A method for monitoring road pavement construction quality

Abstract

The present invention relates to the field of image processing and analysis technology, and specifically to a method for monitoring the quality of road pavement construction, which specifically includes obtaining an original road surface image before road construction and a quality control road surface image after road construction; generating a flatness evaluation window corresponding to a pixel point based on the original road surface image and the quality control road surface image, and obtaining a concave and convex edge reflecting a road surface depression or protrusion in the flatness evaluation window, and finally obtaining a flatness evaluation result corresponding to the pixel point based on the analysis of the concave and convex edge, thereby realizing the construction quality evaluation of the road surface area. The technical solution provided by the present disclosure can eliminate the adverse effects of large pieces of asphalt on the flatness evaluation of pixel points through the technical means of image processing, and realize the automated and precise monitoring and evaluation of the road construction quality.

Description

Technical Field

[0001] The present disclosure relates to the technical field of image processing and analysis, and in particular to a method for monitoring the quality of road pavement construction. Background Art

[0002] Pavement quality monitoring is an extremely important step during pavement construction. Through monitoring, problems and defects in the construction process can be discovered and resolved early. Timely discovery and correction of problems can avoid subsequent remedial measures and repeated construction, thereby reducing additional cost expenditures.

[0003] Most existing road construction uses asphalt concrete for paving. The external temperature, the fluidity of asphalt concrete, and the vibration of the road roller will affect the flatness of the road during construction. Therefore, in the process of monitoring the quality of road construction, the flatness of the road surface is an important link. In the implementation process, the existing technology generally calculates the flatness of each pixel point in the road surface acquisition image in its neighborhood, thereby reflecting the road surface flatness corresponding to the overall road surface acquisition image. However, due to the particularity of asphalt paving materials, there are irregularly distributed large and small pieces of asphalt in the paved road surface. Among them, large pieces of asphalt appear to be relatively flat before and after the road construction, which is easy to interfere with the flatness evaluation of nearby pixels, resulting in the acquired road surface flatness often not being able to intuitively reflect the actual situation of the road surface, resulting in evaluation deviation. Summary of the invention

[0004] In order to solve the technical problems raised in the background technology, the purpose of the present disclosure is to provide a method for monitoring the quality of road pavement construction. The technical solutions adopted are as follows:

[0005] A first aspect of the present disclosure provides a method for monitoring road pavement construction quality, which may specifically include the following steps:

[0006] Obtaining original pavement images before pavement construction and quality control pavement images after pavement construction;

[0007] Based on the preset neighborhood window, the pixel points in the quality control road surface image are obtained, and the neighborhood grayscale changes compared with the original road surface image are compared to generate the flatness evaluation window corresponding to the pixel points;

[0008] In the flatness evaluation window, the concave and convex edges corresponding to the depressions and / or convexities of the road surface are obtained, and all the concave and convex edges are taken as a concave and convex edge set;

[0009] According to the concave and convex edge set, the flatness evaluation result corresponding to the pixel point is obtained;

[0010] Based on the flatness evaluation results corresponding to all pixel points in the quality-controlled pavement image, the construction quality evaluation results of the pavement area corresponding to the quality-controlled pavement image are obtained.

[0011] In a possible implementation of the first aspect, the process of acquiring the original road surface image and the quality control road surface image includes the following steps:

[0012] Based on a preset image acquisition device, before the pavement construction begins, a captured image including at least a portion of the pavement construction area is acquired as an original pavement image;

[0013] Based on the preset sensor equipment, continuously sense whether the road construction equipment has completely passed through the road construction area;

[0014] When the sensing device senses that the road construction equipment has completely passed through the road construction area, based on the image acquisition device, a captured image is acquired as a quality control road surface image;

[0015] The image acquisition areas corresponding to the original road surface image and the quality control road surface image are the same.

[0016] In a possible implementation of the first aspect, the process of obtaining the grayscale change of the neighborhood of the pixel point includes the following steps:

[0017] Taking the pixel point as the center, based on the preset neighborhood extension direction, the neighborhood window is divided into multiple neighborhood blocks, and the area of ​​each neighborhood block is the same;

[0018] Obtain the grayscale change between each pixel point in the neighborhood block and the corresponding position of the original road surface image to generate a corresponding pixel grayscale change value;

[0019] The pixel grayscale change values ​​of all pixels in the neighborhood block are summed to obtain the neighborhood grayscale change value corresponding to the neighborhood block. The neighborhood grayscale change includes the neighborhood grayscale change value corresponding to each neighborhood block.

