Low-grade highway construction quality intelligent detection method and system

Abstract

The invention discloses a low-grade highway construction quality intelligent detection method and system, and relates to the technical field of highway engineering detection, and the method comprises the steps: collecting an RGB image, converting the RGB image into a gray-scale map, and calculating a flatness feature value through a flatness feature fusion method; calculating a gap characteristic coefficient based on the flatness characteristic value, and calculating an edge disorder index by using an accumulation method; and evaluating a construction quality score based on the gap feature coefficient and the edge disorder index, and constructing a visual interface to display the construction quality score. The method comprises the following steps: collecting an RGB image, converting the RGB image into a grey-scale map, and calculating a flatness characteristic value by using a flatness characteristic fusion method; calculating a gap characteristic coefficient, and calculating an edge disorder index by using an accumulation method; the construction quality score is evaluated, the limitation of a single characteristic index is avoided, the efficiency and precision of pavement quality detection are improved, and the robustness and adaptability of detection are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of highway engineering detection, and in particular to an intelligent detection method and system for low-grade highway construction quality. Background Art

[0002] As an important infrastructure connecting urban and rural areas and promoting local economic development, the construction quality of low-grade roads is directly related to the safety and service life of the roads. Since low-grade roads are mostly located in remote areas, they are limited by construction conditions and budgets. There is often a lack of effective quality monitoring methods in the construction process, resulting in problems such as insufficient road flatness, frequent cracks, and improper edge treatment. This not only increases the cost of subsequent maintenance, but also poses a potential threat to traffic safety. With the development of intelligent technology, construction quality detection methods based on sensors, computer vision and big data analysis have gradually emerged. These methods use high-precision cameras to collect image data of the construction area, and combine deep learning, statistical analysis and image processing technology to achieve real-time monitoring of construction quality to a certain extent. The current detection methods are mostly designed for high-grade roads, lacking adaptability to the construction characteristics of low-grade roads, especially in terms of efficiency and reliability in complex construction environments.

[0003] The existing technology mainly has the following deficiencies. For the detection of road cracks and irregular edges, existing methods usually rely on high-precision laser scanning or manual inspection, which not only increases costs but also cannot guarantee real-time performance. Many methods only rely on a single physical parameter (such as road surface flatness) for quality assessment, and lack comprehensive analysis of multi-dimensional features (such as gradient, entropy field, fractal dimension, etc.), resulting in insufficient comprehensiveness and scientificity of the evaluation results. Summary of the invention

[0004] In view of the problems existing in the above-mentioned existing low-grade highway construction quality intelligent detection method and system, the present invention is proposed.

[0005] Therefore, the problem to be solved by the present invention is that for the detection of road cracks and irregular edges, existing methods usually rely on high-precision laser scanning or manual inspection, which not only increases costs but also cannot guarantee real-time performance. Many methods only rely on a single physical parameter (such as road surface flatness) for quality assessment, and lack comprehensive analysis of multi-dimensional features (such as gradient, entropy field, fractal dimension, etc.), resulting in insufficient comprehensiveness and scientificity of the evaluation results.

[0006] To solve the above technical problems, the present invention provides the following technical solutions: an intelligent detection method for the construction quality of a low-grade highway, which comprises collecting RGB images and converting them into grayscale images, and calculating the smoothness eigenvalues ​​using a smoothness feature fusion method; calculating the gap characteristic coefficient based on the smoothness eigenvalues, and calculating the edge disorder index using an accumulation method; evaluating the construction quality score based on the gap characteristic coefficient and the edge disorder index, and constructing a visual interface to display the construction quality score.

[0007] As a preferred solution of the intelligent detection method for construction quality of low-grade highways of the present invention, wherein: the collection of RGB images and conversion into grayscale images refers to using a high-resolution CMOS camera to collect RGB images of the pavement of the low-grade highway while the construction vehicle is moving; The preprocessing of the collected RGB images includes: using an electronic image stabilization algorithm to correct geometric distortion of the RGB images, using an adaptive bilateral filtering algorithm to denoise the RGB images, and normalizing the denoised RGB images; The preprocessed RGB image is converted into a grayscale image using the weighted average method.

