Introduction:
With the rapid growth of my country's economy, road construction has developed by leaps and bounds. During the use of roads after they are built, due to human factors such as the increase in traffic vehicles or natural factors such as wind and earthquakes, as well as the degradation of the properties of the materials themselves, the road's quality has deteriorated. Service life and driving safety are affected. Therefore, road inspection is a key part of road operation and maintenance management. However, the scale of road infrastructure construction is huge, which brings a huge amount of work to road maintenance. At present, road maintenance is faced with pain points such as high costs, a small number of inspection vehicles, insufficient inspection personnel, low inspection frequency, low detection accuracy, and scattered data.
On April 26, 2022, the Ministry of Transport issued the "14th Five-Year Plan Highway Maintenance Management Development Outline" (hereinafter referred to as the "Outline"). The "Outline" points out that it is necessary to "explore and promote new non-destructive testing equipment, develop and promote the application of economical and efficient automated testing equipment. Strengthen the decision-making analysis of various testing and monitoring data, and form a data-driven maintenance scientific decision-making working mechanism", and puts forward the detection and monitoring of highways. Higher requirements. According to the instructions of the "Outline", during the "14th Five-Year Plan" period, highway inspection and monitoring work in various places needs to introduce economical and non-destructive automated inspection equipment and data decision-making analysis methods to enhance the intelligence and digitalization level of the work.
Currently, highway inspection work in many areas is mainly based on manual driving inspections. Inspection tasks are completed through manual patrols, mobile phone photography, human eye recognition, on-site check-in and other methods. This method is usually accompanied by problems such as low daily inspection mileage, low disease detection efficiency, and high impact from weather. Therefore, traditional detection methods can no longer meet the requirements of the "Outline" and cannot keep up with the rapid expansion of the current road network. It is necessary to rely on advanced technology to improve highway inspection methods and achieve new breakthroughs in highway detection methods.
Industry Status:
01Positioning error of detection data in urban environment
In an urban environment, due to the influence of high-rise buildings, overpasses, tree canopies, tunnels, etc., the RTK signal is very unstable and the coordinates drift seriously, resulting in abnormal objects processed in the internal industry, but the accurate location cannot be found on the ground; secondly, the position of the radar image after drift Overlap leads to more "false exceptions" in internal processing. In addition, inaccurate positioning can also lead to:
① Reduce the efficiency of field data collection;
② Position correlation and spatial and temporal analysis cannot be done between survey lines.
02The test parameters are single and the test radar frequency is single
①The parameter is single and cannot provide more auxiliary information for abnormal identification;
②The radar frequency is single and cannot take into account the resolution of shallow signals and the effectiveness of deep signals. .
03Difficult to manage numbers
①Data processing and anomaly identification, purely manual mode, too inefficient;
②Lack of unified management platform, lack of coordination between municipal, district and sub-district offices, possible duplication Construction situation, a waste of financial funds;
③For the same project, between different construction entities, whether it is original data, processed data, disease information cards, results reports, etc., the formats are the same Without unification, data management is difficult.
GIT-RDM full life cycle road disease detection platform
Guangdong Ground and Air Intelligent Technology Co., Ltd. GIT-RDM full life cycle road disease detection platform integrates satellite navigation + inertial navigation combined positioning technology to solve the industry pain points of unstable RTK signals and poor positioning accuracy in urban field data collection, and high-speed driving road testing The average actual error is less than 20 cm. The round-trip survey line data is spliced, segmented and three-dimensionally modeled according to roads. Massive data is fully automatically processed and numbered and archived according to road sections, allowing for long-term management. The platform supports the mobile phone to download the list of abnormal objects to be retested and verified and assign it to the field work team. The verification results and image data are uploaded to the cloud platform in real time through the mobile terminal, realizing a closed loop of hidden danger detection and verification.
