A highway pavement detection and evaluation system

Abstract

The application discloses a highway pavement detection and evaluation system and relates to the technical field of intelligent detection, which comprises a detection module, a compensation module and an evaluation module, a first detection point is arranged, a first detection result is obtained through the first detection, a second detection point is arranged, a second detection result is obtained through the second detection, a first compensation point is obtained, a first compensation result is obtained through repeated first detection, the pavement state is evaluated, and a pavement state grade is obtained. The double positioning reduces the false positive rate, the secondary retesting reduces the depth error, the image detection completes the preliminary screening of most areas, the ultrasonic wave is only focused on the key areas, the labor and equipment costs are reduced, the sound wave speed database and the detection parameters can be updated through the cloud, the image module can upgrade the AI model, the ultrasonic wave module can be compatible with multiple modes such as the shear wave and the surface wave, and is suitable for different surface layer materials and climate conditions.

Description

Technical Field

[0001] This invention relates to the field of intelligent detection technology, and more particularly to a highway pavement detection and evaluation system. Background Technology

[0002] In recent years, highway pavement inspection has developed rapidly. Based on lightweight models such as YOLO, CenterNet, and RoadID, real-time identification of defects such as cracks, potholes, and network cracks has been achieved. It has evolved from two-dimensional image recognition to three-dimensional reconstruction. Line structured light and digital cameras have become new directions for low-cost and lightweight technologies. It has evolved from laser profilers to vehicle vibration inversion. New sensors such as smartphones and cushion-type accelerometers have lowered the threshold. By optimizing the flight path through annealing algorithms, efficient and automated slope defect inspection has been achieved.

[0003] Currently, Chinese invention patent CN120182625A discloses a method and system for detecting road surface defects. This method obtains a first weight and a second weight through a weight allocation model; acquires a first initial feature set corresponding to a real-time road surface image and a second initial feature set corresponding to a real-time road surface infrared image; performs feature complementation on the first and second initial feature sets to convert them into a first final feature set and a second final feature set, respectively; and obtains the final image to be used through the first final feature set, the second final feature set, the first weight, and the second weight. Based on illumination characteristics, the distribution of light and shadow is identified, forming a weighted combination between real-time road surface images and real-time road surface infrared images. This adaptively eliminates interference from external environmental factors, improving the accuracy of defect identification and reducing the error rate. However, related technologies do not perform preliminary road surface detection based on intuitive road surface images, nor do they perform refined detection based on the preliminary detection results. This is not conducive to the comprehensiveness of road surface condition data acquisition, nor to the speed of data processing. Furthermore, the lack of collaborative analysis between coarse and refined detection results hinders the scientific rigor and accuracy of road surface condition assessment, thus presenting certain limitations. Summary of the Invention

[0004] The technical problem solved by this invention is that related technologies do not perform preliminary detection of the road surface based on intuitive images of the road surface, and do not perform refined detection based on the preliminary detection results. This is not conducive to the comprehensiveness of obtaining road surface condition data, nor is it conducive to the speed of data processing. Furthermore, the lack of collaborative analysis between coarse and refined detection results is not conducive to the scientificity and accuracy of road surface condition assessment, and thus has certain limitations.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a highway pavement detection and evaluation system, comprising a detection module, a compensation module, and an evaluation module;

[0006] The detection module sets a first detection point and a first detection parameter based on the road surface image, performs a first detection at the first detection point, and obtains a first detection result.

[0007] The compensation module sets a second detection point and a second detection parameter based on the first detection result, performs a second detection at the second detection point to obtain a second detection result, obtains a first compensation point based on the second detection result, and repeats the first detection at the first compensation point to obtain a first compensation result.

[0008] The evaluation module assesses the road surface condition based on the first and second detection results to obtain the road surface condition level.

[0009] As a preferred embodiment of the highway pavement detection and evaluation system of the present invention, the logic for setting the first detection point based on the pavement image includes:

[0010] The road surface image is represented as a highway surface image. A standard crack image is retrieved, and the first feature quantity of the standard crack image is extracted. The second feature quantity of the highway surface image is also extracted. Both the first and second feature quantities are represented as shape features. The first similarity between the first and second feature quantities is calculated using the cosine similarity formula. The first value is set as the similarity threshold. The first similarity at each second feature quantity of the highway surface image is iterated. The first similarity is compared with the first value. When the first similarity is less than the first value, the corresponding second feature quantity is deleted and the process moves to the next second feature quantity. When the first similarity is greater than or equal to the first value, the corresponding second feature quantity is retained.

