Intelligent identification method and system for luminous performance of road surface self-luminous material
By combining a drone platform with multimodal sensing technology and deep learning algorithms, efficient and accurate performance monitoring of self-luminous pavement materials has been achieved, solving the identification problems of unstable brightness and complex environments, and supporting automated maintenance decisions.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-25
- Publication Date
- 2026-06-19
AI Technical Summary
In existing technologies, the luminescence performance of self-luminous pavement materials is easily affected by environmental factors, resulting in unstable brightness. Furthermore, there is a lack of efficient and reliable performance monitoring and evaluation methods, making it difficult to achieve large-scale, high-frequency, and objective state perception, and the recognition accuracy is low in complex environments.
A drone platform was used to simultaneously collect luminous images of the road surface and lidar point cloud data. Combined with the deep learning U-Net model and RANSAC algorithm, pixel-level segmentation and three-dimensional geometric feature analysis were performed to generate a comprehensive evaluation report and identify the causes of luminous performance degradation.
It achieves 24/7 automated inspection, improves inspection efficiency by 10 times, provides highly objective inspection results, accurately diagnoses the causes of performance degradation, supports forward-looking maintenance decisions, and has an identification accuracy rate of over 95%.
Smart Images
Figure CN122238276A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent operation and maintenance and non-destructive testing technology for intelligent transportation infrastructure, and in particular to an intelligent identification method and system for the luminescence performance of self-luminous materials on road surfaces. Background Technology
[0002] With the deepening of the "carbon peaking and carbon neutrality" strategy, the green and low-carbon transformation of transportation infrastructure has become an inevitable requirement for the industry's development. As an important carrier of energy consumption, energy conservation and emission reduction during the operation phase of road infrastructure are particularly crucial. Currently, urban roads mainly rely on traditional street lighting, which suffers from high construction costs, high operating energy consumption, and high maintenance costs; while rural roads, due to their remote locations and weak power supply infrastructure, generally lack lighting facilities, posing significant traffic safety hazards.
[0003] To address these challenges, self-luminous pavement materials have emerged. These materials absorb natural or ambient light during the day and autonomously release visible light at night through a long afterglow effect, providing drivers with continuous visual guidance and significantly improving the safety and guidance of nighttime driving. However, the practical application of these materials still faces two major technological bottlenecks: First, the luminescent properties of the material itself are easily affected by various factors. Environmental conditions and material aging can lead to unstable brightness and shortened afterglow time, making it difficult to objectively and quantitatively assess performance degradation. Second, there is a lack of efficient and reliable performance monitoring and evaluation methods. Currently, on-site assessment of paved self-luminous pavements almost entirely relies on manual nighttime inspections. This method suffers from low efficiency, strong subjectivity, and poor data traceability, making it difficult to achieve large-scale, high-frequency, and objective condition perception, severely restricting the large-scale application of the material and precise maintenance decisions.
[0004] While some intelligent detection systems have emerged in the field, such as road surface defect detection systems based on vehicle platforms or unmanned inspection vehicles, these systems generally focus on identifying physical structural defects such as cracks and potholes, without addressing the identification and quantitative assessment of the functional luminescence state of materials. Furthermore, in real road environments, specular reflections from rainwater film on the road surface, interference from roadside lights or vehicle taillights, and other sporadic artificial light sources can severely affect image acquisition quality. This makes it difficult for existing methods to effectively distinguish between ambient reflected light, interfering light sources, and the material's effective self-luminous signals, resulting in a significant decrease in accuracy and insufficient reliability in complex environments.
[0005] In summary, the industry currently lacks a technical solution capable of automatically, accurately, and efficiently identifying the luminescent properties of self-luminous pavement materials, while also being able to adapt to complex environmental interference. Therefore, this invention proposes an intelligent identification system and method for the luminescent properties of self-luminous pavement materials based on UAV-borne multimodal sensing. Summary of the Invention
[0006] In view of the above problems, this disclosure provides an intelligent identification method and system for the luminescence performance of self-luminous pavement materials to overcome or at least partially solve the above problems. The purpose is to solve the problems of low efficiency and insufficient accuracy caused by relying on manual detection in the prior art, break through the difficulty of extracting effective luminescence information in complex environments, and provide core technical support for realizing intelligent management and long-term performance evaluation of self-luminous pavement.
