possible material of underground pipelines. These sensors are usually installed on mobile platforms such as carts or vehicles to enable rapid scanning throughout the survey area. The data collected in this way includes not only the specific location of the pipeline, but also other key attributes, such as the type of pipeline (water supply, drainage, electricity, etc.), diameter size, etc. In order to ensure the accuracy and completeness of the data, preliminary quality control and preprocessing of the collected data are also required, such as removing noise and filling missing values.

[0081] S102: Optimize the data processing flow through machine learning algorithms to reduce computing resource requirements while maintaining high accuracy. After collecting a large amount of raw data, the next task is to efficiently process and analyze this data. The key to this step is to use machine learning techniques to automatically identify and classify pipeline features, thereby reducing the need for manual intervention. For example, a deep learning model, such as a convolutional neural network (CNN), can be trained to automatically extract pipeline contours and location information from GPR images. In addition, clustering algorithms can be used to identify similar pipeline features to further increase the speed of data processing. In order to reduce the demand for computing resources, model compression techniques such as pruning and quantization can be used to reduce the size and computational complexity of the model while maintaining high prediction accuracy. This approach not only speeds up data processing, but also ensures that these algorithms can be run even on lower-performance hardware devices.

[0082] S103: Combined with Geographic Information System (GIS) for spatial analysis, automatically build a three-dimensional model and verify its accuracy. After completing the data processing, the next task is to integrate the processed data into the Geographic Information System (GIS) platform to build a detailed three-dimensional underground pipeline model. The GIS platform provides powerful spatial analysis tools that can help us better understand the spatial relationship between pipelines and generate a three-dimensional model based on this. For example, the spatial interpolation technology in GIS can be used to estimate the exact location of the pipeline in three-dimensional space. In addition, the visualization function of GIS can be used to check the rationality of the model and verify the accuracy of the model by comparing it with known ground features. For example, if there are known buildings or other landmarks in an area, the accuracy of the model can be evaluated by comparing whether these landmarks are consistent with the pipeline locations in the model.

[0083] S104: Implement model simplification technology and intelligent identification to correct wrong connections or missing parts, ensuring the quality of the final model while reducing processing time. After building a preliminary 3D model, it needs to be optimized and corrected to ensure the quality of the model. This step involves two aspects of work: model simplification and error detection and correction. Model simplification technology can reduce the complexity of the model, thereby reducing the time consumption of subsequent processing. For example, Multi-Resolution Representation (MRR) can be used to create model versions with different levels of detail, so that the appropriate model accuracy can be selected in different application scenarios. On the other hand, intelligent identification technology can help automatically detect and correct wrong connections or missing parts in the model. For example, connectivity analysis in graph theory can be used to identify unreasonable pipeline connections, and proximity analysis can be used to infer the reasonable location of the missing parts. Through the application of these technologies, the processing time can be significantly shortened while ensuring the quality of the model.

[0084] Through the above steps, the problem of how to improve data processing efficiency and modeling accuracy while reducing hardware and software requirements can be solved. Throughout the process, by selecting lightweight sensors and optimized data processing algorithms, the dependence on high-performance computing resources is reduced; and through the spatial analysis capabilities and model simplification technology of the GIS platform, high-quality three-dimensional models can be obtained even under limited computing conditions. This comprehensive solution not only improves work efficiency, but also reduces the overall cost of the project, making the automatic modeling technology of underground pipelines more popular and practical.

[0085] Next, refer to Figure 2 Describe the specific steps of using lightweight sensors to collect spatial location and attribute data of underground pipelines:

[0086] S201: Use lightweight sensors to emit electromagnetic wave signals of specific frequencies. This process involves selecting the appropriate frequency to ensure that the signal can penetrate the ground and be reflected by the underground pipeline. The design of the sensor needs to take into account the signal's penetration and resolution to ensure that pipelines of different depths and materials can be accurately detected.

[0087] S202: Receive the reflected signal and record the signal strength and reflection time. The signal strength received by the sensor will vary depending on the pipeline material, soil conditions, and the distance between the pipeline and the sensor. The reflection time is directly related to the actual depth of the pipeline. This information is crucial for subsequent data processing.

[0088] S203: Calculate the depth of the underground pipeline based on the reflection time. The known signal propagation speed is usually close to the speed of light and the reflection time, so the round trip time of the signal can be calculated, and then the approximate depth of the pipeline can be calculated. This step requires a precise time measurement device to ensure the accuracy of the calculation results.