[0020] In a possible implementation of the first aspect, the process of generating the flatness evaluation window includes the following steps:

[0021] Normalize the neighborhood grayscale change value corresponding to each neighborhood block;

[0022] Determine whether the grayscale change values ​​of the neighborhood after normalization are all greater than or equal to the first preset threshold:

[0023] If not, the neighborhood block with the smallest neighborhood grayscale change value is removed from the neighborhood window, and the set of remaining neighborhood blocks is used as the flatness evaluation window.

[0024] In a possible implementation of the first aspect, the process of obtaining the concave-convex edge set includes the following steps:

[0025] Based on the edge detection algorithm, the edge distribution in the flatness evaluation window is obtained, and the edge distribution includes the shape, position and extension direction of each detected pending edge;

[0026] In the flatness evaluation window, according to the relative position relationship between the pending edge and the neighboring edge and the shape of the pending edge, the concave and convex edges are screened out from all the pending edges to obtain a concave and convex edge set.

[0027] In a possible implementation of the first aspect, in the process of screening the concave and convex edges according to the relative position relationship between the edge to be determined and the neighboring edge and the shape of the edge to be determined, the following steps are included:

[0028] Perform linear fitting on the edge to be determined to obtain a fitting edge straight line corresponding to the edge to be determined;

[0029] Obtain the average angle between the fitted edge straight line of the undetermined edge and the fitted edge straight line of the corresponding neighborhood edge, and generate the corresponding first judgment variable according to the average angle. The value of the first judgment variable is in the interval (0,1), and the larger the value of the first judgment variable is, the closer the average angle is to 180°;

[0030] Obtain the similarity between the fitted edge straight line and the corresponding undetermined edge, and generate the corresponding second judgment variable according to the similarity. The value of the second judgment variable is in the interval (0,1), and the larger the value of the second judgment variable is, the higher the similarity between the undetermined edge and the fitted edge straight line is.

[0031] The concave-convex edge evaluation value corresponding to the pending edge is obtained, where the concave-convex edge evaluation value is equal to the product of the first judgment variable and the second judgment variable, and the pending edge whose concave-convex edge evaluation value is greater than the second preset threshold is taken as the concave-convex edge.

[0032] In a possible implementation of the first aspect, the process of obtaining the flatness evaluation result includes the following steps:

[0033] Obtain the corresponding brightness value of each pixel in the quality control road surface image in the HSV space;

[0034] Get the best neighborhood edge corresponding to the concave and convex edges;

[0035] At least one flatness evaluation sequence is constructed between the concave-convex edge and the best neighboring edge, wherein the starting point of the flatness evaluation sequence is a pixel point on the concave-convex edge, the end point of the flatness evaluation sequence is a pixel point on the best neighboring edge, and the flatness evaluation sequence is perpendicular to the fitting edge straight line corresponding to the concave-convex edge;

[0036] According to the change of the brightness value of the flatness evaluation sequence and the weight of each flatness evaluation sequence, the edge flatness corresponding to the concave and convex edge is obtained;

[0037] The flatness evaluation result includes the average value of the edge flatness corresponding to all the concave and convex edges in the flatness evaluation window.

[0038] In a possible implementation of the first aspect, the process of obtaining the weight of the flatness evaluation sequence includes the following steps:

[0039] According to the starting point of the flatness evaluation sequence, an edge segment is intercepted on the concave-convex edge, and the starting point of the flatness evaluation sequence is located at the midpoint of the edge segment;

[0040] Get the fitted edge segment straight line corresponding to the edge segment;

[0041] The weight of the flatness evaluation sequence is calculated according to the angle between the fitted edge segment straight line and the flatness evaluation sequence. The closer the angle between the fitted edge segment straight line and the flatness evaluation sequence is to 90°, the greater the weight of the flatness evaluation sequence.

[0042] In a possible implementation of the first aspect, the process of obtaining the construction quality assessment result includes the following steps:

[0043] Traverse the quality control road surface image to obtain the flatness evaluation results corresponding to all pixel points;

[0044] When the flatness evaluation results corresponding to all pixel points meet the preset flatness standard, a first construction quality evaluation is generated.