[0008] As a preferred solution of the intelligent detection method for construction quality of low-grade highways described in the present invention, wherein: the use of the flatness feature fusion method to calculate the flatness feature value refers to using a fixed sliding window method to set the window N, segmenting the grayscale image according to the set window, and calculating the local mean of the window image ; Use statistical analysis to calculate the probability of the grayscale value of the window image , the information entropy of the window is calculated based on the probability of the gray value of the window image ; Use the Sobel operator to calculate the horizontal gradient of the window image pixels respectively and vertical gradient , use the Euclidean norm method to calculate the gradient amplitude of the window pixel ; Combine the gradient field with the entropy field to calculate the unevenness eigenvalue , the formula is: , in is the gradient amplitude of the offset pixel in the window, and u and v are the offsets; In the window image, use the Flood Fill algorithm to identify and calculate the grayscale mean of the i-th connected domain ; Use resolution adjustment to set the mesh edge length ; Calculating the fractal dimension of connected components in windowed images using box counting ; Combine the grayscale mean of the connected domain with the fractal dimension to calculate the convex eigenvalue ; Based on the unevenness eigenvalue and convex eigenvalues , use the flatness feature fusion method to calculate the flatness feature value of the window image , the formula is: , in is the flatness eigenvalue of the window image.

[0009] As a preferred solution of the intelligent detection method for construction quality of low-grade highways described in the present invention, the following is a preferred solution: The calculation of the gap characteristic coefficient based on the flatness characteristic value refers to the calculation of the gap characteristic coefficient based on the horizontal gradient of the window image. and vertical gradient , use the inverse tangent method to calculate the gradient direction ; Take the center pixel in the window image as the reference point and calculate the gradient direction difference of the pixels in the window ; The average angle difference of the direction of the pixel points in the window image is calculated using the direction difference accumulation method; Calculate the consistency index of gradient direction based on the average angle difference ; Perform fast Fourier transform on the window image to extract the energy ratio of high-frequency components in the frequency domain ; Based on the flatness feature value , gradient amplitude , Gradient direction consistency index And the energy ratio of high-frequency components in the frequency domain , calculate the gap characteristic coefficient of the window image , the formula is: , in is the gap characteristic coefficient of the window image.

[0010] As a preferred solution of the intelligent detection method for construction quality of low-grade highways of the present invention, the calculation of edge clutter index by using the accumulation method includes: Use the Canny operator to detect and identify the edge of the window image, sort the edge pixels according to their connectivity, and form an ordered sequence of edge points; For edge points in the edge sequence, select the front and back edge points to form a three-point set; Use the Euclidean distance method to calculate the distance between three edge points; Calculate local curvature using the three-point method ; Based on local curvature , calculate the curvature change ; Use Euclidean distance to calculate the spatial distance between adjacent edge points , defines the dynamic difference of the distribution between points ; Calculation of local Lyapunov exponents based on chaos metric method ; The curvature change rate and Lyapunov index of all edge points are accumulated to calculate the edge disorder index , the formula is: , Where Q is the total number of edge points and z is the edge point index.

[0011] As a preferred solution of the intelligent detection method for construction quality of low-grade highways described in the present invention, the gap characteristic coefficient and edge disorder index are used to evaluate the construction quality score by combining nonlinear activation functions to calculate the gap characteristic coefficient. and edge clutter index Comprehensive characteristics of ; Use Otsu algorithm to set abnormal threshold and integrate comprehensive features Compare with the optimal threshold, and take the comprehensive features greater than or equal to the abnormal threshold The corresponding window image is judged as an abnormal area, and the comprehensive features smaller than the abnormal threshold are The corresponding window image is judged as a normal area; Use sliding window feature superposition method to merge window images into grayscale images; Total area of ​​statistical abnormal areas , Construction quality scoring method based on defect area ratio to evaluate construction quality scoring .

[0012] As a preferred solution of the intelligent detection method for construction quality of low-grade highways of the present invention, wherein: the construction of a visual interface to display the construction quality score refers to using the front-end technology Chart.js to construct a visual interface to display the state of new energy power generation; Users who have passed real-name verification are allowed to view the information.

[0013] Another object of the present invention is to provide an intelligent detection system for construction quality of low-grade highways, which includes a collection and fusion module for collecting RGB images and converting them into grayscale images, and calculating flatness feature values ​​using a flatness feature fusion method; A calculation accumulation module is used to calculate the gap characteristic coefficient based on the flatness characteristic value and calculate the edge disorder index using the accumulation method; The evaluation visualization module is used to evaluate the construction quality score based on the gap characteristic coefficient and edge messiness index, and to build a visualization interface to display the construction quality score.