[0007] The objective of this invention can be achieved through the following technical solutions: The first aspect of the present invention provides an intelligent identification method for the luminescent properties of self-luminous road surface materials, comprising the following steps: S101. Simultaneously collect luminous images of the road surface, lidar point cloud data, and carrier pose data through a drone platform; wherein, when collecting the luminous images, a camera equipped with a ring polarizing filter is used to suppress specular reflection light. S102. Process the luminous image to obtain the luminous brightness distribution information of the road surface; S103. Process the lidar point cloud data and extract the three-dimensional geometric feature information of the road surface; S104. Perform spatiotemporal registration and fusion analysis on the luminous intensity distribution information and the three-dimensional geometric feature information, and diagnose the causes of luminous performance decay based on preset diagnostic rules to obtain diagnostic results. S105. Based on the diagnostic results, generate a comprehensive evaluation report on the self-luminous performance of the road surface.
[0008] further, The preset diagnostic rules include: if the luminance value of a certain area is lower than the first threshold and the road surface elevation value of that area is lower than the second threshold, it is determined that physical wear has caused performance degradation; if the luminance value of a certain area is lower than the first threshold, but the road surface elevation value of that area is normal, it is determined that material aging has caused performance degradation.
[0009] further, In step S102, processing the luminescent image includes: using a semantic segmentation algorithm based on a deep learning U-Net model to perform pixel-level segmentation of the image to obtain effective luminescent regions; and calculating a quantized brightness index for the effective luminescent regions, wherein the quantized brightness index includes at least one of average grayscale value, brightness uniformity index, and effective luminescent area ratio; the average grayscale value is the arithmetic mean of the grayscale values of all pixels within the target region, and the average grayscale value is expressed as:
[0010] Where N is the total number of pixels in the region. Let be the grayscale value of the i-th pixel; The brightness uniformity index assesses the consistency of luminescence in the region; the brightness uniformity index is expressed as:
[0011] in, It is the standard deviation of the brightness in that area; The closer the value is to 1, the better the uniformity.
[0012] further, The processing of the luminescent image in step S102 further includes: performing fusion processing on multiple frames of images continuously acquired in the same area, and dynamically determining the brightness threshold for segmentation based on the global statistical characteristics of the image.
[0013] further, In step S103, processing the lidar point cloud data includes: segmenting the ground point cloud using the RANSAC algorithm to remove non-road surface point clouds; gridding the road surface point cloud and calculating the elevation value of each grid cell to identify road surface depression areas; the average elevation value... Represented as:
[0014] Where n is the number of points falling into the grid. It is the average elevation value of the point. This represents the elevation value of the j-th point.
[0015] further, In step S101, the acquisition is achieved by synchronously triggering the optical camera and lidar mounted on the UAV; and between steps S101 and S102, there is also a real-time warning step, including: during the flight of the UAV, the acquired single-frame image is quickly analyzed by the airborne edge computing unit, and if it is found that the overall brightness of the image is lower than the alarm threshold, a real-time alarm signal is immediately issued.
[0016] Furthermore, in step S104, the fusion analysis also incorporates visible light images collected during the day for damage identification; the identified road surface damage locations are correlated with the nighttime luminous intensity distribution information to establish a mapping relationship between physical damage and luminous performance degradation; wherein, the mapping relationship is expressed as: .
[0017] A second aspect of the present invention provides an intelligent identification system for the luminescent properties of self-luminous road surface materials, comprising: Intelligent carrier mobile platform for industrial-grade drones equipped with RTK-GNSS modules and IMU; The multimodal sensing system, integrated on the UAV, includes a high-sensitivity global shutter camera for acquiring images of road surface illumination, and a multi-line lidar for acquiring three-dimensional point clouds of the road surface. The core of data processing and intelligent recognition includes an edge computing unit deployed on the UAV itself and a remote cloud server; the edge computing unit is used for image preprocessing and real-time anomaly warning, and the cloud server is used to run image segmentation, point cloud processing and multi-source data fusion diagnostic algorithms. The evaluation result output module is used to receive the comprehensive evaluation report and display it visually.
[0018] Furthermore, In an intelligent identification system for the luminescence performance of self-luminous road surface materials, the RTK-GNSS module and IMU are used to provide spatiotemporal synchronization information with centimeter-level accuracy for the luminescent images and point cloud data.