[0045] In a possible implementation of the first aspect above, the process of obtaining the construction quality assessment result further includes the following steps:

[0046] When the flatness evaluation results corresponding to at least some of the pixels do not meet the preset flatness standard, obtain the flatness evaluation windows of all the pixels whose flatness evaluation results do not meet the preset flatness standard, so as to merge them to obtain an uneven area set;

[0047] Determine whether the coverage area of ​​the uneven area set exceeds the coverage area of ​​the road construction area of ​​the preset proportion in the quality control road image:

[0048] If not, a second construction quality evaluation is generated;

[0049] If so, a third construction quality evaluation is generated;

[0050] Among them, the road surface flatness corresponding to the first construction quality is better than the road surface flatness corresponding to the second construction quality evaluation, and the road surface flatness corresponding to the second construction quality evaluation is better than the road surface flatness corresponding to the third construction quality.

[0051] Compared with the background technology, the present invention has the following beneficial effects:

[0052] The technical solution provided by the present disclosure can realize automatic and precise monitoring and evaluation of the road construction quality through image processing technology based on the acquired original road surface images and quality control road surface images after the road surface construction is completed. In the specific image processing process, the embodiment provided by the present disclosure combines the specific construction scene of asphalt pavement flattening, targets the influence of different asphalt material shapes on the road surface image, eliminates the adverse influence of large pieces of asphalt on the flatness evaluation of pixel points, and further optimizes the image edge detection and edge classification screening scheme based on the screened flatness evaluation window, which can accurately and efficiently obtain the road surface flatness evaluation results based on the distribution of concave and convex edges in a specific evaluation window, and has a popularizable value. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 According to an embodiment of the present disclosure, a flow chart of a method for monitoring road pavement construction quality is provided.

[0055] Figure 2 According to an embodiment of the present disclosure, a schematic diagram of a scenario for acquiring an original road surface image and a quality control road surface image is provided.

[0056] Figure 3 According to an embodiment of the present disclosure, a flowchart of obtaining grayscale changes in a neighborhood of a pixel is provided.

[0057] Figure 4 According to an embodiment of the present disclosure, a schematic diagram of dividing a neighborhood window to obtain neighborhood blocks is provided.

[0058] Figure 5 According to an embodiment of the present disclosure, a flow chart of obtaining a flatness evaluation window according to a neighborhood grayscale change value is provided.

[0059] Figure 6 According to an embodiment of the present disclosure, a schematic flow chart of obtaining a concave-convex edge set is provided.

[0060] Figure 7 According to an embodiment of the present disclosure, a schematic diagram of a process for screening concave and convex edges according to the relative position relationship between the edge to be determined and the neighboring edge and the shape of the edge to be determined is provided.

[0061] Figure 8 According to an embodiment of the present disclosure, a schematic flow chart of obtaining a flatness evaluation result is provided.

[0062] Fig. 9 A schematic diagram of a flatness evaluation sequence is provided according to an embodiment of the present disclosure.

[0063] Fig.10 According to an embodiment of the present disclosure, a schematic flow chart of obtaining the weight of a flatness evaluation sequence is provided.

[0064] Fig.11 According to an embodiment of the present disclosure, a schematic diagram of an angle between a fitted edge segment straight line and a flatness evaluation sequence is provided. DETAILED DESCRIPTION

[0065] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of a road pavement construction quality monitoring method proposed by the present invention, its specific implementation method, structure, characteristics and effects, in conjunction with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0066] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0067] In view of the problems raised in the background technology, the present disclosure provides a road pavement construction quality monitoring method, which can provide a targeted image processing method for the asphalt pavement paving and flattening process. Specifically, in some embodiments of the present disclosure, Figure 1 FIG. 1 is a flow chart showing a method for monitoring the quality of road pavement construction. Figure 1 As shown, this method may specifically include the following steps:

[0068] Step 100: Acquire the original road surface image before road surface construction and the quality control road surface image after road surface construction. The specific acquisition of the original road surface image and the quality control road surface image will be described later.

[0069] Step 200: Based on the preset neighborhood window, the pixel points in the quality control road surface image are obtained, and the neighborhood grayscale changes compared with the original road surface image are obtained to generate a flatness evaluation window corresponding to the pixel points. It can be understood that the flatness evaluation window different from the preset neighborhood window is generated to eliminate the adverse effects of the large asphalt blocks mentioned in the background technology section on the flatness evaluation of the pixel points. The specific acquisition of the flatness evaluation window will be explained later.