[0089] S204: Determine whether it is a valid signal based on the signal strength and the preset threshold. In actual operation, if the signal strength is greater than the preset threshold, it is determined to be a valid signal. The selection of the preset threshold needs to comprehensively consider environmental factors and pipeline characteristics to reduce the false alarm rate and improve the detection accuracy. The identification of valid signals helps to filter out interference signals from non-pipeline objects, thereby improving the accuracy of automatic modeling of underground pipelines.

[0090] For example, in practical applications, suppose we use a lightweight sensor that can transmit electromagnetic wave signals with a frequency of 100MHz. When the sensor transmits the signal to the ground, the signal is reflected back by the underground pipeline. The sensor records the intensity of the reflected signal as -50dBm and the reflection time as 10 nanoseconds. According to the average propagation speed of electromagnetic waves in the soil, which is about two-thirds of the speed of light, about 2×10^8 meters per second, we can calculate that the depth of the pipeline is about 1 meter. At the same time, if the preset threshold is set to -60dBm, then since the received signal strength is greater than this threshold, it can be determined that this signal is a valid signal, further confirming the existence of the pipeline and its approximate depth.

[0091] Next, refer to Figure 3 Describe the specific steps of the flowchart that receives the reflected signal and records the signal strength and reflection time:

[0092] S301: Start the timer and begin recording time. In this stage, the system records the start time of signal transmission by transmitting a signal and starting the timer at the same time. This step ensures the accurate measurement of the subsequent reflection time.

[0093] S302: Monitor the arrival of the reflected signal. The system continuously monitors the reflected signal from the underground pipeline. Once the reflected signal is received, it proceeds to the next step to record the reflection time. This step is crucial to capture effective reflected signals.

[0094] S303: Stop the timer and record the reflection time. When the reflection signal is detected, the system immediately stops the timer and records the time from the transmission signal to the reception of the reflection signal. This is a key step in calculating the reflection time.

[0095] S304: Determine whether it is a valid reflection based on the reflection time and the preset time range. The system compares the recorded reflection time with the preset time range. If the reflection time falls within the preset time range, it is determined to be a valid reflection. This step helps to filter out non-pipeline reflection signals and improve detection accuracy.

[0096] For example, in actual applications, suppose the system transmits a radar signal for detecting underground pipelines, and the preset effective reflection time range is 0.5 to 5 milliseconds. When the system receives a reflected signal and records the reflection time as 2 milliseconds, since this time falls within the preset effective range, the system will determine that this reflection is a valid reflection, and further process this data to build an underground pipeline model.

[0097] Next, refer to Figure 4 Describe the specific steps of starting the timer and recording time in the automatic modeling method of underground pipelines based on GIS and artificial intelligence technology:

[0098] S401: Initialize the timer. In this step, the system will set a timer to record the time consumption of the entire underground pipeline automatic modeling process. This usually involves initializing the timer object or function in the software to ensure that it is ready to start counting.

[0099] S402: Set the timer to the start state. Once the timer is initialized, the next operation is to set it to the start state, which means that the timer will start recording time. This can be achieved by calling a specific API function or setting an internal flag to ensure that the timer can accurately track time.

[0100] S403: Start recording time. After setting the start state of the timer, the system will immediately start recording time. This step ensures that from this moment on, all processing time related to automatic modeling of underground pipelines is accurately recorded.

[0101] S404: Stop timing when the recorded time exceeds the preset maximum waiting time. In order to prevent indefinite time recording, the system needs to set a maximum waiting time threshold. If the recorded time exceeds this threshold, the timer will automatically stop timing. This mechanism helps to avoid unnecessary waste of resources and can be used as a means of fault detection. For example, when the system encounters a long period of unresponsiveness, timely measures can be taken.

[0102] Through the above steps, the time consumption in the automatic modeling process of underground pipelines can be effectively monitored and managed, thereby improving the efficiency and reliability of the system.

[0103] Next, refer to Figure 5Describe the specific steps of setting the timer to the start state in the automatic modeling method of underground pipelines:

[0104] S501: Setting the state flag of the timer to the start state. This operation is usually implemented through software programming, that is, defining a variable in the program as the state flag of the timer, and setting its value to a specific value indicating the start state.

