A method and system for monitoring road paving construction indicators

Abstract

The present invention is applicable to the field of paving monitoring technology, and provides a method and system for monitoring road paving construction indicators, including the following steps: receiving manually monitored indicator information and visually monitored indicator information, where the visually monitored indicator information includes a number of road paving images; detecting the integrity and compliance of the manually monitored indicator information to generate manually monitored indicator information; geometrically correcting the road paving images, identifying calibration length crossbars, and determining the paving width based on the calibration length crossbars, where each road paving image contains calibration length crossbars; analyzing the road paving images based on computer vision technology to determine the paving surface quality, where the paving surface quality includes flatness, crack conditions, and pothole conditions; and determining visually monitored indicator information based on the paving width, flatness, crack conditions, and pothole conditions. The present invention combines manual monitoring and visual monitoring, utilizes their respective advantages, and achieves comprehensive coverage and monitoring of the construction site.

Description

Technical Field

[0001] The present invention relates to the technical field of paving monitoring, and specifically relates to a method and system for monitoring road paving construction indicators. Background Art

[0002] Road paving construction is a key link to ensure road quality. Traditional road paving construction monitoring mainly relies on manual on-site monitoring. The traditional method is not only time-consuming and laborious, but also greatly affected by human factors, making it difficult to ensure the accuracy and consistency of data. In recent years, with the progress of technology, especially the development of computer vision technology, new means have been provided for the monitoring of road paving construction. However, relying solely on visual monitoring or manual monitoring has its own limitations. Although manual monitoring can directly obtain the actual data of the construction site, in the face of large-scale, complex and changeable construction sites, it is often difficult for manual monitoring to achieve full coverage and real-time monitoring. At the same time, the accuracy and integrity of the monitoring data are also easily affected by various factors such as the skill level and work attitude of the monitoring personnel.

[0003] For monitoring based on computer vision technology, although it can achieve real-time monitoring and automated data processing, in practical applications, due to the complexity and diversity of road paving construction, the accuracy and stability of the visual monitoring results will be affected. Therefore, it is necessary to provide a method and system for monitoring road paving construction indicators to solve the above problems. Summary of the Invention

[0004] Aiming at the deficiencies of the existing technology, the purpose of the present invention is to provide a method and system for monitoring road paving construction indicators to solve the problems existing in the above background art.

[0005] The present invention is implemented as follows. A method for monitoring road paving construction indicators, the method includes the following steps:

[0006] Receive manual monitoring indicator information and visual monitoring indicator information, where the manual monitoring indicator information includes several construction indicators and monitoring data, and the visual monitoring indicator information includes several road paving images;

[0007] Detect the integrity and compliance of the manual monitoring indicator information to generate manual indicator monitoring information;

[0008] Perform geometric correction on the road paving images in the visual monitoring indicator information, identify the calibration length crossbar, and determine the paving width based on the calibration length crossbar. Each road paving image contains the calibration length crossbar;

[0009] Analyze the road paving images based on computer vision technology to determine the paving surface quality, and the paving surface quality includes flatness, crack conditions and pothole conditions;

[0010] Determine the visual index monitoring information according to the paving width, flatness, crack condition and pothole condition.

[0011] Another object of the present invention is to provide a road paving construction index monitoring system, the system comprising:

[0012] A monitoring index information module, configured to receive manual monitoring index information and visual monitoring index information, the manual monitoring index information includes a plurality of construction indexes and monitoring data, and the visual monitoring index information includes a plurality of road paving images;

[0013] A manual index monitoring module, configured to detect the integrity and compliance of the manual monitoring index information, and generate manual index monitoring information;

[0014] A paving width determination module, configured to perform geometric correction on the road paving images in the visual monitoring index information, identify the calibration length crossbar, and determine the paving width based on the calibration length crossbar. Each road paving image includes a calibration length crossbar;

[0015] A surface quality determination module, configured to analyze the road paving images based on computer vision technology to determine the paving surface quality, and the paving surface quality includes flatness, crack condition and pothole condition;

[0016] A visual index monitoring module, configured to determine the visual index monitoring information according to the paving width, flatness, crack condition and pothole condition.