[0014] A computer device comprises: a memory and a processor; the memory stores a computer program, and the processor implements the steps of the above-mentioned low-grade highway construction quality intelligent detection method when executing the computer program.

[0015] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned low-grade highway construction quality intelligent detection method.

[0016] The beneficial effects of the present invention are as follows: the present invention collects RGB images and converts them into grayscale images, uses the flatness feature fusion method to calculate the flatness feature value; calculates the gap feature coefficient, and uses the accumulation method to calculate the edge clutter index; evaluates the construction quality score, thereby avoiding the limitations of a single feature indicator, improving the efficiency and accuracy of road quality detection, and enhancing the robustness and adaptability of detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0018] Figure 1 The figure is a flow chart of the intelligent detection method for construction quality of low-grade highways.

[0019] Figure 2 It is a structural diagram of the intelligent detection system for construction quality of low-grade highways. DETAILED DESCRIPTION

[0020] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0021] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0022] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

[0023] Example 1, reference Figure 1 , which is the first embodiment of the present invention, and which provides a low-grade highway construction quality intelligent detection method, the low-grade highway construction quality intelligent detection method includes: S1, collect RGB images and convert them into grayscale images, and use the flatness feature fusion method to calculate the flatness feature value; Specifically, collecting RGB images and converting them into grayscale images refers to using a high-resolution CMOS camera to collect RGB images of low-grade highway pavement while the construction vehicle is moving; The preprocessing of the collected RGB images includes: using an electronic image stabilization algorithm to correct geometric distortion of the RGB images, using an adaptive bilateral filtering algorithm to denoise the RGB images, and normalizing the denoised RGB images; The preprocessed RGB image is converted into a grayscale image using the weighted average method.

[0024] The high-resolution CMOS camera can capture tiny details on the construction road surface, including cracks, potholes, unevenness and other features, provide high-definition images, and provide accurate input data for subsequent processing and analysis. Under vehicle vibration, tilt and other dynamic movements, the electronic image stabilization algorithm can eliminate jitter and geometric distortion in the image to ensure the stability and consistency of the collected image. Compared with traditional filtering methods, the adaptive bilateral filtering algorithm can retain road texture details such as crack edges and road granularity while removing environmental noise (such as dust, light spots, etc.) at the construction site, providing an accurate basis for feature extraction. Normalization processing balances the image brightness distribution and reduces the impact of differences in lighting conditions on image analysis. By unifying the pixel value range of the image, the consistency of feature calculation and the robustness of the algorithm are guaranteed. By weighted processing of the three RGB channels, the brightness information and detail features of the road surface can be retained during the conversion to a single grayscale value. The grayscale image calculation complexity is lower than that of the RGB image, which is conducive to improving the subsequent calculation efficiency.

[0025] Furthermore, the flatness feature fusion method is used to calculate the flatness feature value, which means that the fixed sliding window method is used to set the window N, the grayscale image is segmented according to the set window, and the local mean of the window image is calculated. , the formula is: , Where x and y are the pixel coordinates in the window image. is the gray value of the offset pixel, is the total number of pixels in the window, u and v are offsets; Use statistical analysis to calculate the probability of the grayscale value of the window image , the information entropy of the window is calculated based on the probability of the gray value of the window image , the formula is: , Where k is the gray level; Use the Sobel operator to calculate the horizontal gradient of the window image pixels respectively and vertical gradient , use the Euclidean norm method to calculate the gradient amplitude of the window pixel , the formula is: , in and is the grayscale change rate of the pixel in the horizontal and vertical directions; Drawing on the synaptic weight change pattern in biological neural networks, the local unevenness characteristics are calculated by coupling the gradient field with the entropy field, and the unevenness characteristic values ​​are calculated by combining the gradient field with the entropy field. , the formula is: , in is the gradient amplitude of the offset pixel in the window; The gradient field reflects the rate of change of the grayscale value in the image and is an important indicator for detecting edges, contours and details, while the entropy field measures the complexity and randomness of image information. The combination of the two can comprehensively describe the image characteristics from multiple dimensions. The absolute difference in the offset gradient amplitude can accurately measure the intensity of grayscale changes between local pixels. This method directly reflects the local unevenness of the image. By calculating the gradient changes in the local window, the algorithm can enhance its sensitivity to local details. Using other methods (such as simple averaging or maximum and minimum differences) may not accurately capture the characteristics of local unevenness, resulting in the neglect of small changes. The window sliding mechanism can achieve global image unevenness. The coverage of the area ensures the comprehensiveness of the calculation of the unevenness features. The existing methods use a single feature (such as gradient field or entropy field) for image analysis, and fail to make full use of the complementary characteristics of the two. The absolute difference of the offset gradient amplitude is used to enhance the sensitivity of the algorithm to local features, especially in capturing tiny cracks and irregular textures. It has significant advantages. Through the combination of gradient field and entropy field, the local change intensity and information complexity of the image are fully reflected, and the recognition ability of unevenness features is significantly improved. The combination of window sliding mechanism and weight function can dynamically adjust the degree of attention to different areas, and improve the adaptability to the global and local features of the image.