The platform is compatible with all mainstream 3D radars in the industry and enables fully automated processing of the massive data generated by 3D radars. Data display, abnormal identification and annotation, retesting and verification realize the entire process online, and construction results can be exported with one click, which greatly reduces the internal workload and greatly improves construction efficiency and project benefits. Multi-parameter and multi-dimensional road data are bound to precise positioning information to achieve spatio-temporal correlation of survey line data. Use AI technology to identify and model infrastructure such as pipelines, identify hidden dangers such as cavities, and identify high-risk areas to achieve full life cycle detection and management of road hazards and solve the problem of unclear distribution of hidden risks, unknown causes of hidden dangers, and post-event Dealing with issues such as passivity.
Platform features
01Transportation work amount
All data can be collected at one time on-site through the intelligent data collection module built into the tablet computer. The online model realizes fully automated processing of massive data without manual intervention, which not only improves data collection efficiency but also reduces labor costs.
02Manipulator humanization
The data entry interface is simple and clear, easy to operate and supports functions such as zooming in and out, providing an underground space information interface for smart cities.
03One-click summary of data
One-click operation is performed when uploading data on the PC. All data, clear drawings, and graphical reports are formed at once, avoiding repeated entry and no need for secondary adjustments.
04Comprehensive testing data
Centralized and unified management of all elements and full life cycle data provides a foundation for massive training of AI models and analysis of high-risk road sections. This platform is used in road inspection work, including the complete workflow of the road inspection business and a set of related reports. The platform is highly integrated.
05Realize data sharing
With multi-role division and collaboration in collection, processing, identification and verification, managers can view data anytime and anywhere through mobile devices, realizing data sharing within their authority.
Road disease detection process
01New project and task assignment
Create a new project based on the project task list, improve the project information, add tasks, and assign the tasks to the corresponding person in charge.
02Scenery exploration
An on-site survey is carried out at the beginning of the project operation. On-site survey data and image data are uploaded through the field App. The platform generates a survey report with one click and plans the detection scene based on the on-site survey information.
03Radar data collection
The global navigation satellite system + inertial navigation positioning method is used, combined with vehicle-mounted three-dimensional ground penetrating radar, to quickly detect diseased objects around the pipeline. Real-time splicing of three-dimensional radar measurement harnesses, real-time uploading of positioning trajectories and radar data to the platform. Add supervisor supplementary testing and branch pipe detection tasks through the platform, use radar for data collection, and use the App to collect field data and upload data and image records.
04Fully automatic modeling of massive data
The online model enables fully automated processing of massive data. The round-trip survey line data is spliced, segmented, three-dimensionally modeled, and numbered and archived by road. Divide it into 25-meter, 50-meter or 100-meter-long blocks along the road direction, breaking the massive data into parts for easy management; the block angle is dynamically adjusted with the road to ensure that the survey line view is placed horizontally from left to right to the maximum extent. Comply with data analysis habits; use blocks as units for task management, identification, verification and long-term monitoring;
05Road anomaly identification
The data is downloaded to the local disk for identification processing according to road conditions, so that the working position can be quickly read without any requirements and can be operated remotely; the platform combines high-precision location information to perform spatial correlation calculations and model reconstruction of all elements, based on the full space The model's underground facility identification and disease hazard abnormal data are screened, filtered, evaluated and analyzed, and detailed data information of each road disease hazard point is identified and marked.
06Exception point retest verification
The entire process of field retest and verification can be used to view abnormal points and upload data and image data through the App, and to quickly compare the abnormal points of on-site retest verification, and quickly view cross-sectional images and different horizontal slices through the App. As well as the timely upload of endoscopic videos and operating images, after the disease verification is completed, the risk of the disease can be assessed on-site, achieving paperless operations, multi-role collaborative work, and reducing clerical errors and improper storage of paper documents. The recorded data is missing or distorted due to the following conditions.
07Output of road disease results
The platform exports radar line maps, disease distribution maps, disease information cards, results reports and related results data with one click, allowing relevant managers to intuitively understand the details of road diseases, understand road performance conditions at a glance, and take appropriate measures in a timely manner based on road disease information. Carry out repair and maintenance according to the measures taken to realize intelligent daily inspection. This not only extends the service life of the road, but also saves overall road maintenance costs.