[0017] Compared with the prior art, the beneficial effects of the present invention are:

[0018] The present invention combines manual monitoring and visual monitoring, utilizes their respective advantages, realizes the full coverage and real-time monitoring of the construction site, detects the integrity and compliance of the manual monitoring index information, and generates manual index monitoring information. In this way, the integrity and reliability of the data can be ensured; by performing geometric correction on the road paving images in the visual monitoring index information, automatically identifying the calibration length crossbar, automatically determining the paving width based on the calibration length crossbar, and analyzing the road paving images according to computer vision technology to determine the paving surface quality, the advantages of computer vision technology are fully utilized, and the accuracy and stability of the visual monitoring results are improved. Description of the Drawings

[0019] Figure 1 It is a flowchart of a road paving construction index monitoring method.

[0020] Figure 2 It is a flowchart of generating manual index monitoring information in a road paving construction index monitoring method.

[0021] Figure 3 It is a flowchart for identifying and calibrating the length crossbar in a method for monitoring road paving construction indicators.

[0022] Figure 4 It is a flowchart for determining the paving width in a method for monitoring road paving construction indicators.

[0023] Figure 5 It is a flowchart for determining the paving surface quality in a method for monitoring road paving construction indicators.

[0024] Figure 6 It is a schematic structural diagram of a road paving construction indicator monitoring system. Specific implementation manners

[0025] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0026] The following describes the specific implementation of the present invention in detail with reference to specific embodiments.

[0027] As Figure 1 shown, an embodiment of the present invention provides a method for monitoring road paving construction indicators, and the method includes the following steps:

[0028] S100, receiving artificial monitoring indicator information and visual monitoring indicator information, where the artificial monitoring indicator information includes several construction indicators and monitoring data, and the visual monitoring indicator information includes several road paving images;

[0029] S200, detecting the integrity and compliance of the artificial monitoring indicator information to generate artificial indicator monitoring information;

[0030] S300, geometrically correcting the road paving images in the visual monitoring indicator information, identifying and calibrating the length crossbar, and determining the paving width based on the calibrated length crossbar. Each road paving image includes a calibrated length crossbar;

[0031] S400, analyzing the road paving images based on computer vision technology to determine the paving surface quality, where the paving surface quality includes flatness, crack conditions and pothole conditions;

[0032] S500, determining visual indicator monitoring information according to the paving width, flatness, crack conditions and pothole conditions.

[0033] It should be noted that road paving construction is a key link to ensure road quality. Relying solely on visual monitoring or manual monitoring has its own limitations. Although manual monitoring can directly obtain the actual data of the construction site, in the face of large-scale, complex and changeable construction sites, it is often difficult to achieve full coverage and real-time monitoring. At the same time, the accuracy and integrity of the monitoring data are also easily affected by various factors such as the skill level and work attitude of the monitoring personnel. For the monitoring based on computer vision technology, although it can achieve real-time monitoring and automated data processing, in actual applications, due to the complexity and diversity of road paving construction, the accuracy and stability of the visual monitoring results will be affected. The embodiments of the present invention aim to solve the above problems.

[0034] In the embodiments of the present invention, manual monitoring and visual monitoring are combined to utilize their respective advantages to achieve full coverage and real-time monitoring of the construction site. Specifically, first, the manual monitoring index information and the visual monitoring index information need to be obtained. The manual monitoring index information includes several construction indexes and monitoring data, which can provide direct and accurate information of the construction site. The visual monitoring index information includes several road paving images, and the indexes with high accuracy are automatically monitored by using the road paving images. Then, the integrity and compliance of the manual monitoring index information are detected to generate manual index monitoring information. In this way, the integrity and reliability of the data can be ensured, which helps to reduce the influence of human factors on the monitoring results and improve the accuracy and consistency of the data. Then, the road paving images in the visual monitoring index information are geometrically corrected, and the calibration length crossbar is automatically identified, and the paving width is automatically determined based on the calibration length crossbar.

[0035] It should be noted that each road paving image contains a calibration length crossbar, and the calibration length crossbar has a certain length value. Finally, the road paving images are analyzed according to computer vision technology to determine the paving surface quality. The paving surface quality includes flatness, crack condition and pothole condition. The visual index monitoring information is determined according to the paving width, flatness, crack condition and pothole condition. In this way, the advantages of computer vision technology are fully utilized, and the accuracy and stability of the visual monitoring results are improved.

[0036] As Figure 2 shown, as a preferred embodiment of the present invention, the step of detecting the integrity and compliance of the manual monitoring index information to generate the manual index monitoring information specifically includes:

[0037] S201, determining whether the construction indexes in the manual monitoring index information are complete to obtain the manual index integrity;

[0038] S202, determining whether the monitoring data corresponding to each construction index complies with the specifications to obtain the manual index compliance;

[0039] S203. Generate manual index monitoring information based on the integrity and compliance of manual indexes.