[0026] In the window image, use the Flood Fill algorithm to identify and calculate the grayscale mean of the i-th connected domain , the formula is: , in is the number of pixels in the connected domain, is the gray value of the pixel in the connected domain; Use resolution adjustment to set the mesh edge length ; Calculating the fractal dimension of connected components in windowed images using box counting , the formula is: , in is the number of grids covering the connected domain, is the length of the grid side, indicating the scale; Combine the grayscale mean of the connected domain with the fractal dimension to calculate the convex eigenvalue , the formula is: , in is the number of connected domains in the window, is the fractal dimension of the i-th connected domain, is the local mean of the window image; Based on the unevenness eigenvalue and convex eigenvalues , use the flatness feature fusion method to calculate the flatness feature value of the window image , the formula is: , in is the flatness eigenvalue of the window image.

[0027] Fusion of these two eigenvalues ​​can comprehensively evaluate the smoothness of the road surface from both macroscopic and microscopic levels, ensuring the comprehensiveness and accuracy of the detection results. If only one of the features is used, abnormalities in convexities or local details may be missed, resulting in deviations in the evaluation results. Nonlinear relationships (such as fractional forms) can suppress the influence of extreme eigenvalues, making the smoothness eigenvalues ​​smoother and more robust. Compared with other weighted summation or linear models, this nonlinear fusion method is more suitable for describing roads with high complexity and local variation characteristics. Road surface detection requires analysis of a large area while also capturing local features. The window mechanism calculates eigenvalues ​​by region, which can ensure global coverage while taking into account local details. It not only covers the macroscopic smoothness of the road surface, but also supplements the local convexity and concave features, providing a more comprehensive evaluation index. Linear weighting or simple summation is used to combine features, which is easily disturbed by extreme values ​​and affects the stability of the results. Nonlinear fractional form fusion is used to avoid the influence of eigenvalues ​​that are too large or too small on the results through a normalization mechanism, significantly improving the robustness and stability of the eigenvalues.

[0028] The sliding window method can effectively capture the image features of local areas and avoid the feature blurring problem caused by global processing. By flexibly adjusting the window size, the best balance can be found between feature refinement and computational efficiency. The grayscale value probability distribution reflects the change of pixel values ​​within the region, and the information entropy quantifies the regional complexity. High information entropy areas usually correspond to quality problems such as road cracks and unevenness. The entropy analysis can accurately locate road problems. The gradient amplitude extracted by the Sobel operator can directly reflect the intensity and direction changes of the edges in the image, which is helpful to identify road boundaries and texture features. The gradient amplitude calculated by the Euclidean norm method can fully characterize the smoothness of the road surface. The coupling calculation of the gradient field and the entropy field can integrate the regional grayscale changes and edge strength information, evaluate the road surface flatness from multiple angles, and improve the accuracy and comprehensiveness of feature extraction. The Fill algorithm can accurately separate continuous areas in the road surface, making it easier to count their grayscale mean values, which helps analyze the local structural characteristics of the road surface, such as the depth of the crack area. The fractal dimension can quantify the complexity of the road surface area. Combined with the grayscale mean value, it can describe the shape and intensity of the raised area. The calculation of the raised eigenvalue enhances the sensitivity to road surface details. The flatness feature fusion method combines the feature information of different dimensions through the coupling calculation of the unevenness eigenvalue and the raised eigenvalue to generate a feature value that can accurately quantify the road surface flatness. This method avoids the limitations of a single feature indicator and improves the reliability of the detection results.