[0105] S502: After setting the timer to the start state, wait for the reflected signal to return. This step involves the interaction of hardware and software, that is, the sensor or detection device sends a signal and waits for the signal to be reflected from the target object such as an underground pipeline. In practical applications, ultrasonic, radar or laser technologies can be used to send and receive reflected signals. For example, an ultrasonic probe is used to transmit a pulse signal underground, and the reflected signal is captured by the same probe or other receiving device.

[0106] S503: Read the timer status after receiving the reflected signal. When the reflected signal is successfully received, the system needs to check the status of the timer to determine the time required for the signal to go back and forth. This step can be implemented by programming, for example, by reading the value of the timer status flag to determine whether the timer is still in the start state.

[0107] S504: If the timer status flag is still in the start state, continue to wait until the signal returns. This means that if the timer is not stopped before receiving the reflected signal, that is, the status flag is not changed, the system will continue to wait until the signal returns. This logic can be implemented through a loop structure, such as using `; while timerStatus == START`; loop until the timer status changes or the signal returns. This ensures accurate measurement of the signal round trip time, thereby improving the accuracy of underground pipeline positioning.

[0108] Next, refer to Figure 6 Describe the specific steps of setting the status flag of the timer to the start state in the automatic modeling method of underground pipelines:

[0109] S601: Clear the timer status flag. In this stage, the system first needs to ensure that the timer status flag is reset to the initial state, that is, usually set its value to 0 or a specific initial value. This step is to ensure that subsequent operations can accurately reflect the state changes of the timer.

[0110] S602: Set the timer status flag to the start state. After completing the clearing operation, the system will set the timer status flag to a value indicating the start state. This value can be a binary bit, a specific value, or an option in an enumeration type, depending on the implementation of the system. In this way, the system can clearly indicate that the timer is in the start state.

[0111] S603: Monitor the change of the timer status flag. In order to ensure that the timer can correctly respond to external events or changes in internal logic, the system needs to continuously monitor the status of the timer status flag. This can be achieved through software interrupts, hardware interrupts or other forms of polling mechanisms. The purpose of monitoring is to detect any changes that may affect the timer status in a timely manner and make corresponding processing.

[0112] S604: If the timer status flag has not changed, the start state is maintained. This means that once the timer is set to the start state, the timer will continue to remain in the start state unless a specific event triggers a change in the status flag. This mechanism helps to ensure that the timer can operate stably without external intervention, thereby ensuring the continuity and accuracy of the entire underground pipeline automatic modeling process.

[0113] Suppose we are developing an embedded system for automatic modeling of underground pipelines, which uses a microcontroller as the core processor. We can use a register of the microcontroller to represent the status flag of the timer. First, when the system is initialized, the status flag is cleared by writing 0 to the register S601. Then, when the timer needs to be started, a non-zero value such as 1 is written to the register to set the status flag to the start state S602. Then, the interrupt service routine of the microcontroller is used to periodically check the value of the register to monitor the change of the status flag S603. Finally, if the value of the status flag does not change within a period of time, the system will continue to keep the timer started until a new instruction or event occurs that causes the status flag to change S604. In this way, it can be ensured that the timer can run stably without external intervention, thereby supporting the smooth progress of the automatic modeling process of underground pipelines.

[0114] Next, refer to Figure 7 Describe the specific steps to optimize the data processing flow through machine learning algorithms, reduce computing resource requirements while maintaining high accuracy:

[0115] S701: Use the preprocessing module to perform preliminary cleaning on the collected data. This process aims to remove invalid or redundant information, such as missing values, outliers, and duplicate records, to ensure the effectiveness and accuracy of subsequent analysis. For example, in actual operation, you can set a threshold to identify and remove abnormal data points that are beyond the normal range, or use interpolation to fill in missing values, thereby improving data quality.

[0116] S702: Use feature selection algorithm to select the feature set with the greatest impact on model building. This step evaluates the correlation between each feature and the target variable and selects the most relevant features for modeling to avoid noise introduced by irrelevant features. For example, a correlation coefficient-based method, such as the Pearson correlation coefficient, can be used to measure the strength of the linear relationship between the feature and the target, and the most influential feature subset can be selected accordingly.

[0117] S703: Apply lightweight model training technology to reduce computing resource consumption. In order to reduce computing costs and improve efficiency, lightweight machine learning algorithms or model compression technologies can be used. For example, pruning technology can be used to reduce the complexity of decision trees, or knowledge distillation methods can be used to transfer knowledge from large models to smaller models, thereby reducing memory usage and computing time.