[0040] In the embodiment of the present invention, after receiving the manual monitoring index information, it will automatically determine whether the construction indexes in the manual monitoring index information are complete, that is, whether the construction indexes are complete, so as to obtain the integrity of the manual indexes. Then, it will automatically determine whether the monitoring data corresponding to each construction index meets the specifications, that is, whether the construction indexes are within the corresponding qualified ranges, so as to obtain the compliance of the manual indexes, and then generate the manual index monitoring information.

[0041] As Figure 3 shown, as a preferred embodiment of the present invention, the steps of geometrically correcting the road paving image in the visual monitoring index information and identifying the calibration length crossbar specifically include:

[0042] S301. Compare the road paving image with the reference template image to determine a number of control points;

[0043] S302. Calculate the correction parameters based on the control points and the affine transformation model, and perform geometric correction;

[0044] S303. Retrieve the color features and contour features of the calibration length crossbar, and identify the calibration length crossbar based on the color features and contour features.

[0045] In the embodiment of the present invention, in order to better perform subsequent processing, it is necessary to geometrically correct the road paving image. Specifically, the road paving image will be compared with the reference template image, which is set in advance. A number of control points are determined, and the control points will be evenly distributed throughout the image. The more the number, the better the correction effect usually is. The control points include the intersections of roads, the corners of buildings, etc. Then, the correction parameters are calculated based on the control points and the affine transformation model, and geometric correction is performed. Check whether the corrected image meets the requirements, including whether the distortion is eliminated and whether the original information of the image is maintained, etc. Finally, the color features and contour features of the calibration length crossbar are retrieved, and the calibration length crossbar in the image is automatically identified based on the color features and contour features.

[0046] As a preferred embodiment of the present invention, the step of comparing the road paving image with the reference template image to determine a number of control points specifically includes the following sub-steps:

[0047] Convert the road paving image and the reference template image into grayscale images to obtain the road grayscale image and the template grayscale image respectively;

[0048] The SIFT algorithm is used to perform feature detection on the road grayscale image and the template grayscale image, obtaining the key point feature description of the road paving image and the key point feature description of the reference template image;

[0049] Based on K-nearest neighbor matching, for each key point feature description in the road paving image, at least two key point feature descriptions of the reference template image are matched, obtaining a number of K-nearest neighbor matching pairs;

[0050] The Lowe's ratio test is used to filter the K-nearest neighbor matching pairs, obtaining a number of filtered matching pairs;

[0051] The confidence score of a number of filtered matching pairs is calculated to obtain the first calculation result, and the first calculation result is normalized to a probability distribution;

[0052] Based on the RANSAC algorithm, weighted sampling and iterative calculation are performed on a number of filtered matching pairs using the probability distribution. After the iteration is completed, the optimal homography matrix is obtained;

[0053] Using the optimal homography matrix, the four corner points of the road paving image are perspectively transformed to obtain the transformed corner point coordinates, and the transformed corner point coordinates are mapped onto the reference template image to obtain the corresponding four control points on the reference template image.

[0054] In the embodiments of the present invention, technologies such as SIFT and RANSAC are combined, which can effectively cope with noise, occlusion, and scale changes in the image, ensuring the reliability of feature matching. And on the basis of the traditional RANSAC, a weighted mechanism of matching confidence is added, so that high-confidence matching pairs have a higher selection probability in model fitting. By using confidence information, the accuracy and robustness of model estimation are improved.

[0055] As a preferred embodiment of the present invention, the step of calculating correction parameters based on the control points and the affine transformation model and performing geometric correction specifically includes the following sub-steps:

[0056] The control points of the road paving image and the reference template image are respectively described by two sets of linear equations to represent the transformation relationship of horizontal and vertical coordinates, so as to obtain a coordinate matrix;

[0057] The coordinate matrix is decomposed by singular value decomposition to obtain affine transformation parameters, and an affine transformation matrix is constructed according to the affine transformation parameters;

[0058] Each pixel coordinate in the road paving image is transformed using the affine transformation matrix to obtain new coordinates;

[0059] The pixel value of the new coordinates is calculated using the bilinear interpolation method, and then post-processing and boundary adjustment are performed in sequence to obtain the transformed image.