[0019] In an intelligent recognition system for the luminescence properties of self-luminous road surface materials, a high-sensitivity global shutter camera is a scientific-grade CMOS camera with a quantum efficiency of over 70% in the 500-600nm wavelength band.
[0020] In an intelligent recognition system for the luminescence properties of self-luminous road surface materials, a high-sensitivity global shutter camera is equipped with an adjustable-angle ring polarizing filter.
[0021] In an intelligent recognition system for the luminescence properties of self-luminous road surface materials, the edge computing unit incorporates a lightweight neural network model for rapid preliminary image recognition.
[0022] A third aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed by one or more processors, cause the processors to perform the intelligent identification method for the luminescent properties of self-luminous road surface materials as described in the first aspect.
[0023] The technical solution proposed in this application can bring the following beneficial effects: 1. The method described in this invention uses drones to replace manual labor, enabling 24 / 7 automated inspection. The inspection efficiency is more than 10 times higher than that of manual labor, greatly reducing labor costs and safety risks.
[0024] 2. The method described in this invention features a fully automated data acquisition and processing flow, ensuring highly objective and repeatable test results. All data are accompanied by precise spatiotemporal labels, facilitating the establishment of long-term digital archives of pavement performance and providing a solid data foundation for tracking material performance degradation trends and evaluating maintenance effectiveness.
[0025] 3. The method described in this invention innovatively integrates luminescence characteristics with road surface geometry and visual defects through multi-source information fusion. It can not only determine "whether it emits light" but also diagnose "why it does not emit light", enabling in-depth analysis of the causes of performance degradation and supporting forward-looking maintenance decisions.
[0026] 4. The system described in this invention employs a lightweight deep learning model (such as U-Net), which is not simply applied but specifically optimized for the characteristics of road surface luminescence images in low illumination and complex backgrounds. It achieves a leap from "human eye observation" to pixel-level precise quantization, with a stable recognition accuracy of over 95% for luminescence brightness and uniformity, avoiding the fluctuations and uncertainties of subjective judgment.
[0027] 5. The system solution described in this invention is complete, with close integration of software and hardware, and can be directly applied to the intelligent management and maintenance upgrade of existing road facilities.
[0028] The above description is merely an overview of the technical solution disclosed herein. In order to better understand the technical means of this disclosure and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this disclosure more apparent and understandable, specific embodiments of this disclosure are described below. Attached Figure Description
[0029] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the scope of this disclosure. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a flowchart illustrating the steps of an intelligent identification method for the luminescent properties of self-luminous road surface materials provided in the embodiments of this specification. Figure 2 This is a schematic diagram of the structure of an intelligent identification system for the luminescence performance of self-luminous road surface materials provided in the embodiments of this specification; Figure 3 This is a flowchart illustrating the intelligent identification method for the luminescent properties of self-luminous road surface materials provided in the embodiments of this specification. Figure 4 This is a schematic diagram illustrating the effect of the luminance recognition algorithm provided in the embodiments of this specification, wherein (a) represents the original road surface luminance image; (b) represents the grayscale distribution image; (c) represents the luminance heat map; and (d) represents the defect detection result image. Figure 5 This is a schematic diagram illustrating the principle of wear assessment using multi-source data fusion provided in the embodiments of this specification. Detailed Implementation
[0030] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art. The technical solutions provided by various embodiments of this application will be described in detail below with reference to the accompanying drawings.
[0031] This application provides a system and method for the automatic identification and evaluation of the luminescent performance of self-luminescent road surface materials, enabling large-scale, high-efficiency, high-precision, and quantifiable assessment. It aims to address the core problems of low efficiency, poor accuracy, and lack of automated functional monitoring methods in manual inspection, achieving automated, intelligent, and quantitative evaluation of the self-luminescent performance of road surfaces. Specifically, it solves the problems of low signal-to-noise ratio and difficulty in identification of luminescent images due to weak illumination and background interference in complex nighttime environments; it achieves precise quantification of luminescent performance (brightness, uniformity), rather than merely qualitative judgment; and it enables preliminary diagnosis of performance degradation, distinguishing whether brightness reduction is due to physical wear or chemical aging of the material, providing a basis for precise maintenance decisions.