[0118] S704: Adjust model parameters based on feature importance scores. During the training process, model parameters are dynamically adjusted according to the importance of features to achieve the goal of saving resources and ensuring model performance. For example, the feature importance score Feature Importance Score (FIS) in the random forest algorithm can be used. This score reflects the contribution of the feature to the prediction result. Based on these scores, model parameters can be adjusted in a targeted manner, such as increasing the weight of important features and reducing the impact of unimportant features, thereby optimizing the performance of the overall model.

[0119] Next, refer to Figure 8 Describe the specific steps of the feature selection algorithm in the automatic modeling method of underground pipelines based on GIS and artificial intelligence technology:

[0120] S801: Calculate the correlation coefficient between each feature and the target variable. This process involves statistical analysis to assess the importance of each feature by quantifying the strength of the relationship between each feature and the target variable. Specifically, statistics such as the Pearson correlation coefficient can be used to measure this linear correlation to determine which features are more likely to have a significant impact on model predictions.

[0121] S802: Set a correlation coefficient threshold T, where the feature is retained when the correlation coefficient is greater than T. This step aims to filter out features that are weakly associated with the target variable by setting a threshold, thereby reducing model complexity and improving training efficiency. For example, if the threshold T is set to 0.7, only features with a correlation coefficient greater than 0.7 with the target variable will be retained for subsequent model construction.

[0122] S803: Feature screening is performed based on the correlation coefficient threshold T, where the value range of T is between 0.5 and 0.9. This means that in actual operation, the threshold T needs to be adjusted according to the specific situation to achieve the best feature selection effect. For example, in a specific application scenario, if it is found that retaining too many features leads to overfitting of the model, the threshold T can be appropriately increased, such as set to 0.8, to further reduce the number of features; conversely, if the model performs poorly and more feature information is needed to improve the prediction ability, the threshold T can be lowered, such as set to 0.6, to increase the richness of the feature set.

[0123] Next, refer to Fig. 9 Describe the specific steps of setting the correlation coefficient threshold T in the automatic modeling method of underground pipelines:

[0124] S901: Determine a candidate threshold set C. This process involves defining a series of possible threshold options based on previous experience, theoretical analysis, or preliminary experimental results. These thresholds will be used in subsequent performance evaluation to determine which threshold can provide the best model performance.

[0125] S902: Traverse each threshold in the candidate threshold set C. This means that for each threshold in the set C, the subsequent performance evaluation step needs to be performed. This step ensures that all candidate thresholds are considered and the optimal solution can be found through comparison.

[0126] S903: Calculate the performance indicators of the model at different thresholds based on the candidate threshold set C. This usually involves using statistical methods such as cross-validation to evaluate the performance of the model at different thresholds. The performance indicator can be accuracy, recall, F1 score or other metrics suitable for specific application scenarios.

[0127] S904 is to select the threshold that makes the performance index optimal as T. After completing the performance evaluation of all candidate thresholds, those thresholds that can make the performance index optimal, such as the highest accuracy or the maximum F1 score, are selected as the final correlation coefficient threshold T. This selection ensures that the model can achieve the best performance level in practical applications.

[0128] Through the above steps, the most suitable value for the correlation coefficient threshold T can be effectively determined, thereby optimizing the performance of the automatic modeling method for underground pipelines.

[0129] Next, refer to Fig.10 , describing the specific steps of calculating the performance indicators of the model at different thresholds:

[0130] S1001: First, select candidate thresholds for evaluating the performance of the underground pipeline modeling method based on GIS and artificial intelligence technology. This step involves defining a series of thresholds that will be tested to determine the optimal threshold for the model. The selected threshold should cover a range that is likely to contain the optimal value to suit the specific application.

[0131] Next, a cross-validation technique is applied to evaluate the performance of the model at each candidate threshold. Cross-validation ensures the robustness of the evaluation by partitioning the dataset into multiple subsets, training the model on different combinations of the subsets, and testing it on the remaining data. This process is repeated for each threshold, generating a set of performance metrics for each iteration.

[0132] S1002: Calculate the average accuracy (ACC) for each threshold based on the results of the cross-validation process. ACC is the average of the accuracy scores in all cross-validation folds, providing a single metric that represents the overall performance of the model at a given threshold.

[0133] S1003: Compare the calculated average accuracy (ACC) with the predefined minimum accuracy requirement. If the ACC of a certain threshold exceeds this minimum value, then record the threshold as a potential optimal threshold candidate.