[0011] Set the geometric center point of the region where each of the retained second features is located as the first detection point.

[0012] In a preferred embodiment of the highway pavement inspection and evaluation system of the present invention, the first detection parameter includes ultrasonic amplitude and ultrasonic frequency, and the setting logic of the first detection parameter includes:

[0013] The method involves acquiring the road filling material, average temperature, and average humidity, where the average temperature represents the daily average temperature of the area where the road is located, and the average humidity represents the daily average humidity of the area where the road is located. It also involves retrieving a detection database, inputting the road filling material, average temperature, and average humidity into the database, matching the historical ultrasonic amplitude and historical ultrasonic frequency corresponding to the road filling material, average temperature, and average humidity, calculating the first average value of each ultrasonic amplitude, and calculating the second average value of the historical ultrasonic frequency. Finally, the first average value and the second average value are set as the corresponding ultrasonic amplitude and ultrasonic frequency in the first detection parameters.

[0014] In a preferred embodiment of the highway pavement inspection and evaluation system of the present invention: after the first inspection parameters are set, a start inspection signal is sent to the inspection device at the first inspection point. After receiving the start inspection signal, the inspection device starts to emit ultrasonic waves perpendicular to the horizon at the corresponding first inspection point. After a first time period, a first inspection result is obtained, which is expressed as the echo duration.

[0015] As a preferred embodiment of the highway pavement inspection and evaluation system of the present invention, the method for setting the second inspection point based on the first inspection result includes:

[0016] Select any first detection point, obtain the echo duration at the detection point, retrieve the sound velocity database, input the road filling material into the sound velocity database, match the sound velocity corresponding to the road filling material, the sound velocity is represented as the speed at which the ultrasonic wave travels in the road filling material, retrieve the crack depth formula, input the sound velocity and echo duration into the crack depth formula, and calculate the crack depth at the first detection point according to the crack depth formula.

[0017] By iterating through the echo durations at each first detection point, the crack depth at each first detection point is obtained.

[0018] Sort the crack depths in ascending order. When the crack depths are the same, keep only one crack depth and delete the others to obtain an increasing numerical sequence with no duplicate values. Obtain the first detection point corresponding to the numerical sequence and set the maximum crack depth of the corresponding first detection point as the second detection point.

[0019] In a preferred embodiment of the highway pavement inspection and evaluation system described in this invention, the second detection parameter includes ultrasonic direction, ultrasonic amplitude, and ultrasonic frequency. The setting logic for ultrasonic amplitude and ultrasonic frequency is the same as that for the first detection parameter. The ultrasonic direction includes all directions within 360 degrees parallel to the horizontal plane.

[0020] After the second parameter is set, a second detection is performed at the second detection point to obtain the second detection result, which is represented as the echo amplitude.

[0021] As a preferred embodiment of the highway pavement inspection and evaluation system described in this invention, the logic for obtaining the first compensation point based on the second inspection result includes:

[0022] Connect any two second detection points to obtain a first connecting line. Traverse each connecting line and set the direction of the ultrasonic wave parallel to each connecting line as the first direction. Obtain any echo amplitude value at the first direction.

[0023] The echo amplitude is plotted as an acoustic curve, with time on the horizontal axis and echo amplitude on the vertical axis. The number of extreme points of the acoustic curve is counted and recorded as the first number. The number of second detection points in the first direction is counted and recorded as the second number. The first number and the second number are compared. When the first number equals the second number, the process jumps to the next acoustic curve in the first direction. When the first number is greater than the second number, the first distance between each extreme point and the starting point is calculated according to the acoustic travel distance formula. The second distance between each second detection point of the first connecting line and the starting point is obtained. The position of the extreme point where the second distance is greater than the first distance is set as the first compensation point.

[0024] Obtain any echo value not located in the first direction, plot the echo value as an acoustic wave curve, and when the acoustic wave curve does not have an extreme value, jump to the next ultrasonic wave direction. When the acoustic wave curve has an extreme value, set the position of each extreme value point as the first compensation point.

[0025] In a preferred embodiment of the highway pavement inspection and evaluation system of the present invention, the first inspection is repeated at the first compensation point to obtain a first compensation result, wherein the first compensation result is expressed as the crack depth at the first compensation point.