[0032] The first aspect of the technical solution of the present invention provides an intelligent identification method and system for the luminescence performance of road surface self-luminous materials. The system includes an intelligent carrier mobile platform, a multimodal sensing system, a data processing and intelligent identification core, and an evaluation result output module.
[0033] The method includes: simultaneously acquiring road surface luminescence images, 3D point clouds, and high-precision pose data via UAVs; providing real-time anomaly warnings at the edge; segmenting luminescent regions and quantifying brightness indicators using a U-Net model in the cloud; processing point cloud data to extract road surface geometric features; and finally generating a health map that integrates spatial location, performance level, and causal inference. Figure 1 As shown, it includes the following steps: S101. Simultaneously collect luminous images of the road surface, LiDAR point cloud data, and carrier pose data via a UAV platform. When collecting the luminous images, a camera equipped with a ring polarizer is used to suppress specular reflection. The inspection area, path, and time schedule are set on the control platform, and the UAV initiates autonomous inspection according to the plan. During the inspection, the UAV flies autonomously according to the plan, with the optical camera, LiDAR, and RTK-GNSS / IMU simultaneously triggered to collect luminous images of the road surface, 3D point clouds, and high-precision pose data, respectively. All data includes a unified timestamp and spatial coordinates.
[0034] S102. Process the luminous image to obtain the luminous brightness distribution information of the road surface.
[0035] During drone flight intervals, the edge computing unit performs rapid calculations on the captured single-frame images. A lightweight neural network method is used to determine if there are any severe brightness anomalies. If an anomaly is detected, a real-time alarm mechanism is immediately triggered; simultaneously, regardless of whether an alarm is triggered, all raw data is continuously uploaded to the cloud.
[0036] S103. Process the lidar point cloud data and extract the three-dimensional geometric feature information of the road surface.
[0037] S104. Perform spatiotemporal registration and fusion analysis on the luminous intensity distribution information and the three-dimensional geometric feature information, and diagnose the causes of luminous performance degradation based on preset diagnostic rules to obtain diagnostic results.
[0038] S105. Based on the diagnostic results, generate a comprehensive evaluation report on the self-luminous performance of the road surface.
[0039] In a preferred embodiment of the present invention, a 5-kilometer-long section of cement-based long afterglow luminescent material (mainly composed of SrAl2O4:Eu) is used. 2+ , Dy 3+ Taking a main urban road where a water film forms after rain, easily causing specular reflection, as an example, the test was conducted in complete darkness, two hours after sunset. A Matrice 350 RTK drone equipped with an adjustable ring polarizing filter and a D-RTK 2 high-precision GNSS mobile station was used to achieve real-time centimeter-level positioning. A DJI Zenmuse P1 full-frame aerial survey camera with a 35mm lens was used. The settings were: ISO 1600, exposure time 0.5 seconds, and global shutter mode to avoid rolling shutter effect. A DJI Zenmuse L1 LiDAR, integrated into the same drone platform, was triggered synchronously with the camera. The U-Net architecture based on the PyTorch framework was used. This model was trained using 10,000 labeled images of luminous road surfaces to achieve pixel-level segmentation of luminous areas. The PCL library was used for point cloud denoising, ground segmentation, and meshing, with a mesh size of 1cm × 1cm. A preset brightness dimming threshold was used. Road surface wear threshold The first threshold is the brightness dimness threshold. The second threshold is the road surface wear threshold. .
[0040] further, The preset diagnostic rules include: if the luminance value of a certain area is lower than the first threshold and the road surface elevation value of that area is lower than the second threshold, it is determined that physical wear has caused performance degradation; if the luminance value of a certain area is lower than the first threshold, but the road surface elevation value of that area is normal, it is determined that material aging has caused performance degradation.
[0041] If a region appears bright in an optical image, but point cloud data shows that the region has a depression, then the region is marked as "suspected reflective interference," and its brightness value is not included in the self-luminous performance evaluation. If a bright region in a nighttime image has an artificial light source (such as a street lamp pole) identified at the corresponding position in a daytime visible light image, then the nighttime bright region is determined to be "artificial light source interference" and is removed.