[0134] S1004: Select the best threshold T based on the recorded thresholds. Among the recorded thresholds, select the threshold that produces the highest average accuracy (ACC) as the best threshold T. This step ensures that the final model uses the threshold that provides the best balance in identifying underground pipelines.

[0135] By following these steps, the method systematically identifies the most effective thresholds for underground pipeline modeling algorithms, ensuring that the model performs optimally when deployed in real-world scenarios.

[0136] Next, refer to Fig.11 , describe the specific steps if the average accuracy ACC is higher than the predetermined minimum accuracy:

[0137] S1101: Set a minimum accuracy requirement (LAR) as a benchmark for evaluating the performance of automated modeling methods. This step involves defining a threshold that represents the minimum acceptable accuracy of the model's predictions. The LAR should be determined based on the specific application requirements and the expected reliability of the model.

[0138] S1102: Compare the average accuracy (ACC) of the model at a given threshold with the predefined LAR. This comparison is critical to determine whether the current threshold can generate a model that meets the expected accuracy. If the ACC is lower than the LAR, it indicates that the model performance does not meet the minimum standard set by the user or the system.

[0139] S1103: If ACC is greater than or equal to LAR, record the current threshold. This step ensures that the system retains the threshold that produces an acceptable accuracy model. Recording this threshold is very important for future reference and use in the final model configuration.

[0140] S1104: If ACC is lower than LAR, continue to iterate the next threshold. This iterative method allows the system to systematically search for the best threshold that maximizes the model accuracy and meets the minimum accuracy requirement.

[0141] For example, let's say we are developing a system that uses GIS and AI techniques to automatically detect and model underground pipelines. We define the LAR as 90% to ensure high confidence in the model output. During the training phase, we evaluate different thresholds for the AI ​​model. At a specific threshold, the ACC is calculated to be 92%. Since 92% is greater than or equal to the LAR of 90% that we set, we record that threshold. However, if at another threshold the ACC is only 88%, which is lower than the LAR, we will not record that threshold, but continue to test the next threshold in the sequence until we find the best threshold that satisfies the LAR.

[0142] In actual operation, when this device is used, firstly, the underground pipeline is non-invasively detected and data collected through lightweight sensors. These sensors can efficiently obtain the spatial location information of the pipeline and related attribute data, such as material, diameter, etc. Subsequently, the collected data is transmitted to the back-end processing system, where the pre-trained machine learning algorithm begins to play a role. By cleaning, classifying and pre-processing the data, the algorithm can automatically identify which data is valid and which data needs further processing or elimination. This process not only reduces unnecessary computing resource consumption, but also ensures the accuracy of subsequent modeling. Next, the geographic information system GIS is used to perform spatial analysis on the data that has been preliminarily processed. Through the powerful functions of GIS, the pipeline data can be combined with the existing topographic information to achieve accurate positioning in three-dimensional space, and on this basis, a three-dimensional model of the underground pipeline is automatically constructed. In order to ensure the accuracy and practicality of the model, the system will also automatically perform a series of verification steps, such as checking whether the relative position relationship between the pipelines in the model is reasonable and whether there are cross-overlapping problems. Finally, in order to further improve the quality of the model and reduce processing time, the system will use model simplification technology to optimize the complex pipeline structure, and use intelligent recognition algorithms to correct incorrect connections or missing parts in the model. This series of intelligent processing not only improves the overall quality of the model, but also greatly shortens the time cycle from data collection to model generation, making the entire underground pipeline automatic modeling process more efficient and reliable.

[0143] The methods, programs, systems, devices, etc. of the embodiments of the present invention may be executed or implemented in a single or multiple networked computers, or may be practiced in a distributed computing environment. In the embodiments of this specification, in these distributed computing environments, tasks may be performed by remote processing devices connected via a communication network.

[0144] Those skilled in the art should understand that the embodiments of the present specification may be provided as methods, systems or computer program products. Therefore, those skilled in the art may imagine that the functional modules / units or controllers and related method steps described in the above embodiments may be implemented in software, hardware or a combination of software / hardware.

[0145] Unless explicitly stated, the actions or steps of the methods, programs, and embodiments of the present invention do not have to be performed in a specific order and can still achieve the desired results. In some implementations, multitasking and parallel processing are also possible or may be advantageous.