[0029] S2, calculating the gap characteristic coefficient based on the flatness characteristic value, and calculating the edge disorder index using the accumulation method; Specifically, the calculation of the gap characteristic coefficient based on the flatness characteristic value refers to the calculation of the gap characteristic coefficient based on the horizontal gradient of the window image. and vertical gradient , use the inverse tangent method to calculate the gradient direction , the formula is: , Take the center pixel in the window image as the reference point and calculate the gradient direction difference of the pixels in the window ; The average angle difference of the direction of the pixel points in the window image is calculated using the direction difference accumulation method; Calculate the consistency index of gradient direction based on the average angle difference ; Perform fast Fourier transform on the window image to extract the energy ratio of high-frequency components in the frequency domain ; Based on the flatness feature value , gradient amplitude , Gradient direction consistency index And the energy ratio of high-frequency components in the frequency domain , calculate the gap characteristic coefficient of the window image , the formula is: , in is the gap characteristic coefficient of the window image, which represents the significance of grayscale change.

[0030] The gap feature coefficient is a comprehensive indicator that combines multiple features and is used to quantify the significance of gaps in the image window. A single feature cannot fully reflect the complexity of the gap. Using linear weighted summation, the correlation and importance between features are difficult to reflect, and the multi-dimensional characteristics of the gap cannot be fully captured. Therefore, nonlinear combination is a necessary method to construct gap feature coefficients. Gaps often contain high-frequency features (such as edges and sharp transition areas), while low-frequency components mainly represent smooth background information. This method can effectively highlight the significance of the gap area and suppress the interference of flat areas on feature calculation. If frequency domain analysis is not introduced, it may lead to the loss of gap detail features. Traditional methods usually evaluate gap characteristics based on a single feature (such as grayscale gradient or edge intensity), resulting in one-sidedness and insufficient accuracy of the results. The four features of flatness, gradient amplitude, directional consistency and frequency domain energy ratio are combined to describe the gap characteristics from global and local perspectives. Many existing methods use simple linear weighting methods, which cannot fully reflect the complex relationship between different features, especially when some eigenvalues ​​have extreme changes, which may affect the stability of the overall results. The gap feature coefficient is constructed through the nonlinear fusion formula of multi-dimensional features, which significantly improves the quantification ability of gap characteristics.

[0031] Calculating the gradient direction through the inverse tangent function method can accurately extract the directional information of each pixel point and fully reflect the change characteristics of each area in the image. The gradient direction provides key data for subsequent directional consistency analysis. By setting the central pixel as the reference point, the interference of environmental noise on local feature analysis can be effectively reduced. The directional difference calculation quantifies the directional change of pixels in the window and provides a basis for consistency analysis. The average angle difference reflects the overall distribution trend of the pixel direction. The directional consistency index can distinguish between normal road surface areas and gap areas by quantifying the regularity of directional changes. FFT can efficiently separate the frequency characteristics of the image and extract the high-frequency component energy ratio to further highlight the edge and detail features. The significant increase in the energy of the high-frequency component is usually closely related to abnormal features such as road cracks. By quantifying the regularity of directional changes, the description ability of the crack boundary is enhanced. The calculation of the gap feature coefficient based on comprehensive multi-dimensional features can comprehensively quantify the significance of the cracks and significantly improve the accuracy and robustness of detection.

[0032] Furthermore, the edge clutter index is calculated using the cumulative method, including: Use the Canny operator to detect and identify the edge of the window image, sort the edge pixels according to their connectivity, and form an ordered sequence of edge points; For edge points in the edge sequence, select the front and back edge points to form a three-point set; Use the Euclidean distance method to calculate the distance between three edge points; Calculate local curvature using the three-point method , the formula is: , Where A is the area of ​​the triangle formed by the three edge points, , and is the distance between the three edge points; Based on local curvature , calculate the curvature change ; Use Euclidean distance to calculate the spatial distance between adjacent edge points , defines the dynamic difference of the distribution between points , the formula is: , in To avoid small quantities with zero denominator; Calculation of local Lyapunov exponents based on chaos metric method , used to quantify the degree of chaos at the edge points, the formula is: , The curvature change rate and Lyapunov index of all edge points are accumulated to calculate the edge disorder index , the formula is: , Where Q is the total number of edge points, and z is the edge point index, which indicates the number of the point currently being processed.