[0042] During data acquisition: In the mission planning software, the UAV's flight altitude was set to 30 meters, its flight speed to 5 m / s, its forward overlap rate to 80%, and its lateral overlap rate to 70%. The UAV flew automatically along the preset route, with GNSS, IMU, P1 camera, and L1 lidar strictly synchronized according to the system synchronization signal (PPS). The following data was collected: high-resolution luminous images of the road surface and high-precision 3D laser point clouds (point density up to 160 points / square meter); median filtering was applied to three consecutively acquired images from the same geographical location to effectively eliminate random high-light noise caused by raindrops or temporary puddles; the precise latitude, longitude, altitude, and attitude angle corresponding to each image frame and each group of point clouds were recorded. Simultaneously, brightness segmentation thresholds were set. Instead of being globally fixed, the brightness is dynamically calculated based on the global brightness distribution of the currently acquired batch of images (taking the 90th percentile of all pixel grayscale values as the benchmark) to adapt to the possible changes in overall brightness after rain.
[0043] further, In step S102, processing the luminous image includes: performing grayscale conversion, contrast enhancement, and normalization on the acquired luminous image in the cloud to improve image quality. A semantic segmentation algorithm based on the deep learning U-Net model is used to segment the image at the pixel level to obtain effective luminous regions; and a quantized brightness index of the effective luminous regions is calculated. The quantized brightness index includes at least one of the following: average grayscale value, brightness uniformity index, and effective luminous area ratio. The average grayscale value is the arithmetic mean of the grayscale values of all pixels within the target area (e.g., road segment K2+300 to K2+350), expressed as:
[0044] Where N is the total number of pixels in the region. Let be the grayscale value of the i-th pixel; in this embodiment, it is calculated as follows: =15, significantly lower than the overall average of 85; The brightness uniformity index assesses the consistency of luminescence in the region; the brightness uniformity index is expressed as:
[0045] in, It is the standard deviation of the brightness in that area; The closer the value is to 1, the better the uniformity.
[0046] further, The processing of the luminescent image in step S102 further includes: performing fusion processing on multiple frames of images continuously acquired in the same area, and dynamically determining the brightness threshold for segmentation based on the global statistical characteristics of the image.
[0047] further, In step S103, processing the LiDAR point cloud data includes: denoising and filtering the LiDAR point cloud, then segmenting the ground point cloud using the RANSAC algorithm to remove non-road surface point clouds, such as vegetation and vehicles; meshing or voxelizing the road surface point cloud (mesh size set to 1cm×1cm), calculating the elevation value of each mesh cell to identify road surface depression areas; calculating the elevation value, surface roughness, or curvature of each mesh cell, and identifying areas with continuous negative deviations or abnormal undulations by analyzing these geometric features. Specifically, after completing the point cloud meshing, the system traverses each mesh and accumulates the elevation values of all points within it. Then divide by the total number of points n in the grid to get the average elevation value.
[0048] Among them, the average elevation value Represented as:
[0049] Where n is the number of points falling into the grid. It is the average elevation value of the point. This represents the elevation value of the j-th point, usually the Z-coordinate. For example... Figure 4 As shown in the figure, the analysis reveals a continuous low-lying area in the dimly lit section from K2+300 to K2+350, with an average elevation of Average elevation of the surrounding reference area The difference, i.e., the wear depth .
[0050] further, In step S101, the acquisition is achieved by synchronously triggering the optical camera and lidar mounted on the UAV; and between steps S101 and S102, there is also a real-time warning step, including: during the flight of the UAV, the acquired single-frame image is quickly analyzed by the airborne edge computing unit, and if it is found that the overall brightness of the image is lower than the alarm threshold, a real-time alarm signal is immediately issued.
[0051] Furthermore, such as Figure 5As shown, in step S104, the fusion analysis also incorporates visible light images collected during the day for damage identification; the identified road surface damage locations are correlated with the luminous brightness distribution information at night to establish a mapping relationship between physical damage and luminous performance degradation. Specifically, the high-precision timestamps and spatial coordinates provided by RTK-GNSS and IMU are used to fuse the obtained luminance recognition results with the three-dimensional point cloud data acquired by lidar; by analyzing the geometric features of areas with weak or no luminance, a correlation model between luminance attenuation and physical wear is established. The mapping relationship is expressed as follows: .
[0052] If a certain area emits dim light, and the corresponding point cloud shows obvious geometric depressions at that location (such as wear depth exceeding a set threshold), it is determined that the performance degradation is caused by physical wear. If the geometric smoothness of the area is good, it is determined that it is caused by chemical aging of the material.