[0060] In the embodiments of the present invention, since the control points directly correspond between the known original and target positions, using the control points can ensure the accuracy of image transformation. By calculating the transformation relationships of these points, an accurate transformation model can be obtained; and using singular value decomposition to solve the linear equations can ensure the accurate calculation of the affine transformation parameters, perform well in dealing with rank-deficient matrices and noise interference, and can provide a stable least-squares solution to ensure the accuracy of geometric correction.

[0061] As Figure 4 shown, as a preferred embodiment of the present invention, the steps of determining the paving width based on the calibrated length crossbar specifically include:

[0062] S304. According to the identified calibrated length crossbar, calculate the pixel length Lp of the calibrated length crossbar in the road paving image.

[0063] S305. Calculate the crossbar length ratio Sc, where Sc = Lr / Lp, and Lr represents the actual length of the calibrated length crossbar.

[0064] S306. Identify the paving edges parallel to the calibrated length crossbar, determine the pixel width Wp between the edges, and calculate the actual paving width Wr, where Wr = Wp × Sc.

[0065] In the embodiments of the present invention, according to the identified calibrated length crossbar, the pixel length Lp of the calibrated length crossbar in the road paving image is determined. The actual length Lr of the calibrated length crossbar is known, so that the crossbar length ratio Sc can be calculated, where Sc = Lr / Lp. Then, identify the paving edges parallel to the calibrated length crossbar. Edge detection algorithms or line fitting algorithms can be used to identify the edges, determine the pixel width Wp between the edges, and finally calculate the actual paving width Wr, where Wr = Wp × Sc.

[0066] As Figure 5 shown, as a preferred embodiment of the present invention, the steps of analyzing the road paving image based on computer vision technology to determine the paving surface quality specifically include:

[0067] S401. Traverse each pixel point in the road paving image, count the gray values, calculate the average value and standard deviation of the gray values to obtain the overall flatness.

[0068] S402. Detect the crack area in the road paving image according to the threshold segmentation algorithm, and determine the crack situation according to the width, length and gray value characteristics of the cracks.

[0069] S403. Detect the pothole area in the road paving image according to the edge detection algorithm, and determine the pothole situation according to the shape, size and depth characteristics of the potholes.

[0070] In the embodiments of the present invention, in order to determine the flatness, crack conditions, and pothole conditions, each pixel point in the road paving image is automatically traversed, the gray values are statistically counted, the average value and standard deviation of the gray values are calculated to obtain the overall flatness. Further, the image can be divided into multiple small regions, the gray value distributions of the small regions are compared, and the regions with large differences are found, which may represent uneven regions. Then, the crack regions in the road paving image are detected according to the threshold segmentation algorithm, and the crack conditions are determined according to the width, length, and gray value characteristics of the cracks;

[0071] Further, machine learning algorithms (such as K-means clustering, support vector machines, etc.) can be used to further classify and confirm the cracks, and the cracks are classified according to the severity, such as minor cracks, moderate cracks, and severe cracks. Then, the pothole regions in the road paving image are detected according to the edge detection algorithm, and the pothole conditions are determined according to the shape, size, and depth characteristics of the potholes. Further, the detected potholes are quantitatively evaluated, such as calculating the area, depth, and volume of the potholes, and the potholes are classified according to the evaluation results, such as small potholes, medium potholes, and large potholes.

[0072] As a preferred embodiment of the present invention, the steps of traversing each pixel point in the road paving image, statistically counting the gray values, calculating the average value and standard deviation of the gray values, and obtaining the overall flatness specifically include the following sub-steps:

[0073] Apply Fourier transform to the road gray image to convert it to the frequency domain and perform spectral centering to obtain the spectrum;

[0074] Filter the spectrum using a filter, and then convert the filtered spectrum back to the spatial domain through inverse Fourier transform to obtain the filtered image;

[0075] Perform multi-scale smoothing on the filtered image in a way that each scale corresponds to a different Gaussian blur degree to obtain a multi-scale smoothed image;

[0076] Calculate the difference between the filtered image and the multi-scale smoothed image to obtain the differences at different scales;

[0077] Calculate the standard deviation of the differences at different scales to obtain the local standard deviations at different scales;

[0078] Accumulate the local standard deviations at different scales and perform normalization processing to obtain the normalized local standard deviations;

[0079] Calculate the overall gray mean value of the filtered image to obtain the global mean value;

[0080] The degree of change in each part of the image is measured by the normalized local standard deviation. The global mean is used as a reference for the overall brightness of the image. The global mean and the normalized local standard deviation are weighted and combined in a linear combination to obtain the overall flatness.