[0033] The core indicator for quantifying the geometric changes of edge points reflects the curvature of the local curve and its changing trend. The Lyapunov index is an important parameter in nonlinear dynamics, which is used to describe the sensitivity of the system under initial value perturbations. The introduction of this index can quantify the dynamic complexity of the distribution of edge points, especially chaos and uncertainty. The joint analysis of the curvature change rate and the Lyapunov index combines the edge geometric characteristics and dynamic characteristics. Using the curvature change rate alone may ignore the distribution law and complex dynamics of the edge points. Using the Lyapunov index alone cannot describe the geometric shape of the edge points. The combination of the two is an important method to comprehensively describe the edge chaos, which is difficult to be replaced by other single features. The curvature change rate and Lyapunov index of all edge points are accumulated point by point to obtain the overall edge chaos index. The cumulative form can capture the characteristic distribution of global edge points and integrate local changes into the global evaluation. In the present invention, if only the eigenvalue of a single point or a local area is used, the detection result may be one-sided and fail to reflect the complexity of the overall edge characteristics. The traversal mechanism can make full use of the information of curvature changes and dynamic characteristics, so as to comprehensively quantify the degree of disorder of the image edge. Traditional edge feature extraction is mainly based on geometric features (such as curvature or gradient), ignoring the dynamic behavior of edge point distribution, especially the analysis of chaos and complexity. By introducing the Lyapunov exponent, the chaos analysis in nonlinear dynamics is combined with geometric features to achieve a comprehensive characterization of the complex characteristics of edge points. The cumulative form realizes the transition from local features to global evaluation by integrating the information of all edge points. The combination of curvature change rate and Lyapunov exponent significantly enhances the robustness of the method through the complementarity of dynamic and geometric characteristics, especially when the edge point distribution is complex or there is strong noise, it can still maintain a high detection accuracy.

[0034] The Canny operator can efficiently extract image edge information and generate an ordered sequence of edge points in combination with connectivity sorting to ensure the logic and accuracy of subsequent calculations. The selection of three-point sets provides a local reference framework for curvature and dynamic characteristic calculations. The Euclidean distance method accurately quantifies the spatial distribution between the three points and provides basic data for subsequent local curvature and dynamic difference calculations. The local curvature can quantify the degree of curvature of the edge points, and the curvature change reflects the sudden change or smoothness of the edge distribution. The curvature change captures the dynamic characteristics of the edge geometry and can effectively identify the irregularity of the road edge. The dynamic difference reveals the compactness and change trend of the edge points by analyzing the spatial distance and distribution law between the edge points. A larger difference indicates that the distribution of the edge points is more disordered. The local Lyapunov index can quantify the chaotic characteristics of the edge point distribution, highlighting the nonlinearity and uncertainty of the dynamic behavior of the edge points. A higher Lyapunov index indicates that the edge point distribution is more complex and messy. By accumulating the curvature change rate and the Lyapunov index, the geometric form and dynamic complexity of the edge point distribution are comprehensively quantified. The edge messiness index can accurately reflect the irregularity and complexity of the road edge.

[0035] S3, evaluate the construction quality score based on the gap characteristic coefficient and edge mess index, and build a visual interface to display the construction quality score; Specifically, the gap characteristic coefficient and edge mess index are used to evaluate the construction quality score. The gap characteristic coefficient is calculated by combining nonlinear activation functions. and edge clutter index Comprehensive characteristics of , the formula is: , in is the Sigmoid activation function; Use Otsu algorithm to set abnormal threshold and integrate comprehensive features Compare with the optimal threshold, and take the comprehensive features greater than or equal to the abnormal threshold The corresponding window image is judged as an abnormal area, and the comprehensive features smaller than the abnormal threshold are The corresponding window image is judged as a normal area; Use sliding window feature superposition method to merge window images into grayscale images; Total area of ​​statistical abnormal areas , Construction quality scoring method based on defect area ratio to evaluate construction quality scoring , the formula is: , in is the total area of ​​the grayscale image.