[0026] As a preferred embodiment of the highway pavement detection and evaluation system of the present invention, the system includes: obtaining a first detection result and a second detection result, evaluating the pavement condition, obtaining an evaluation value, and classifying the pavement condition level according to the evaluation value. The pavement condition level includes a first level, a second level, and a third level, and the pavement condition represented by the first level, the second level, and the third level is progressively better.

[0027] The calculation logic for the evaluation value includes:

[0028] Calculate the first sum of each first detection result and the second detection result, and set the first sum as the evaluation value.

[0029] As a preferred embodiment of the highway pavement inspection and evaluation system of the present invention, the logic for classifying the pavement condition levels includes:

[0030] The second and third values ​​are set as grade classification values. The evaluation value is compared with the grade classification value. When the evaluation value is less than or equal to the second value, the road surface condition level is set to the third level. When the evaluation value is greater than the second value and less than or equal to the third value, the road surface condition level is set to the second level. When the evaluation value is greater than the third value, the road surface condition level is set to the first level.

[0031] The beneficial effects of this invention are as follows: First, a road surface image is used for rapid screening across the entire area, followed by point-by-point verification using ultrasound in suspected areas. This dual positioning reduces the false alarm rate. Depth is calculated using a sound velocity database and echo duration, and a second measurement is performed at the deepest crack, further reducing depth error and meeting the millimeter-level accuracy requirements for maintenance decisions. Image detection completes the initial screening of most areas, while ultrasound focuses only on key areas, reducing manpower and equipment costs. The first and second detection results are mapped to road surface condition levels, directly connecting to the highway maintenance management system to achieve seamless integration of disease classification, cost calculation, and construction order dispatch. The sound velocity database and detection parameters can be updated via the cloud, the image module can be upgraded with an AI model, and the ultrasound module is compatible with multiple modes such as shear waves and surface waves, adapting to different surface materials and climatic conditions. Attached Figure Description

[0032] Figure 1 This is a basic flowchart of a highway pavement detection and evaluation system provided in one embodiment of the present invention. Detailed Implementation

[0033] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0034] Example, refer to Figure 1 As an embodiment of the present invention, a highway pavement detection and evaluation system is provided, including a detection module, a compensation module and an evaluation module;

[0035] The detection module sets a first detection point and first detection parameters based on the road surface image, performs a first detection at the first detection point, and obtains a first detection result.

[0036] The compensation module sets a second detection point and a second detection parameter based on the first detection result, performs a second detection at the second detection point to obtain a second detection result, obtains a first compensation point based on the second detection result, and repeats the first detection at the first compensation point to obtain a first compensation result.

[0037] The assessment module evaluates the road surface condition based on the first and second detection results to obtain the road surface condition level.

[0038] This invention first uses road surface images for rapid full-area screening, then uses ultrasound to verify each point in suspicious areas. This dual positioning reduces the false alarm rate. Depth is calculated using a sound velocity database and echo duration, and a second measurement is performed at the deepest crack, further reducing depth error and meeting the millimeter-level accuracy requirements for maintenance decisions. Image detection completes the initial screening of most areas, while ultrasound focuses only on key areas, reducing manpower and equipment costs. The first and second detection results are mapped to road surface condition levels, directly connecting to the highway maintenance management system to achieve seamless integration of defect classification, cost calculation, and construction order dispatch. The sound velocity database and detection parameters can be updated via the cloud, the image module can be upgraded with an AI model, and the ultrasound module is compatible with multiple modes such as shear waves and surface waves, adapting to different surface materials and climatic conditions.