[0053] In this embodiment, since the data in this region simultaneously satisfies =15<30 and =5.2mm>3mm, the system automatically determines that the performance degradation at this point is due to "physical wear".
[0054] By integrating brightness, uniformity, geometric, and disease information, a weighted scoring model is used to calculate the overall health score for each region. Based on the health score, K-means clustering analysis is employed for performance grading, automatically generating a comprehensive visual report that includes a brightness distribution heatmap, defect location markers, and health score levels, while also providing maintenance recommendations.
[0055] Finally, the system automatically generates a "luminous performance health map" for the road section, visually displaying the results in the form of a heat map, and outputs a comprehensive report that includes the specific location, quantitative data, cause diagnosis, and maintenance recommendations such as "severe wear, priority milling and repaving is recommended".
[0056] In this embodiment, data collection for a 5-kilometer road section was completed in approximately 18 minutes, and data processing and analysis on the cloud server took approximately 5 minutes, all without human intervention. The system successfully identified the area from K2+300 to K2+350 as a performance degradation zone and accurately diagnosed the cause as physical wear, consistent with the results of subsequent on-site manual inspection.
[0057] To further enhance detection robustness in complex environments such as after rain and with sporadic light sources, this invention also includes targeted optimizations at the hardware, algorithm, and data fusion levels. In terms of hardware, the camera is equipped with an adjustable ring polarizing filter to suppress specular reflection. In terms of algorithms, multi-frame fusion technology and dynamic threshold segmentation are employed to effectively overcome water film reflection and transient specular interference, ensuring the accuracy of brightness information extraction. In terms of data fusion, high-precision image registration is achieved, and the diagnostic rules incorporate cases marked as potentially causing reflective interference or artificial light source interference, eliminating such data.
[0058] A second aspect of the present invention provides an intelligent identification system for the luminescent properties of self-luminous road surface materials, comprising: Intelligent carrier mobile platform for industrial-grade drones equipped with RTK-GNSS modules and IMU; The multimodal sensing system, integrated on the UAV, includes a high-sensitivity global shutter camera for acquiring images of road surface illumination, and a multi-line lidar for acquiring three-dimensional point clouds of the road surface. The core of data processing and intelligent recognition includes an edge computing unit deployed on the UAV itself and a remote cloud server; the edge computing unit is used for image preprocessing and real-time anomaly warning, and the cloud server is used to run image segmentation, point cloud processing and multi-source data fusion diagnostic algorithms. The evaluation result output module is used to receive the comprehensive evaluation report and display it visually.
[0059] This example provides an intelligent identification system for the luminescence performance of self-luminous road surface materials. The intelligent mobile carrier is a drone. The multimodal sensing system is divided into a main sensor and an auxiliary sensor. The main sensor is a multispectral camera, and the auxiliary sensors are lidar, RTK-GNSS modules, and IMU. The core of data processing and intelligent identification is an edge measurement unit set up inside the system and a cloud server set up at a remote location. The evaluation result output module is equipped with a wireless communication module and a data visualization interface.
[0060] The intelligent carrier mobile platform utilizes a drone, which serves as the mobile foundation and carries all subsequent sensing and processing modules, enabling flexible movement within the detection area. Employing an industrial-grade drone equipped with an RTK-GNSS module and a high-precision IMU ensures flight stability and centimeter-level positioning accuracy. It can autonomously inspect according to preset paths or commands, providing a stable benchmark for image acquisition and guaranteeing the accuracy of material luminescence identification and positioning in different areas.
[0061] The multimodal sensing system is integrated into the intelligent carrier mobile platform, including a main sensor. This main sensor is a high-sensitivity, global shutter scientific-grade CMOS camera paired with an adjustable ring polarizer. Its quantum efficiency should be higher than 70% in the 500-600nm wavelength band to adapt to low-light environments and avoid motion blur. Since self-emitting light is weak, the monochrome camera eliminates the Bayer filter, resulting in significantly higher photosensitivity than a color camera, capturing more photons. The global shutter prevents rolling shutter effect during carrier movement, ensuring image geometric accuracy. The ring polarizer, by adjusting the polarization angle, effectively suppresses specular reflections from road surface water films, improving the image signal-to-noise ratio to cope with special environments. Auxiliary sensors include a lidar, an RTK-GNSS module, and an IMU. The lidar uses a multi-line lidar to acquire high-precision 3D point cloud data of the road surface, assisting in the analysis of road surface smoothness and geometric wear, and features low cost and small data volume. The RTK-GNSS and IMU combination provides high-precision carrier position and attitude information for data stitching and precise defect localization.
