Highway disease automatic identification method based on big data and cloud computing
By deploying sensors along the highway and using drone inspection technology, combining big data and cloud computing technology, automatic identification and early warning of highway diseases are achieved, the problems of inefficient and low accuracy of traditional inspections are solved, and the scientificity and efficiency of highway maintenance are improved.
Patent Information
- Application Number
- CN202510050588.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-13
- Publication Date
- 2025-05-30
AI Technical Summary
Traditional highway disease inspections rely on manual observation, are inefficient and have low accuracy, making it difficult to meet the high standards for highway maintenance.
The automatic identification method of highway disease based on big data and cloud computing is adopted to collect multi-source data through sensors deployed along the highway, and comprehensive data is obtained using drone inspection technology to perform data preprocessing, feature extraction, disease identification and classification, distribution analysis and early warning decisions.
It has achieved all-round and deep perception of highway diseases, improved the accuracy and efficiency of disease identification, and the recognition accuracy rate can reach more than 95%, significantly improving the scientificity and efficiency of highway maintenance.
Smart Images
Figure CN120070127A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of intelligent transportation and highway maintenance, and specifically relates to an automatic highway disease recognition method based on big data and cloud computing. Background Art
[0002] As a key foundation of the transportation system, the safety and durability of highways are directly related to the smoothness of traffic and the travel safety of citizens. However, various diseases will inevitably occur during the use of highways, such as cracks, potholes, spalling, etc. These diseases not only affect the flatness of the highway and the driving comfort, but also may pose safety hazards to vehicles and pedestrians. The traditional highway disease inspection mode mainly relies on direct on-site observation and manual recording by workers. This method is not only inefficient and consumes a large amount of human resources, but also is extremely vulnerable to the subjective judgment and experience level of the inspection personnel. Due to individual differences and fatigue and other factors, the inspection results are often inconsistent, resulting in large errors and omissions in the process of discovering, identifying, and diagnosing highway diseases. According to incomplete statistics, it takes at least several weeks or even longer to complete a complete inspection cycle for a comprehensive disease detection on a 100-kilometer-long highway using the traditional inspection method. Moreover, even at such an inspection intensity, the accuracy rate of disease identification is often less than 80%, which is difficult to meet the high standards of highway maintenance requirements. Summary of the Invention
[0003] The purpose of the present invention is to provide an automatic highway disease recognition method based on big data and cloud computing to solve the problems raised in the above background art.
[0004] To achieve the above purpose, the present invention provides the following technical solution: An automatic highway disease recognition method based on big data and cloud computing, and the steps of the automatic recognition method are as follows: Step 1, multi-source data collection; use various sensors deployed along the highway to collect image data, vibration signals, and displacement change information related to highway diseases; at the same time, with the help of the drone inspection technology means, obtain more comprehensive highway disease data, including areas that are difficult to reach or observe; Step 2, data preprocessing; on the cloud computing platform, perform preprocessing operations such as data cleaning, denoising, and format unification on the collected data to ensure the quality and consistency of the data; Step 3, feature extraction: use edge detection and texture analysis image processing technologies to extract features from the image data and identify the characteristic information of road surface cracks and pothole diseases; perform time-frequency analysis on the sensor data to extract vibration characteristics and displacement change information related to diseases, and combine multi-source data to construct a comprehensive feature vector to provide input for subsequent disease recognition; Step 4, disease recognition and classification: Based on the extracted feature vectors, a classification model is constructed to automatically identify and classify highway diseases; the classification model is trained and verified using the cross-validation method to improve the accuracy and generalization ability of the identification, and the identification results are verified and calibrated to ensure the accuracy of the identification; Step Five, Disease Distribution Analysis: Utilize the parallel computing power of the cloud computing platform to conduct spatial distribution analysis on the identification results to generate a disease distribution map; combine geographic information system technology to display the location, scope, and development trend information of the diseases, and then conduct statistical analysis on the disease distribution to mine the laws and characteristics of the diseases, providing a scientific basis for highway maintenance; Step Six, Early Warning and Decision Support: According to the disease identification results and distribution analysis, set an early warning threshold to automatically generate disease early warning information; push the early warning information to the highway management department through the mobile terminal or Web platform to assist it in making timely and effective maintenance decisions, providing maintenance suggestions and optimization plans to guide the implementation of highway maintenance work.
[0005] Preferably, the cleaning process in the data preprocessing includes detecting outliers in the data, and the formula for outlier detection is: Outlier detection = where, is the data point, is the mean value, is the standard deviation, is the threshold; utilize big data processing technology to efficiently store and manage massive data, providing a basis for subsequent analysis.
[0006] Preferably, the step four includes the evaluation of the recognition accuracy rate, and the formula for the evaluation of the recognition accuracy rate is: Accuracy rate = × 100% Preferably, the sensors deployed along the highway include but are not limited to cameras, vibration sensors, and displacement sensors.
[0007] Preferably, in the feature extraction of step three, it also includes: during the feature extraction process, a convolutional neural network deep learning model is simultaneously adopted to automatically learn the key features in the image, facilitating the time-frequency analysis of the sensor data.