[0070] Step 300: In the smoothness evaluation window, the concave and convex edges corresponding to the depressions and / or protrusions of the road surface are obtained, and all the concave and convex edges are taken as a concave and convex edge set. It is understandable that in the obtained smoothness evaluation window, if the existing edge detection algorithm is directly used to obtain the edge, the concave and convex edges representing the concave and convex conditions of the road surface and the material edges representing the edges of the asphalt material will be obtained at the same time, and the existing edge detection algorithm is difficult to effectively distinguish the concave and convex edges from the material edges. Therefore, it is necessary to further screen the concave and convex edges representing the concave and convex conditions of the road surface through the solution provided by the embodiment of the present disclosure, so as to achieve an accurate evaluation of the smoothness of the road surface. The specific screening of the concave and convex edges will be explained later.

[0071] Step 400: Obtain the flatness evaluation result corresponding to the pixel point according to the concave-convex edge set. Specifically, the plane drop corresponding to the concave-convex edge can be judged by the change in the brightness value of the area near the concave-convex edge, thereby intuitively reflecting the flatness evaluation result after the road surface construction. The specific implementation of the flatness evaluation result will be described later.

[0072] Step 500: Based on the flatness evaluation results corresponding to all the pixels in the quality-controlled road surface image, the construction quality evaluation results of the road surface area corresponding to the quality-controlled road surface image are obtained.

[0073] It can be understood that through the above steps 100 to 500, after the road construction is completed, the road construction quality can be automatically and accurately monitored and evaluated through image processing technology based on the acquired original road surface image and quality control road surface image. In the specific image processing process, the embodiment provided by the present disclosure combines the specific construction scene of asphalt pavement flattening, especially for whether there are uneven conditions on the road surface, based on the screened flatness evaluation window, further optimizes the image edge detection and edge classification screening scheme, and can accurately and efficiently obtain the road surface flatness evaluation result based on the distribution of concave and convex edges in a specific evaluation window. The specific implementation of the above steps 100 to 500 will be further explained and illustrated below.

[0074] In the above embodiment, specifically, in the implementation process of the above step 100, the original road surface image and the quality control road surface image can be automatically acquired by setting an image acquisition device and a sensor device. Figure 2 FIG. 1 shows a schematic diagram of a scenario for obtaining an original road surface image and a quality control road surface image. Figure 2 As shown in the figure, an image acquisition device 001 can be pre-set, and the acquisition area corresponding to the image acquisition device 001 is as follows: Figure 2As shown in part 001a, image acquisition is performed before the pavement construction begins to obtain the original pavement image. It can be understood by those skilled in the art that the acquisition area 001a can be entirely located in the pavement construction area, and the pavement construction effect of the current time can be reflected by regional sampling; the acquisition area 001a can also include the pavement construction area and part of the non-construction area, which can further reflect the construction treatment effect of the boundary of the pavement construction area. Those skilled in the art can select the appropriate sampling area setting according to actual needs and the configuration performance of the image acquisition device, and no limitation is made here.

[0075] In such Figure 2 The scene shown also includes a sensor device 002, the monitoring area of ​​which is aligned with the acquisition area 001a of the image acquisition device 001, for continuously sensing whether the road construction equipment 003, such as a road roller, has completely passed through the road construction area, that is, Figure 2 The acquisition area 001a in the image is shown in FIG. 1 . It can be understood by those skilled in the art that the sensing device 002 can be a sensor such as an infrared sensor that can detect dynamic objects, which is not limited here. Figure 2 In the scene shown, there is an electrical connection and / or data transmission between the sensing device 002 and the image acquisition device 001. When the sensing device 002 senses that the road construction device 003 has completely passed through the road construction area, the image acquisition device 001 can collect images again by aiming at the collection area 001a to obtain a quality control road image. Since the image collection areas corresponding to the original road image and the quality control road image are the same, the road construction quality can be evaluated by the changes in the pixels in the original road image and the quality control road image.

[0076] In the above embodiment, specifically, during the implementation of the above step 200, Figure 3 FIG. 1 shows a schematic diagram of a process for obtaining the grayscale change of a pixel point in the neighborhood. Figure 3 As shown, the following steps may be specifically included:

[0077] Step 201: With the pixel point as the center, based on the preset neighborhood extension direction, divide the neighborhood window into multiple neighborhood blocks, and the area of ​​each neighborhood block is the same. It can be understood that for the selected pixel point, a neighborhood window can be determined in advance according to the preset settings. The neighborhood window is generally configured as a square or a rectangle or other shape similar to the original road surface image, and the selected pixel point is usually located at the center of the neighborhood window. Since the embodiment provided by the present disclosure needs to exclude the influence of large pieces of asphalt on the smoothness calculation, and thus needs to exclude areas in the neighborhood window that may contain large pieces of asphalt, the preset neighborhood window can be divided into blocks first. Specifically, the neighborhood window can be divided into multiple neighborhood blocks with the selected pixel point as the center through multiple neighborhood extension directions to obtain multiple neighborhood blocks of the same area. Exemplarily, Figure 4 A schematic diagram of dividing a neighborhood window to obtain a neighborhood block is shown, such as Figure 4 As shown, the neighborhood window 004 is square, the pixel 005 is located at the center of the neighborhood window 004, and the neighborhood extension directions a1 to a8 divide the neighborhood window 004 into eight neighborhood blocks 0041 of the same size. Those skilled in the art can select the size and shape of the neighborhood window and the division of the neighborhood blocks according to actual needs, which are not limited here.

[0078] Step 202: Obtain the grayscale change between each pixel in the neighborhood block and the corresponding position of the original road surface image to generate a corresponding pixel grayscale change value. It is understandable that when calculating the grayscale change of each pixel in each neighborhood block, the grayscale change of a single pixel can be characterized by the grayscale change of its adjacent neighborhood pixels. Specifically, the pixel grayscale change value between a single pixel in the quality control road surface image and the corresponding position of the original road surface image can be calculated by the following formula:

[0079] ;

[0080] in, Characterizing Neighborhood Blocks in Quality Control Road Images Pixels on Among the 8 corresponding neighboring pixels, Gray value of neighboring pixels; Representation in the original road image and the neighborhood block Pixels on The corresponding position point, among its corresponding 8 neighboring pixels, Gray value of neighboring pixels; Characterizing Neighborhood Blocks Pixels on The pixel grayscale change value.

[0081] Step 203: Sum the pixel grayscale change values ​​of all pixels in the neighborhood block to obtain the neighborhood grayscale change value corresponding to the neighborhood block. Correspondingly, the neighborhood grayscale change of the pixel point includes the neighborhood grayscale change value corresponding to each neighborhood block. It can be understood that for a neighborhood block containing For a neighborhood block of pixels, the corresponding neighborhood grayscale change value can be expressed as . Combining with the actual application scenario, it can be understood that after a large piece of asphalt is rolled by a road roller or other road construction equipment, the shape and size of the main body will not change much, and the corresponding image grayscale change is also small; while for a small piece of asphalt, since there are certain gaps in it, the density of the entire small piece of asphalt will be greater after being rolled by a road roller or other road construction equipment, and the gaps will become smaller or even eliminated, and the corresponding image grayscale change will also be relatively large. Therefore, for the neighborhood grayscale change value In terms of neighborhood grayscale change value The larger the value of, the larger the current neighborhood block The possibility of multiple small pieces of asphalt is relatively greater; the grayscale change value of the neighborhood The smaller the value of, the smaller the current neighborhood block The possibility of large pieces of asphalt on the surface is relatively greater.

[0082] In the above embodiment, further, Figure 5 FIG. 1 shows a flow chart of obtaining a flatness evaluation window according to a neighborhood grayscale change value. Figure 5 As shown, the following steps may be specifically included:

[0083] Step 204: Normalize the neighborhood grayscale change values ​​corresponding to each neighborhood block. It is understandable that in order to uniformly evaluate and screen the grayscale changes of the neighborhood blocks of the same pixel, the obtained neighborhood grayscale change values ​​may be normalized in advance, which is not limited here.

[0084] Step 205: Determine whether the neighborhood grayscale change values ​​after normalization are all greater than or equal to the first preset threshold: If not, remove the neighborhood block with the smallest neighborhood grayscale change value in the neighborhood window, and use the remaining set of neighborhood blocks as the flatness evaluation window. It can be understood that the first preset threshold can be preset based on historical image processing experience. If there is a situation where the neighborhood grayscale change value after normalization is less than the first preset threshold, it means that there are large pieces of asphalt in the neighborhood block. Large pieces of asphalt are less affected by road construction equipment such as rollers and are relatively flat, which can easily interfere with the overall flatness performance of nearby pixels in the road surface flatness evaluation. Therefore, these neighborhood blocks need to be excluded. In the specific implementation of the above step 205, by screening out the neighborhood block with the smallest neighborhood grayscale change value, the neighborhood block that is most likely to contain large pieces of asphalt can be excluded, so that it will not affect the subsequent flatness evaluation.