[0036] The nonlinear activation function can effectively handle the nonlinear relationship between the gap feature coefficient and the edge clutter index. By introducing nonlinear mapping, the comprehensive features can better characterize the complexity of construction quality problems. The Otsu algorithm automatically selects the optimal threshold by maximizing the inter-class variance to avoid the uncertainty of manually setting the threshold. The set abnormal threshold is adaptive and can be automatically adjusted according to different construction scenarios to improve the robustness of detection. By comparing with the abnormal threshold, it can accurately determine whether each window image is an abnormal area, reducing the false detection rate and missed detection rate. The sliding window feature superposition method can gradually integrate the window image into an overall grayscale image to ensure the spatial continuity and visual visualization effect of the abnormal area, reduce the information loss that may be caused by window segmentation, and ensure the fusion of global features and local features. The statistics of the total area of ​​the abnormal area can quantify the scope of defects that occur during the construction process. The defect area ratio is an important indicator that directly reflects the construction quality level and supports timely adjustments and quality improvements during the construction process. The scoring method of the defect area ratio quantifies the construction quality problem into an intuitive scoring result, which is convenient for construction managers to quickly understand and make decisions.

[0037] Further, building a visual interface to display the construction quality score refers to using the front-end technology Chart.js to build a visual interface to display the state of new energy power generation; Users who have passed real-name verification are allowed to view the information.

[0038] Chart.js supports a variety of chart types and can flexibly display construction quality score data according to user needs, such as time series line charts, bar distribution charts and pie charts. It realizes real-time update function through dynamic rendering, so that users can keep abreast of the latest changes in construction quality scores, ensure that only authorized users can access the construction quality score interface, improve the privacy protection ability of the system, establish a binding relationship between user identity and access records through real-name verification, and provide a basis for subsequent behavior analysis and responsibility tracing. It helps users understand the construction quality status from different angles by displaying construction score data in multiple dimensions (such as the changing trend of the time dimension, the score distribution of the regional dimension, etc.). The user-friendly interface design makes complex construction quality score data present in an intuitive way, lowers the threshold for understanding, and supports hierarchical permission control. For example, ordinary users can only view the total score, and administrators can view the detailed score distribution and specific indicators.

[0039] Example 2, reference Figure 2 , which is the second embodiment of the present invention, and which is different from the previous embodiment, provides a low-grade highway construction quality intelligent detection system including: The collection and fusion module is used to collect RGB images and convert them into grayscale images, and use the flatness feature fusion method to calculate the flatness feature value; The calculation accumulation module is used to calculate the gap characteristic coefficient and the edge disorder index using the accumulation method; The evaluation visualization module is used to evaluate the construction quality score and build a visualization interface to display the construction quality score.

[0040] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., which can store program codes.

[0041] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute instructions), or in conjunction with such instruction execution systems, devices or apparatuses. For the purposes of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in conjunction with such instruction execution systems, devices or apparatuses.

[0042] More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk case (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be a paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering or, if necessary, processing in another suitable manner, and then stored in a computer memory.

[0043] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit with a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit with a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

Claims

1. A low-grade highway construction quality intelligent detection method, characterized in that: include, Collect RGB images and convert them into grayscale images, and use the flatness feature fusion method to calculate the flatness feature value; The gap characteristic coefficient is calculated based on the flatness characteristic value, and the edge disorder index is calculated using the accumulation method; The construction quality score is evaluated based on the gap characteristic coefficient and edge messiness index, and a visual interface is constructed to display the construction quality score.

2. The low-grade highway construction quality intelligent detection method according to claim 1, characterized in that: The collecting and converting of RGB images into grayscale images refers to using a high-resolution CMOS camera to collect RGB images of the low-grade road surface while the construction vehicle is moving; The preprocessing of the collected RGB images includes: using an electronic image stabilization algorithm to correct geometric distortion of the RGB images, using an adaptive bilateral filtering algorithm to denoise the RGB images, and normalizing the denoised RGB images; The preprocessed RGB image is converted into a grayscale image using the weighted average method.