[0039] In practical implementation, the traditional "edge + grayscale" algorithm generated 1829 false alarms per km. This system, based on shape-cosine similarity filtering, reduced the false alarms to 52 per km, a decrease of 97.2%. Manual core drilling verified 376 cracks, with the system missing only 6, a miss rate of 1.6%. The width of all missed cracks was <0.2mm, below the 0.3mm intervention threshold in maintenance standards, which is acceptable. After segmentation of the original image, 12438 "suspicious blocks" were generated. After filtering with a similarity threshold of 0.75, only 1011 first detection points were retained, a reduction of 91.9%. The vehicle inspection speed increased from 30km / h to 75km / h, and the total time decreased from 114 minutes to 46 minutes. The average time for shape feature extraction and cosine similarity comparison of 4K images (3840×2160px) is 11ms. On the NVIDIA Jetson AGXXavier vehicle terminal, it can process 6 cameras in parallel, meeting the real-time requirement of 100Hz frame rate. Using RTK-GNSS as the ground truth, the average plane error of the center point is ±8cm, which meets the 20cm diameter coupling window requirement of the subsequent ultrasonic probe. Based on the logic of setting the first detection point based on shape-cosine similarity, the number of subsequent ultrasonic probes is reduced by an order of magnitude while ensuring that no cracks are missed. This realizes a new mode of highway pavement crack detection: "large-scale rapid screening + small-scale accurate retesting".

[0040] The first detection parameters include ultrasonic amplitude and ultrasonic frequency. The setting logic for the first detection parameters includes:

[0041] The average temperature and average humidity of the road filling material are obtained. The average temperature is represented by the daily average temperature of the area where the road is located, and the average humidity is represented by the daily average humidity of the area where the road is located. The detection database is retrieved and the road filling material, average temperature, and average humidity are input into the detection database. The historical ultrasonic amplitude and historical ultrasonic frequency corresponding to the road filling material, average temperature, and average humidity are matched. The first average value of each ultrasonic amplitude is calculated, and the second average value of the historical ultrasonic frequency is calculated. The first average value and the second average value are set as the corresponding ultrasonic amplitude and ultrasonic frequency in the first detection parameters, respectively.

[0042] After the first detection parameters are set, a start detection signal is sent to the detection device at the first detection point. After receiving the start detection signal, the detection device starts to emit ultrasonic waves perpendicular to the horizon at the corresponding first detection point. After the first time period, the first detection result is obtained, which is represented as the echo duration.

[0043] In practice, after setting the amplitude / frequency based on historical averages, the standard deviation of echo duration σt decreased from 8.7µs to 3.1µs (↓64%), and the depth back-calculation error decreased from ±7.4mm to ±2.6mm (↓65%). At a fixed frequency of 200kHz, the high temperature in summer caused asphalt softening, resulting in poor coupling and a success rate of only 71%. After adopting an adaptive frequency (dynamically adjusted from 170 to 260kHz) based on "filling material + temperature and humidity," the first-time coupling success rate increased to 94%. Traditional fixed parameters require 2-3 retransmissions and parameter adjustments, with a single-point time of 4.8s. This system directly calls historical average parameters. The single-point detection time was reduced to 1.9s (↓60%), the overall 64km detection time was reduced from 10.2h to 4.7h, the transmission voltage was reduced from a fixed 180Vpp to an "average-adaptive" 110~150Vpp, the average power consumption was reduced from 38W to 22W (↓42%), the on-board UPS battery life was extended by 1.8h, and the amplitude and frequency were pre-matched by querying historical averages based on three factors: "filling material + daily average temperature + daily average humidity". Without increasing the sensor hardware, the random error of the echo duration can be reduced by more than 60%, which significantly improves the repeatability of crack depth and detection efficiency.

[0044] The methods for setting a second detection point based on the first detection result include:

[0045] Select any first detection point, obtain the echo duration at the detection point, retrieve the sound velocity database, input the highway filling material into the sound velocity database, match the sound velocity corresponding to the highway filling material, the sound velocity is represented as the speed at which the ultrasonic wave travels in the highway filling material, retrieve the crack depth formula, input the sound velocity and echo duration into the crack depth formula, and calculate the crack depth at the first detection point according to the crack depth formula.

[0046] By iterating through the echo durations at each first detection point, the crack depth at each first detection point is obtained.

[0047] Sort the crack depths in ascending order. When the crack depths are the same, keep only one crack depth and delete the others to obtain an increasing numerical sequence with no duplicate values. Obtain the first detection point corresponding to each numerical sequence and set the maximum crack depth of the corresponding first detection point as the second detection point.

[0048] In practical implementation, the traditional "single-point maximum echo" strategy has an average depth deviation of 6.8mm (σ=4.1mm), while this "non-repeating incremental sequence maximum value" strategy has an average depth deviation of 1.9mm (σ=1.3mm), improving accuracy by 72%. Compared with the true value of the drill core, the traditional method has κ=0.61, while this method has κ=0.92 (close to 1.0 for complete consistency). Each second detection point automatically generates a QR code with coordinates, depth, and confidence level. Scanning the code allows viewing the original echo image and drill core photo, achieving 100% traceability. The cloud database contains a total of 3.2GB of original waveforms, supporting AI retraining, and the system accuracy continuously iterates with mileage.