[0146] In this article, multiple embodiments of the present invention are described, but for the sake of simplicity, the description of each embodiment is not exhaustive, and the same or similar features or parts between the embodiments may be omitted. In this article, an embodiment, some embodiments, examples, specific examples, or some examples are intended to be applicable to at least one embodiment or example according to the present invention, rather than all embodiments. The above terms do not necessarily mean to refer to the same embodiment or example. In the absence of mutual contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples.

[0147] The exemplary systems and methods of the present invention have been specifically shown and described with reference to the above embodiments, which are merely examples of the best modes for implementing the present systems and methods. It will be appreciated by those skilled in the art that various changes may be made to the embodiments of the systems and methods described herein when implementing the present systems and / or methods without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. The automatic modeling method of underground pipelines based on GIS and artificial intelligence technology is characterized by: include: Use lightweight sensors to collect spatial location and attribute data of underground pipelines; Optimize data processing through machine learning algorithms, reduce computing resource requirements while maintaining high accuracy; Combined with Geographic Information System (GIS) for spatial analysis, automatically build three-dimensional models and verify their accuracy; Implement model simplification techniques and intelligently identify and correct incorrect connections or missing parts to ensure the quality of the final model while reducing processing time.

2. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 1 is characterized in that: The method of collecting the spatial position and attribute data of the underground pipeline by using the lightweight sensor includes: emitting an electromagnetic wave signal of a specific frequency by using the lightweight sensor; Receive the reflected signal and record the signal strength and reflection time; Calculate the depth of underground pipelines based on reflection time; Whether it is a valid signal is determined based on the signal strength and a preset threshold, wherein when the signal strength is greater than the preset threshold, it is determined to be a valid signal.

3. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 2 is characterized in that: Receiving the reflected signal and recording the signal strength and reflection time includes: Start the timer and begin recording time; monitoring the arrival of reflected signals; Stop the timer and record the reflex time; Whether it is a valid reflection is determined according to the reflection time and a preset time range, wherein when the reflection time falls within the preset time range, it is determined to be a valid reflection.

4. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 3 is characterized in that: Start the timer and begin recording time: Initialize the timer; Set the timer to start state; Start recording time; The timing is stopped when the recording time exceeds the preset maximum waiting time, wherein the timing is stopped when the recording time exceeds the preset maximum waiting time.

5. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 4 is characterized in that: Setting the timer to start state includes: Set the timer status flag to the start state; Wait for the reflected signal to return; Read the timer status after receiving the reflected signal; If the timer status flag is still in the started state, continue to wait until the signal returns, wherein if the timer status flag is still in the started state, continue to wait until the signal returns.

6. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 5 is characterized in that: Setting the timer status flag to the start state includes: Clear the timer status flag; Set the timer status flag to the start state; Monitor the changes of the timer status flag; If the timer status flag bit does not change, the start state is maintained, wherein if the timer status flag bit does not change, the start state is maintained.

7. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 1 is characterized in that: The optimization of data processing flow by machine learning algorithm to reduce computing resource requirements while maintaining high accuracy includes: Use the preprocessing module to perform preliminary cleaning on the collected data; Use feature selection algorithm to select the feature set that has the greatest impact on model construction; Apply lightweight model training technology to reduce computing resource consumption; The model parameters are adjusted based on the feature importance score FIS, which is output by the machine learning model.

8. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 7 is characterized in that: The feature selection algorithm selects the feature set that has the greatest impact on model construction, including: calculating the correlation coefficient between each feature and the target variable; Set a correlation coefficient threshold T, where the feature is retained when the correlation coefficient is greater than T; Feature screening is performed based on the correlation coefficient threshold T, where the value range of T is between 5 and 0.

9.

9. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 8 is characterized in that: Setting the correlation coefficient threshold T includes: Determine a candidate threshold set C; Traverse each threshold in the candidate threshold set C; Calculate the performance index of the model at different thresholds based on the candidate threshold set C; The threshold that makes the performance index optimal is selected as T.

10. The automatic underground pipeline modeling method based on GIS and artificial intelligence technology according to claim 9 is characterized in that: The performance indicators of the model calculated at different thresholds include: For each candidate threshold, the model performance was evaluated using a cross-validation method; Calculate the average accuracy ACC based on the cross-validation results; If the average accuracy ACC is higher than the predetermined minimum accuracy requirement, the current threshold is recorded; The threshold value that gives the highest average accuracy rate ACC among the recorded threshold values ​​is selected as T.

Images