[0081] In the embodiments of the present invention, in order to ensure the accuracy of the overall flatness calculation, the present invention captures multi-scale features of the image by performing smoothing and local standard deviation calculations at different scales. This method can identify surface changes at different scales, making it applicable to various complex scenarios. And by using Fourier transform and frequency domain filtering, it can effectively separate and retain the frequency information of interest. This is very helpful for removing noise and highlighting important details, thereby improving the accuracy of flatness evaluation. Finally, the local changes and the overall features are combined together, balancing local and global information, and can more comprehensively reflect the complex features of the image to more accurately and comprehensively evaluate the overall flatness of the road paving image.

[0082] As a preferred embodiment of the present invention, the step of detecting the crack area in the road paving image according to the threshold segmentation algorithm and determining the crack situation according to the width, length and gray value characteristics of the crack specifically includes the following sub-steps:

[0083] The road gray image is enhanced in contrast by local histogram equalization to obtain an image with enhanced contrast;

[0084] The image with enhanced contrast is adaptively threshold-segmented by the adaptive threshold segmentation method to generate a binary image;

[0085] The binary image is processed by closing operation to obtain a processed image;

[0086] Contour detection is performed on the processed image to obtain a number of contours and corresponding bounding boxes; the length of each contour and the width of the bounding box are calculated, and the average gray value of the contour is calculated. According to the length and width of each contour and the average gray value, the crack situation is determined.

[0087] In the embodiments of the present invention, based on the local histogram equalization technology, the contrast of the image is improved, making the cracks more obvious under different lighting conditions. It helps to improve the accuracy of detection, especially in uneven lighting or dim environments. And the adaptive threshold segmentation determines the threshold according to the local information of each pixel neighborhood, and can more effectively process complex backgrounds and uneven lighting, thereby enhancing the sensitivity and stability of crack detection. By morphological closing operation, noise is removed, the boundaries of the cracks are made smoother, effectively reducing false detections and improving the reliability of the results.

[0088] Such as Figure 6As shown in the figure, an embodiment of the present invention further provides a road paving construction index monitoring system, and the system includes:

[0089] A monitoring index information module 100, configured to receive manual monitoring index information and visual monitoring index information, where the manual monitoring index information includes several construction indexes and monitoring data, and the visual monitoring index information includes several road paving images;

[0090] A manual index monitoring module 200, configured to detect the integrity and compliance of the manual monitoring index information, and generate manual index monitoring information;

[0091] A paving width determination module 300, configured to perform geometric correction on the road paving images in the visual monitoring index information, identify the calibration length crossbar, and determine the paving width based on the calibration length crossbar. Each road paving image includes a calibration length crossbar;

[0092] A surface quality determination module 400, configured to analyze the road paving images based on computer vision technology to determine the paving surface quality, and the paving surface quality includes flatness, crack condition, and pothole condition;

[0093] A visual index monitoring module 500, configured to determine visual index monitoring information according to the paving width, flatness, crack condition, and pothole condition.

[0094] As a preferred embodiment of the present invention, the manual index monitoring module 200 includes:

[0095] An index integrity unit, configured to determine whether the construction indexes in the manual monitoring index information are complete, and obtain the manual index integrity;

[0096] An index compliance unit, configured to determine whether the monitoring data corresponding to each construction index complies with the specification, and obtain the manual index compliance;

[0097] A manual index monitoring unit, configured to generate manual index monitoring information according to the manual index integrity and the manual index compliance.

[0098] As a preferred embodiment of the present invention, the paving width determination module 300 includes:

[0099] A control point determination unit, configured to compare the road paving image with a reference template image to determine several control points;

[0100] A geometric correction unit, configured to calculate correction parameters based on the control points and an affine transformation model, and perform geometric correction;

[0101] A calibration crossbar recognition unit, configured to retrieve the color feature and contour feature of the calibration length crossbar, and recognize the calibration length crossbar based on the color feature and the contour feature.