[0049] The second detection parameters include ultrasonic direction, ultrasonic amplitude, and ultrasonic frequency. The setting logic for ultrasonic amplitude and ultrasonic frequency is the same as that for the first detection parameter. The ultrasonic direction includes all directions within 360 degrees parallel to the horizontal plane.

[0050] After the second parameter is set, a second detection is performed at the second detection point to obtain the second detection result, which is expressed as the echo amplitude.

[0051] In practice, by expanding the ultrasonic direction to a 360° horizontal scan, the system no longer measures only a single "vertical depth," but simultaneously captures the crack's direction, bifurcation, and end extension on the horizontal plane. Maintenance units can thus directly obtain a "crack plan view" without secondary measurements. The vertical direction provides depth, and the horizontal direction provides length and angle; these two values ​​are superimposed to create a three-dimensional crack skeleton in real time in the background. Subsequent measurements, whether for grouting volume estimation or milling area marking, do not require manual re-measurement because the material, temperature, and humidity have already been determined in the previous step. The second test directly reuses the same amplitude and frequency, eliminating the need for on-site calibration. This avoids error drift caused by parameter changes and reduces operational complexity. Some surface scratches or joints may generate echoes in a single vertical direction, but the echo amplitude decays rapidly with angle during a 360° horizontal scan. The system utilizes this difference to automatically eliminate false cracks, reducing false alarms.

[0052] The logic for obtaining the first compensation point based on the second detection result includes:

[0053] Connect any two second detection points to obtain a first connecting line. Traverse each connecting line and set the direction of the ultrasonic wave parallel to each connecting line as the first direction. Obtain any echo amplitude value at the first direction.

[0054] The echo amplitude is plotted as a sound wave curve, with time on the horizontal axis and echo amplitude on the vertical axis. The number of extreme points of the sound wave curve is counted and recorded as the first number. The number of second detection points in the first direction is counted and recorded as the second number. The first number and the second number are compared. When the first number equals the second number, the process jumps to the next sound wave curve in the first direction. When the first number is greater than the second number, the first distance between each extreme point and the starting point is calculated according to the sound wave travel distance formula. The second distance between each second detection point on the first connecting line and the starting point is obtained. The position of the extreme point where the second distance is greater than the first distance is set as the first compensation point.

[0055] Obtain any echo value not located in the first direction, plot the echo value as an acoustic wave curve, and jump to the next ultrasonic wave direction when the acoustic wave curve does not have an extreme value. When the acoustic wave curve has an extreme value, set the position of each extreme value point as the first compensation point.

[0056] The first test is repeated at the first compensation point to obtain the first compensation result, which is represented as the crack depth at the first compensation point.

[0057] In practical implementation, by scanning in a 360° direction and automatically interpolating at the mismatch between the number of extreme values ​​and the number of detection points, the system can map out all the secondary cracks, micro-bifurcations, and internal cavities that are invisible to the naked eye, forming a complete crack network skeleton. This avoids secondary cracking due to missed patching in the later stages. Traditional methods rely on manual experience to decide "whether to densify"; this logic uses extreme value-distance comparison to automatically determine the densification location, achieving adaptive point layout of "dense in complex crack areas and sparse in simple areas," which neither wastes probes nor misses dangerous areas. The first compensation point happens to fall in the high-stress area where the crack extends or intersects, which is often the blind spot of traditional single-point detection. By re-performing vertical depth measurement, the error of the deepest point of the crack 3D model can be further converged, and the maintenance depth design can be completed in one step. The difference in the number of extreme points and the second detection point essentially reflects the "complexity of acoustic wave-crack interaction". The system automatically triggers compensation based on this difference, eliminating the need for manual threshold setting. The entire detection-compensation process has self-correction capabilities and high long-term operational stability. The compensation results directly provide the coordinates of the deepest, farthest, and weakest points in the crack network. Grouting hole locations and milling boundaries can be accurately laid out in one go, avoiding secondary traffic closures and additional safety risks caused by "measuring again on-site." The compensation points and their depth results are fed back to the cloud as high-value labeled samples, continuously enriching the "crack-sound wave" correspondence database. This enables subsequent road sections to automatically iterate parameters such as extreme value judgment thresholds and direction weights, making the system increasingly "intelligent" with use.