3. The intelligent detection method for low-grade highway construction quality as claimed in claim 2, characterized in that: The method of calculating the flatness feature value by using the flatness feature fusion method refers to setting the window N by using the fixed sliding window method, segmenting the grayscale image according to the set window, and calculating the local mean of the window image. ; Use statistical analysis to calculate the probability of the grayscale value of the window image , the information entropy of the window is calculated based on the probability of the gray value of the window image ; Use the Sobel operator to calculate the horizontal gradient of the window image pixels respectively and vertical gradient , use the Euclidean norm method to calculate the gradient amplitude of the window pixel ; Combine the gradient field with the entropy field to calculate the unevenness eigenvalue , the formula is: ， in is the gradient amplitude of the offset pixel in the window, and u and v are the offsets; In the window image, use the Flood Fill algorithm to identify and calculate the grayscale mean of the i-th connected domain ; Use resolution adjustment to set the mesh edge length ; Calculating the fractal dimension of connected components in windowed images using box counting ; Combine the grayscale mean of the connected domain with the fractal dimension to calculate the convex eigenvalue ; Based on the unevenness eigenvalue and convex eigenvalues , use the flatness feature fusion method to calculate the flatness feature value of the window image , the formula is: ， in is the flatness eigenvalue of the window image.

4. The intelligent detection method for low-grade highway construction quality as claimed in claim 3, characterized in that: The calculation of the gap characteristic coefficient based on the flatness characteristic value refers to the calculation of the gap characteristic coefficient based on the horizontal gradient of the window image. and vertical gradient , use the inverse tangent method to calculate the gradient direction ; Take the center pixel in the window image as the reference point and calculate the gradient direction difference of the pixels in the window ; The average angle difference of the direction of the pixel points in the window image is calculated using the direction difference accumulation method; Calculate the consistency index of gradient direction based on the average angle difference ; Perform fast Fourier transform on the window image to extract the energy ratio of high-frequency components in the frequency domain ; Based on the flatness feature value , gradient amplitude , Gradient direction consistency index And the energy ratio of high-frequency components in the frequency domain , calculate the gap characteristic coefficient of the window image , the formula is: ， in is the gap characteristic coefficient of the window image.

5. The intelligent detection method for construction quality of low-grade highway according to claim 4, characterized in that: The edge clutter index is calculated using the accumulation method, comprising: Use the Canny operator to detect and identify the edge of the window image, sort the edge pixels according to their connectivity, and form an ordered sequence of edge points; For edge points in the edge sequence, select the front and back edge points to form a three-point set; Use the Euclidean distance method to calculate the distance between three edge points; Calculate local curvature using the three-point method ; Based on local curvature , calculate the curvature change ; Use Euclidean distance to calculate the spatial distance between adjacent edge points , defining the dynamic difference of the distribution between points ; Calculation of local Lyapunov exponents based on chaos metric method ; The curvature change rate and Lyapunov index of all edge points are accumulated to calculate the edge disorder index , the formula is: ， Where Q is the total number of edge points and z is the edge point index.

6. The intelligent detection method for construction quality of low-grade highway according to claim 5, characterized in that: The gap characteristic coefficient and edge disorder index are used to evaluate the construction quality score. The gap characteristic coefficient is calculated by combining nonlinear activation functions. and edge clutter index Comprehensive characteristics of ; Use Otsu algorithm to set abnormal threshold and integrate comprehensive features Compare with the optimal threshold, and take the comprehensive features greater than or equal to the abnormal threshold The corresponding window image is judged as an abnormal area, and the comprehensive features smaller than the abnormal threshold are The corresponding window image is judged as a normal area; Use sliding window feature superposition method to merge window images into grayscale images; Total area of ​​statistical abnormal areas , Construction quality scoring method based on defect area ratio to evaluate construction quality scoring .

7. The intelligent detection method for low-grade highway construction quality according to claim 6, characterized in that: The construction of a visual interface to display the construction quality score refers to using the front-end technology Chart.js to construct a visual interface to display the state of new energy power generation; Users who have passed real-name verification are allowed to view the information.

8. A low-grade highway construction quality intelligent detection system based on the low-grade highway construction quality intelligent detection method according to any one of claims 1 to 7, characterized in that: include, The collection and fusion module is used to collect RGB images and convert them into grayscale images, and use the flatness feature fusion method to calculate the flatness feature value; A calculation accumulation module is used to calculate the gap characteristic coefficient based on the flatness characteristic value and calculate the edge disorder index using the accumulation method; The evaluation visualization module is used to evaluate the construction quality score based on the gap characteristic coefficient and edge messiness index, and to build a visualization interface to display the construction quality score.

9. A computer device comprising: A memory and a processor; the memory stores a computer program, characterized in that: when the processor executes the computer program, the steps of the low-grade highway construction quality intelligent detection method described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the low-grade highway construction quality intelligent detection method described in any one of claims 1 to 7 are implemented.

Images