[0102] As a preferred embodiment of the present invention, the paving width determination module 300 further includes:

[0103] A pixel length calculation unit, configured to calculate the pixel length Lp of the calibration length crossbar in the road paving image according to the recognized calibration length crossbar;

[0104] A crossbar length ratio unit, configured to calculate the crossbar length ratio Sc, where Sc = Lr / Lp, and Lr represents the actual length of the calibration length crossbar;

[0105] An actual paving width unit, configured to identify the paving edges parallel to the calibration length crossbar, determine the pixel width Wp between the edges, and calculate the actual paving width Wr, where Wr = Wp × Sc.

[0106] As a preferred embodiment of the present invention, the surface quality determination module 400 includes:

[0107] An overall flatness unit, configured to traverse each pixel point in the road paving image, count the gray values, calculate the average value and standard deviation of the gray values, and obtain the overall flatness;

[0108] A crack condition determination unit, configured to detect the crack area in the road paving image according to the threshold segmentation algorithm, and determine the crack condition according to the width, length, and gray value characteristics of the crack;

[0109] A pothole condition determination unit, configured to detect the pothole area in the road paving image according to the edge detection algorithm, and determine the pothole condition according to the shape, size, and depth characteristics of the pothole.

[0110] The above only describes the preferred embodiments of the present invention in detail, and does not limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

[0111] It should be understood that although the steps in the flowcharts of the embodiments of the present invention are shown in sequence according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit, and these steps can be executed in other orders. Moreover, at least a part of the steps in each embodiment may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0112] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0113] After considering the specification and the disclosure of the embodiments, those skilled in the art will readily conceive of other embodiments of the present disclosure. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow the general principles of the present disclosure and include known common knowledge or conventional technical means in the technical field not disclosed in the present disclosure. The specification and the embodiments are only regarded as exemplary, and the true scope and spirit of the present disclosure are pointed out by the claims.