[0062] The core of the data processing and intelligent recognition system includes an edge computing unit deployed on the carrier itself and a remote cloud server. The edge computing unit, deployed on the drone, performs image preprocessing and real-time preliminary identification and analysis of luminous areas, enabling real-time alarms for severe anomalies such as large-area light loss. The cloud server receives all sensor data, runs complex intelligent recognition algorithms, performs multi-source data fusion and in-depth analysis, and generates a comprehensive evaluation report.
[0063] The assessment result output module includes a wireless communication module and a data visualization interface. This module is responsible for transmitting the identification results (such as brightness distribution maps, health scores, and defect location markers) to the control center and visualizing them on a digital map, providing an intuitive basis for maintenance decisions.
[0064] Furthermore, In an intelligent identification system for the luminescence performance of self-luminous road surface materials, the RTK-GNSS module and IMU are used to provide spatiotemporal synchronization information with centimeter-level accuracy for the luminescent images and point cloud data.
[0065] In an intelligent recognition system for the luminescence properties of self-luminous road surface materials, a high-sensitivity global shutter camera is a scientific-grade CMOS camera with a quantum efficiency of over 70% in the 500-600nm wavelength band.
[0066] In an intelligent recognition system for the luminescence properties of self-luminous road surface materials, a high-sensitivity global shutter camera is equipped with an adjustable-angle ring polarizing filter.
[0067] In an intelligent recognition system for the luminescence properties of self-luminous road surface materials, the edge computing unit incorporates a lightweight neural network model for rapid preliminary image recognition.
[0068] A third aspect of the present invention provides a computer-readable storage medium having instructions stored thereon, which, when executed by one or more processors, cause the processors to perform the intelligent identification method for the luminescent properties of self-luminous road surface materials as described in the first aspect.
[0069] This embodiment can divide the method into functional modules based on the above method example. For example, each function can be assigned to a separate module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware. It should be noted that the module division in this embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.
[0070] When each function is divided into modules corresponding to its specific function, it may include: a data acquisition module, a signal preprocessing module, an adaptive decomposition module, a signal reconstruction module, a feature extraction module, a fault diagnosis module, etc. It should be noted that all relevant content of each step involved in the above method embodiments can be referenced from the functional descriptions of the corresponding modules, and will not be repeated here.
[0071] This embodiment also provides a computer-readable storage medium (including but not limited to disk storage, CD-ROM, optical storage, etc.) storing computer program code. When the computer program code is run on a computer, the computer executes the above-mentioned related method steps to realize the intelligent identification method and system for the luminescence performance of road surface self-luminous materials provided in the above embodiment.
[0072] This embodiment also provides a computer program product. When the computer program product is run on a computer, it causes the computer to perform the aforementioned steps to realize the intelligent identification method and system for the luminescent properties of self-luminous road surface materials provided in the above embodiment. The beneficial effects of the above embodiments can be found in the corresponding methods described above, and will not be repeated here.
[0073] Through the above description of the embodiments, those skilled in the art will understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above.
[0074] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of modules or units is merely a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms. In the description of this disclosure, it should be understood that if terms such as "upper," "lower," "front," "rear," "left," and "right" are used to indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, they are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the indicated position or element must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of this disclosure.
[0075] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes the element.
[0076] The above are merely embodiments of this disclosure and are not intended to limit the scope of this disclosure. Various modifications and variations can be made to this disclosure by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of the claims of this disclosure.
Claims
1. A method for intelligently identifying the luminescent properties of self-luminous road surface materials, characterized in that, Includes the following steps: S101. Simultaneously collect luminous images of the road surface, lidar point cloud data, and carrier pose data through an unmanned aerial vehicle (UAV) platform. S102. Process the luminous image to obtain the luminous brightness distribution information of the road surface; S103. Process the lidar point cloud data and extract the three-dimensional geometric feature information of the road surface; S104. Perform spatiotemporal registration and fusion analysis on the luminous intensity distribution information and the three-dimensional geometric feature information, and diagnose the causes of luminous performance decay based on preset diagnostic rules to obtain diagnostic results. S105. Based on the diagnostic results, generate a comprehensive evaluation report on the self-luminous performance of the road surface.