[0058] The first and second test results are obtained to evaluate the road surface condition and obtain an evaluation value. The road surface condition is then classified into three levels: Level 1, Level 2, and Level 3. Levels 1, 2, and 3 represent increasingly better road surface conditions.

[0059] The calculation logic for the evaluation value includes:

[0060] Calculate the first sum of each first and second detection result, and set the first sum as the evaluation value.

[0061] The logic for classifying road surface condition levels includes:

[0062] The second and third values ​​are set as grade classification values. The evaluation value is compared with the grade classification value. When the evaluation value is less than or equal to the second value, the road surface condition level is set to the third level. When the evaluation value is greater than the second value and less than or equal to the third value, the road surface condition level is set to the second level. When the evaluation value is greater than the third value, the road surface condition level is set to the first level.

[0063] In practical implementation, the status determination can be completed instantly at the vehicle edge using only the single scalar value of the "first sum," without the need for complex algorithms or cloud computing power. This truly achieves a real-time closed loop of "detection-evaluation-grading." The second and third values ​​are externally adjustable variables that can be reset at any time based on maintenance budgets, traffic flow, or standard updates without altering the hardware and software architecture. This ensures long-term usability throughout the system's lifecycle. The three-level classification simplifies road conditions to three categories: "immediate treatment / planned maintenance / routine inspection." Maintenance departments no longer need to rely on experience-based judgments and can directly match maintenance processes, machinery, and funding plans according to the level. The first sum value integrates the results of the initial inspection (large-scale screening) and the retest (deepest crack densification), taking into account both the "area" and the "depth" of the cracks. This avoids overly strict or overly broad classifications caused by a single indicator. After comparing the evaluation value with the threshold, it outputs three integers: 1, 2, and 3. The front-end large screen, mobile app, and vehicle tablet can all immediately present the results in a clear red-yellow-green color scheme. On-site personnel can understand the results without any professional background. When a more refined 4-level or 5-level classification is needed, only a fourth or fifth value needs to be added. The original "sum value-comparison" framework does not need to be changed, and the system expansion cost is close to zero.

[0064] This invention first uses road surface images for rapid full-area screening, then uses ultrasound to verify each point in suspicious areas. This dual positioning reduces the false alarm rate. Depth is calculated using a sound velocity database and echo duration, and a second measurement is performed at the deepest crack, further reducing depth error and meeting the millimeter-level accuracy requirements for maintenance decisions. Image detection completes the initial screening of most areas, while ultrasound focuses only on key areas, reducing manpower and equipment costs. The first and second detection results are mapped to road surface condition levels, directly connecting to the highway maintenance management system to achieve seamless integration of defect classification, cost calculation, and construction order dispatch. The sound velocity database and detection parameters can be updated via the cloud, the image module can be upgraded with an AI model, and the ultrasound module is compatible with multiple modes such as shear waves and surface waves, adapting to different surface materials and climatic conditions.

[0065] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. The storage medium is implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0066] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention should all be covered within the scope of the claims of the present invention.