Claims

1. A road paving construction index monitoring method, characterized in that: The method comprises the following steps: Receiving manual monitoring index information and visual monitoring index information, wherein the manual monitoring index information includes a plurality of construction indexes and monitoring data, and the visual monitoring index information includes a plurality of road paving images; Detect the completeness and compliance of manual monitoring indicator information and generate manual indicator monitoring information; Performing geometric correction on the road paving image in the visual monitoring index information, identifying the calibration length crossbar, and determining the paving width based on the calibration length crossbar, wherein each road paving image includes the calibration length crossbar; Analyze road paving images based on computer vision technology to determine the paving surface quality, including flatness, cracks and potholes; Determine visual indicator monitoring information based on paving width, flatness, crack conditions and pothole conditions; The step of detecting the completeness and conformity of the manual monitoring indicator information and generating the manual indicator monitoring information specifically includes: Determine whether the construction indicators in the manual monitoring indicator information are complete, and obtain the completeness of the manual indicators; Determine whether the monitoring data corresponding to each construction indicator meets the specifications and obtain the compliance degree of the artificial indicator; Generate manual indicator monitoring information based on manual indicator completeness and manual indicator compliance; The step of geometrically correcting the road paving image in the visual monitoring index information and identifying the calibration length crossbar specifically includes: Compare the road paving image with the reference template image to determine a number of control points; Calculate correction parameters based on control points and affine transformation model to perform geometric correction; Retrieving the color features and contour features of the calibration length crossbar, and identifying the calibration length crossbar based on the color features and contour features; The step of comparing the road paving image with the reference template image to determine a number of control points specifically includes the following sub-steps: The road paving image and the reference template image are converted into grayscale images to obtain a road grayscale image and a template grayscale image respectively; The SIFT algorithm is used to perform feature detection on the road grayscale image and the template grayscale image to obtain the key point feature description of the road paving image and the key point feature description of the reference template image; Based on K-nearest neighbor matching, each key point feature description in the road paving image is matched with the key point feature description of at least two reference template images to obtain a number of K-nearest neighbor matching pairs; Lowe's ratio test is used to filter the K nearest neighbor matching pairs to obtain several filtered matching pairs; Calculating confidence scores for the plurality of filtered matching pairs to obtain a first calculation result, and normalizing the first calculation result into a probability distribution; Based on the RANSAC algorithm, the probability distribution is used to perform weighted sampling and iterative calculation on several filtered matching pairs. After the iteration is completed, the optimal homography matrix is ​​obtained; Using the optimal homography matrix, the four corner points of the road paving image are perspective transformed to obtain the transformed corner point coordinates, and the transformed corner point coordinates are mapped to the reference template image to obtain the corresponding four control points on the reference template image; The step of calculating the correction parameters based on the control points and the affine transformation model and performing geometric correction specifically includes the following sub-steps: The control points of the road paving image and the reference template image are respectively described by two sets of linear equations for the transformation relationship of the horizontal and vertical coordinates to obtain a coordinate matrix; Decomposing the coordinate matrix by singular value to obtain affine transformation parameters, and constructing an affine transformation matrix according to the affine transformation parameters; Using an affine transformation matrix to transform each pixel coordinate in the road paving image to obtain a new coordinate; The pixel values ​​of the new coordinates are calculated using the bilinear interpolation method, and then post-processing and boundary adjustment are performed in sequence to obtain the transformed image; The step of determining the paving width based on the calibrated length crossbar specifically includes: According to the identified calibrated length crossbar, the pixel length Lp of the calibrated length crossbar in the road paving image is calculated; Calculate the crossbar length ratio Sc, crossbar length ratio Sc=Lr / Lp, where Lr represents the actual length of the crossbar with the calibrated length; Identify the paving edge parallel to the calibrated length crossbar, determine the pixel width Wp between the edges, and calculate the actual paving width Wr, Wr=Wp×Sc; The step of analyzing the road paving image based on computer vision technology to determine the paving surface quality specifically includes: Traverse each pixel in the road paving image, count the grayscale value, calculate the average and standard deviation of the grayscale value, and obtain the overall flatness; The crack area in the road paving image is detected according to the threshold segmentation algorithm, and the crack condition is determined according to the width, length and gray value characteristics of the crack; Detect pothole areas in road paving images based on edge detection algorithms, and determine pothole conditions based on the shape, size, and depth features of the potholes; The step of traversing each pixel point in the road paving image, counting the grayscale value, calculating the average value and standard deviation of the grayscale value, and obtaining the overall flatness specifically includes the following sub-steps: The road grayscale image is converted into the frequency domain by Fourier transform and the spectrum is centered to obtain the spectrum; The spectrum is filtered using a filter, and the filtered spectrum is then converted back to the spatial domain by inverse Fourier transform to obtain a filtered image; In a way that each scale corresponds to a different Gaussian blur degree, a multi-scale smoothing operation is performed on the filtered image to obtain a multi-scale smoothed image; Calculate the difference between the filtered image and the multi-scale smoothed image to obtain the differences at different scales; The standard deviation of the differences at different scales is calculated to obtain the local standard deviation at different scales; The local standard deviations of different scales are accumulated and normalized to obtain the normalized local standard deviation; Calculate the overall grayscale mean of the filtered image to obtain the global mean; The normalized local standard deviation is used to measure the degree of change of each part of the image, and the global mean is used as the reference part of the overall brightness of the image. The global mean and the normalized local standard deviation are weighted and combined by linear combination to obtain the overall flatness; The step of detecting the crack area in the road paving image according to the threshold segmentation algorithm and determining the crack condition according to the width, length and gray value characteristics of the crack specifically includes the following sub-steps: The road grayscale image is subjected to contrast enhancement by using local histogram equalization to obtain a contrast enhanced image; The contrast-enhanced image is segmented using an adaptive threshold segmentation method to generate a binary image; The binary image is processed by using a closing operation to obtain a processed image; Perform contour detection on the processed image to obtain several contours and corresponding bounding boxes; The length of each contour and the width of the bounding box are calculated, and the average gray value of the contour is calculated. The crack condition is determined based on the length of each contour, the width of the bounding box and the average gray value.

2. A road paving construction index monitoring system, the system is applied to the road paving construction index monitoring method according to claim 1, characterized in that: The system comprises: A monitoring index information module, used to receive manual monitoring index information and visual monitoring index information, wherein the manual monitoring index information includes a number of construction indexes and monitoring data, and the visual monitoring index information includes a number of road paving images; The manual indicator monitoring module is used to detect the completeness and compliance of manual monitoring indicator information and generate manual indicator monitoring information; A paving width determination module is used to perform geometric correction on the road paving image in the visual monitoring index information, identify the calibration length crossbar, and determine the paving width based on the calibration length crossbar, and each road paving image contains the calibration length crossbar; A surface quality determination module is used to analyze the road paving images based on computer vision technology to determine the paving surface quality, which includes flatness, cracks and potholes; The visual indicator monitoring module is used to determine the visual indicator monitoring information according to the paving width, flatness, crack conditions and pothole conditions.

Images