2. The intelligent identification method for the luminescent properties of self-luminous road surface materials according to claim 1, characterized in that, The preset diagnostic rules include: if the luminance value of a certain area is lower than the first threshold and the road surface elevation value of that area is lower than the second threshold, it is determined that physical wear has caused performance degradation; if the luminance value of a certain area is lower than the first threshold, but the road surface elevation value of that area is normal, it is determined that material aging has caused performance degradation.
3. The intelligent identification method for the luminescent properties of self-luminous road surface materials according to claim 1, characterized in that, In step S102, processing the luminescent image includes: using a semantic segmentation algorithm based on a deep learning U-Net model to perform pixel-level segmentation of the image to obtain effective luminescent regions; and calculating a quantized brightness index for the effective luminescent regions, wherein the quantized brightness index includes at least one of average grayscale value, brightness uniformity index, and effective luminescent area ratio; the average grayscale value is the arithmetic mean of the grayscale values of all pixels within the target region, and the average grayscale value is expressed as: Where N is the total number of pixels in the region. Let be the grayscale value of the i-th pixel; The brightness uniformity index assesses the consistency of luminescence in the region; the brightness uniformity index is expressed as: in, It is the standard deviation of the brightness in that area.
4. The intelligent identification method for the luminescent properties of self-luminous road surface materials according to claim 1, characterized in that, The processing of the luminescent image in step S102 further includes: performing fusion processing on multiple frames of images continuously acquired in the same area, and dynamically determining the brightness threshold for segmentation based on the global statistical characteristics of the image.
5. The intelligent identification method for the luminescent properties of self-luminous road surface materials according to claim 1, characterized in that, In step S103, processing the lidar point cloud data includes: segmenting the ground point cloud using the RANSAC algorithm to remove non-road surface point clouds; gridding the road surface point cloud and calculating the elevation value of each grid cell to identify road surface depression areas; the average elevation value... Represented as: Where n is the number of points falling into the grid. It is the average elevation value of the point. This represents the elevation value of the j-th point.
6. The intelligent identification method for the luminescent properties of self-luminous road surface materials according to claim 1, characterized in that, In step S101, the acquisition is achieved by synchronously triggering the optical camera and lidar mounted on the UAV; and between steps S101 and S102, there is also a real-time warning step, including: during the flight of the UAV, the acquired single-frame image is quickly analyzed by the airborne edge computing unit, and if it is found that the overall brightness of the image is lower than the alarm threshold, a real-time alarm signal is immediately issued.
7. The intelligent identification method for the luminescent properties of self-luminous road surface materials according to claim 1, characterized in that, In step S104, the fusion analysis also incorporates visible light images collected during the day for damage identification; the identified road surface damage locations are correlated with the nighttime luminous intensity distribution information to establish a mapping relationship between physical damage and luminous performance degradation.
8. An intelligent identification system for the luminescent properties of self-luminous road surface materials, characterized in that, include: Intelligent carrier mobile platform for industrial-grade drones equipped with RTK-GNSS modules and IMU; The multimodal sensing system, integrated on the UAV, includes a high-sensitivity global shutter camera for acquiring images of road surface illumination, and a multi-line lidar for acquiring three-dimensional point clouds of the road surface. The core of data processing and intelligent recognition includes edge computing units deployed on the drone itself and remote cloud servers; The edge computing unit is used for image preprocessing and real-time anomaly warning, and the cloud server is used to run image segmentation, point cloud processing and multi-source data fusion diagnostic algorithms. The evaluation result output module is used to receive the comprehensive evaluation report and display it visually.
9. The intelligent identification system for the luminescent properties of self-luminous road surface materials according to claim 8, characterized in that, The RTK-GNSS module and IMU are used to provide spatiotemporal synchronization information with centimeter-level accuracy for the luminescent images and point cloud data.
10. A computer-readable storage medium, characterized in that, It stores instructions that, when executed by one or more processors, cause the processors to perform the intelligent identification method for the luminescent properties of road surface self-luminescent materials as described in any one of claims 1-7.