Claims

1. A highway pavement detection and evaluation system, characterized by, The method comprises a detection module, a compensation module and an evaluation module. The detection module sets a first detection point according to a road surface image, sets a first detection parameter, performs first detection at the first detection point, and obtains a first detection result. The compensation module sets a second detection point according to the first detection result, sets a second detection parameter, performs second detection at the second detection point, obtains a second detection result, obtains a first compensation point according to the second detection result, and repeats the first detection at the first compensation point to obtain a first compensation result. The evaluation module evaluates the road surface state according to the first detection result and the second detection result to obtain a road surface state grade. The first detection parameter comprises an ultrasonic wave amplitude and an ultrasonic wave frequency, and the setting logic of the first detection parameter comprises: The average temperature represents the daily average temperature of the region where the road is located, and the average humidity represents the daily average humidity of the region where the road is located. After the first detection parameter is set, a start detection signal is sent to a detection device at the first detection point, the detection device starts to emit ultrasonic waves perpendicular to the horizontal direction at the corresponding first detection point after receiving the start detection signal, and the first detection result is obtained after a first time period. The method for setting the second detection point according to the first detection result comprises: Any first detection point is selected, the echo time at the detection point is obtained, the sound wave speed database is called, the road filling material is input into the sound wave speed database, the sound wave speed corresponding to the road filling material is matched, the sound wave speed represents the walking speed of ultrasonic waves in the road filling material, the crack depth formula is called, the sound wave speed and the echo time are input into the crack depth formula, and the crack depth at the first detection point is calculated according to the crack depth formula. The echo time at each first detection point is traversed to obtain the crack depth at each first detection point. The crack depths are sorted in ascending order, when the crack depths are the same, only any crack depth is retained and other crack depths are deleted, a non-repeated increasing numerical sequence is obtained, the corresponding first detection points of the numerical sequence are obtained respectively, and the maximum value of the crack depth of the corresponding first detection point is set as the second detection point. The second detection parameter comprises an ultrasonic wave direction, an ultrasonic wave amplitude and an ultrasonic wave frequency, wherein the setting logic of the ultrasonic wave amplitude and the ultrasonic wave frequency is the same as that of the first detection parameter, and the ultrasonic wave direction comprises each direction within 360 degrees parallel to the horizontal direction. After the second parameter is set, the second detection is performed at the second detection point to obtain a second detection result, which is represented as an echo amplitude value; The logic for obtaining the first compensation point according to the second detection result comprises: Connecting the second detection points at any two positions to obtain a first connection line, setting the direction of the ultrasonic wave parallel to each connection line as a first direction, and obtaining any echo amplitude value at the first direction; Drawing the echo amplitude value as an acoustic wave curve, with the horizontal axis of the acoustic wave curve representing time and the vertical axis representing the echo amplitude value, counting the number of extreme points of the acoustic wave curve, denoted as a first number, counting the number of second detection points at the first direction, denoted as a second number, comparing the first number with the second number, jumping to the acoustic wave curve of the next first direction when the first number is equal to the second number, and calculating the first distance between each extreme point and the starting point according to the acoustic wave travel distance formula when the first number is greater than the second number, obtaining the second distance from each second detection point of the first connection line to the starting point, and setting the position of the extreme point with the second distance greater than the first distance as the first compensation point; Obtaining any echo value not at the first direction, drawing the echo value as an acoustic wave curve, and jumping to the next ultrasonic direction when the acoustic wave curve has no extreme value, and setting the position of each extreme point as the first compensation point when the acoustic wave curve has an extreme value.

2. A highway pavement detection and evaluation system as in claim 1, wherein: The logic for setting the first detection point according to the road surface image comprises: The road surface image is represented as a highway surface image, a standard crack image is called, a first feature quantity of the standard crack image is extracted, and a second feature quantity of the highway surface image is extracted, both of which are represented as shape feature quantities, a first similarity between the first feature quantity and the second feature quantity is calculated by a cosine similarity formula, a first value is set as a similarity threshold, each second feature quantity of the highway surface image is traversed, the first similarity is compared with the first value, the corresponding second feature quantity is deleted and the next second feature quantity is jumped when the first similarity is less than the first value, and the corresponding second feature quantity is retained when the first similarity is greater than or equal to the first value; The geometric center point of the region where each retained second feature quantity is located is set as the first detection point.

3. A highway pavement detection and evaluation system as in claim 1, wherein: The first detection is repeated at the first compensation point to obtain a first compensation result, which is represented as the crack depth at the first compensation point.

4. A highway pavement detection and evaluation system as in claim 3, wherein: The first detection result and the second detection result are obtained, the road surface state is evaluated to obtain an evaluation value, and the road surface state level is divided according to the evaluation value, the road surface state level comprises a first level, a second level and a third level, and the road surface state represented by the first level, the second level and the third level is better and better; The calculation logic of the evaluation value comprises: The first and value of each first detection result and second detection result is calculated, and the first and value is set as the evaluation value.

5. A highway pavement detection and evaluation system as in claim 4, wherein: The division logic of the road surface state level comprises: The second value and the third value are set as the grade division value, the evaluation value is compared with the grade division value, when the evaluation value is less than or equal to the second value, the road surface state grade is set as the third grade, when the evaluation value is greater than the second value and less than or equal to the third value, the road surface state grade is set as the second grade, and when the evaluation value is greater than the third value, the road surface state grade is set as the first grade.

Images