[0008] Preferably, the classification models in the disease identification and classification of step four include but are not limited to support vector machine SVM, random forest, and deep learning models to automatically identify and classify highway diseases.
[0009] Preferably, the cross-validation method of the classification model includes: Step One, Data Division: Divide the original data set into a training set and a test set, which are used to estimate the model parameters, and the test set is used to evaluate the model accuracy; Step 2, training the model: Use the training set data to train the model; Step 3, testing the model: Test the trained model on the test set to evaluate its performance; Step 4, repeated verification: Repeat the above steps for different data partitions to obtain a more reliable model performance evaluation.
[0010] Preferably, this method is applied to various types of roads, including disease inspection and early warning of expressways, national and provincial trunk roads, and rural roads.
[0011] Preferably, both the data collection frequency and the early warning threshold are dynamically adjusted according to historical data and road usage conditions to ensure the timeliness and representativeness of the data.
[0012] The beneficial effects of the present invention are as follows: 1. Through a variety of sensors deployed at key positions on the road, the system of the present invention can collect multi-source data in real time and accurately, including but not limited to road surface images, vibration signals, displacement changes, etc., thereby achieving an all-round and in-depth perception of road diseases. With the help of advanced data processing and analysis technologies, the system can accurately identify various road diseases, providing a scientific basis for road maintenance.
[0013] 2. The present invention makes full use of the efficient data storage and processing capabilities of the cloud computing platform to easily handle the challenges of processing massive multi-source data. The parallel computing and distributed storage technologies of the cloud computing platform make the data processing process faster and more efficient, while ensuring the accuracy and integrity of the data. This efficient data processing method not only improves the overall operation efficiency of the system, but also provides a solid data foundation for subsequent road disease identification and early warning.
[0014] 3. The present invention adopts industry-leading machine learning algorithms and image processing technologies to achieve intelligent identification and fine classification of road diseases. The system can automatically identify and distinguish different types of road diseases, such as cracks, potholes, ruts, etc., greatly improving the accuracy and reliability of identification. In addition, combined with spatial distribution analysis technology, the system can also provide accurate disease early warning information, providing scientific decision-making support for road management departments. Compared with traditional inspection methods, the identification accuracy of the present invention can be as high as over 95%, and the inspection efficiency is increased by several times, injecting new vitality into the intelligent management of road maintenance.
[0015] 4. The method of the present invention exhibits extremely high scalability and flexibility. The system can flexibly adjust and optimize key components such as algorithm models and data processing flows according to the actual road type and disease characteristics requirements. This flexible configuration method enables the system to easily adapt to various complex and changeable road environments, providing a more comprehensive and efficient solution for the intelligent management of road maintenance. Brief Description of the Drawings
[0016] Figure 1 It is a schematic diagram of the steps of the automatic road disease recognition method of the present invention. Detailed Embodiments
[0017] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.
[0018] As Figure 1 shown, the embodiments of the present invention provide an automatic road disease recognition method based on big data and cloud computing. The steps of the automatic recognition method are as follows: Step 1, multi-source data collection; use various sensors deployed along the road to collect image data, vibration signals, and displacement change information related to road diseases; at the same time, with the help of unmanned aerial vehicle inspection technology means, obtain more comprehensive road disease data, including areas that are difficult to reach or observe; the data collection frequency can be dynamically adjusted according to road usage conditions and disease historical data to ensure the timeliness and representativeness of the data.
[0019] Step 2, data preprocessing; perform preprocessing operations such as cleaning, denoising, and format unification on the collected data on the cloud computing platform to ensure the quality and consistency of the data; Step 3, feature extraction: use edge detection and texture analysis image processing technologies to extract features from the image data, and identify the feature information of road surface cracks and pothole diseases; perform time-frequency analysis on the sensor data to extract vibration characteristics and displacement change information related to diseases, and combine multi-source data to construct a comprehensive feature vector to provide input for subsequent disease recognition. Step 4, disease recognition and classification: Based on the extracted feature vector, construct a classification model to automatically recognize and classify road diseases; the classification model is trained and verified using the cross-validation method to improve the accuracy and generalization ability of recognition, and the recognition results are verified and calibrated to ensure the accuracy of recognition; Step 5, Disease Distribution Analysis: Utilize the parallel computing power of the cloud computing platform to conduct spatial distribution analysis on the recognition results, generate a disease distribution map; combine with Geographic Information System (GIS) technology to display the location, scope, and development trend information of the diseases, and then conduct statistical analysis on the disease distribution to explore the laws and characteristics of the diseases, providing a scientific basis for highway maintenance; Step 6, Early Warning and Decision Support: Based on the disease recognition results and distribution analysis, set an early warning threshold, and automatically generate disease early warning information; the early warning threshold can be dynamically adjusted according to historical data and highway usage conditions. Push the early warning information to the highway management department through the mobile terminal or Web platform to assist it in making timely and effective maintenance decisions, providing maintenance suggestions and optimization plans, and guiding the implementation of highway maintenance work.
[0020] Among them, the cleaning process in data preprocessing includes detecting outliers in the data, and the formula relied on for outlier detection is: Outlier detection = Among them, is the data point, is the mean value, is the standard deviation, is the threshold; utilize big data processing technology to efficiently store and manage massive data, providing a basis for subsequent analysis.
[0021] Among them, the recognition accuracy evaluation is included in Step 4, and the recognition accuracy evaluation formula is: Accuracy = × 100% Among them, the sensors deployed along the highway include but are not limited to cameras, vibration sensors, and displacement sensors.
[0022] Among them, in the feature extraction of Step 3, it also includes: simultaneously adopt a Convolutional Neural Network (CNN) deep learning model during the feature extraction process to automatically learn the key features in the image, facilitating time-frequency analysis of the sensor data.
[0023] Among them, the classification models in Step 4, disease recognition and classification, include but are not limited to Support Vector Machine (SVM), random forest, and deep learning models, for automatically recognizing and classifying highway diseases.
[0024] Among them, the cross-validation methods of the classification model include: Step 1, Data Division: Divide the original data set into a training set and a test set (or multiple subsets) to estimate the model parameters, and the test set is used to evaluate the model accuracy; Step 2, Model Training: Use the training set data to train the model; Step 3, Test the model: Test the trained model on the test set to evaluate its performance; Step 4, Repeat verification: Repeat the above steps for different data partitions to obtain a more reliable model performance evaluation.
[0025] Among them, this method is applied to various types of roads, including the disease inspection and early warning of expressways, national and provincial trunk roads, and rural roads.
[0026] Among them, the data collection frequency and early warning threshold are both dynamically adjusted according to historical data and road usage conditions to ensure the timeliness and representativeness of the data.
[0027] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.
[0028] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A method for automatic identification of highway diseases based on big data and cloud computing, characterized by: The steps of the automatic identification method are: Step 1: Multi-source data collection: Use various sensors deployed along the highway to collect image data, vibration signals, and displacement change information related to highway diseases; At the same time, use drone inspection technology to obtain more comprehensive highway disease data, including areas that are difficult to reach or difficult to observe; Step 2: Data preprocessing: The collected data is cleaned, denoised, and formatted in a unified manner on the cloud computing platform to ensure data quality and consistency. Step 3: Feature extraction: Use edge detection and texture analysis image processing technology to extract features from image data and identify the characteristic information of road cracks and potholes. Perform time-frequency analysis on sensor data to extract vibration characteristics and displacement change information related to the disease. Combine multi-source data to construct a comprehensive feature vector to provide input for subsequent disease identification. Step 4: Disease identification and classification: Based on the extracted feature vectors, a classification model is constructed to automatically identify and classify highway diseases. The classification model is trained and verified using a cross-validation method to improve the accuracy and generalization of recognition, and the recognition results are verified and calibrated to ensure the accuracy of recognition. Step 5: Disease distribution analysis: Utilize the parallel computing capability of the cloud computing platform to perform spatial distribution analysis on the identification results and generate a disease distribution map; combine geographic information system technology to display the location, scope, and development trend of the disease, and then conduct statistical analysis on the disease distribution to explore the laws and characteristics of the disease and provide a scientific basis for highway maintenance; Step 6: Early warning and decision support: According to the disease identification results and distribution analysis, set the early warning threshold and automatically generate disease early warning information; The warning information is pushed to the highway management department through mobile terminals or web platforms to assist them in making timely and effective maintenance decisions, provide maintenance suggestions and optimization plans, and guide the implementation of highway maintenance work.
2. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized by: The cleaning process in the data preprocessing includes outlier detection on the data. The outlier detection is based on the formula: Outlier Detection = in, is the data point, is the mean, is the standard deviation, As the threshold; using big data processing technology, massive data can be efficiently stored and managed to provide a basis for subsequent analysis.
3. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized by: The step 4 includes recognition accuracy evaluation, and the recognition accuracy evaluation formula is: Accuracy = ×100%.
4. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized in that: The sensors deployed along the highway include but are not limited to cameras, vibration sensors, and displacement sensors.
5. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized by: The feature extraction in step three also includes: using a convolutional neural network deep learning model in the feature extraction process to automatically learn key features in the image to facilitate time-frequency analysis of sensor data.
6. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized by: The classification models in the step 4 of disease identification and classification include but are not limited to support vector machines (SVM), random forests, and deep learning models, which automatically identify and classify highway diseases.
7. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized in that: The cross-validation method of the classification model includes: Step 1: Divide the data: Divide the original data set into a training set and a test set to estimate the model parameters. The test set is used to evaluate the model accuracy. Step 2: Training the model: Use the training set data to train the model; Step 3: Test the model: Test the trained model on the test set to evaluate its performance. Step 4: Repeat the validation: Repeat the above steps for different data partitions to obtain a more reliable evaluation of the model performance.
8. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized by: This method is applied to disease inspection and early warning of various types of roads, including expressways, national and provincial trunk roads, and rural roads.
9. The method for automatic identification of highway defects based on big data and cloud computing according to claim 1 is characterized by: The data collection frequency and warning threshold are dynamically adjusted based on historical data and road usage to ensure the timeliness and representativeness of the data.