[0085] In some other optional specific implementations of the above embodiments, on the basis of ensuring the integrity of subsequent concave and convex edge acquisition, neighborhood blocks whose neighborhood grayscale change values ​​do not reach the first preset threshold can also be removed from the neighborhood window to exclude neighborhood blocks with large pieces of asphalt as much as possible. Technical personnel in this field can choose the specific screening method of the flatness evaluation window according to actual needs to eliminate neighborhood blocks with smaller neighborhood grayscale change values, which is not limited here.

[0086] In the above embodiment, specifically, in the specific implementation process of the above step 300, Figure 6 FIG. 4 shows a schematic diagram of a process for obtaining a concave-convex edge set. Figure 6 As shown, the following steps may be specifically included:

[0087] Step 301: Based on the edge detection algorithm, obtain the edge distribution in the smoothness assessment window. The edge distribution includes the shape, position and extension direction of each detected pending edge. In an optional specific implementation of the above embodiment, the Canny edge detection algorithm can be used to extract all edges in the smoothness assessment window of the pixel point. Based on the relevant description of the above embodiment, it can be known that the edges directly obtained by the algorithm include both the concave and convex edges that characterize the unevenness of the road surface and the material edges that characterize the edges of the asphalt material. In order to accurately evaluate the smoothness of the road surface, it is necessary to screen out the required concave and convex edges.

[0088] Step 302: In the flatness evaluation window, according to the relative positional relationship between the to-be-determined edge and the neighboring edge and the shape of the to-be-determined edge, the concave and convex edges are screened out from all the to-be-determined edges to obtain a concave and convex edge set. It can be understood that since the material edge is the edge generated by the extrusion deformation of the asphalt particles themselves during the processing process, it is usually shown as irregularly distributed in the flatness evaluation window in the collected image, and since the asphalt particles are mixed with materials such as sand and gravel, the shape of its material edge in the collected image is usually arc-shaped. As for the concave and convex edges that characterize the concave and convex conditions of the road surface, since the convex and / or concave parts of the road surface often have a certain height variation range, the concave and convex edges shown in the collected image are usually consistent with the extension direction of their neighboring edges, and the shape of the concave and convex edges in the collected image is close to a straight line. Therefore, the concave and convex edges can be screened by the above distinction. The following will provide a specific scheme for screening concave and convex edges.

[0089] In the above embodiment, further, Figure 7 The figure shows a flow chart of screening concave and convex edges according to the relative position relationship between the edge to be determined and the neighboring edge and the shape of the edge to be determined. Figure 7 As shown, the following steps may be specifically included:

[0090] Step 3021: Linear fitting is performed on the pending edge to obtain a fitted edge straight line corresponding to the pending edge. In an optional specific implementation of the above embodiment, a least square method can be used to perform linear fitting on all the pending edges in the flatness evaluation window to obtain a corresponding fitted edge straight line, and the specific implementation of the linear fitting method is not limited here.

[0091] Step 3022: Obtain the average angle between the fitted edge straight line of the pending edge and the fitted edge straight line of the corresponding neighboring edge, and generate the corresponding first judgment variable according to the average angle. It can be understood that if the extension direction of the pending edge and its neighboring edge is more consistent, the possibility that the pending edge is a concave-convex edge is greater. Therefore, the angle between the fitted edge straight line and the neighboring edge can be calculated to determine whether the extension direction of the pending edge and its neighboring edge is consistent. Specifically, the first judgment variable can be calculated by the following formula:

[0092] ;

[0093] in, Fitting edge straight line representing the undetermined edge With the neighborhood The angle between the fitted edge lines of the neighborhood edges, Characterizing the Flatness Evaluation Window The number of fitted edge lines of the undetermined edges in , The average angle between the fitted edge straight line of the undetermined edge and the fitted edge straight line of the corresponding neighborhood edge, Characterization in the flatness evaluation window Fitting edge straight line The first judgment variable corresponding to the pending edge. It can be understood that the value of the first judgment variable is in the interval (0,1), and the larger the value of the first judgment variable, the closer the average angle is to 180°, that is, the extension direction of the pending edge and its neighboring edge tends to be consistent.

[0094] Step 3023: Obtain the similarity between the fitted edge straight line and the corresponding pending edge, and generate the corresponding second judgment variable according to the similarity. It can be understood that if the similarity between the fitted edge straight line and the pending edge is higher, the shape of the pending edge is closer to a straight line, and the corresponding pending edge is more likely to be a concave-convex edge. In the specific implementation process, the similarity between the fitted edge straight line and the pending edge can be determined by calculating the average distance between the points on the fitted edge straight line and the corresponding points on the pending edge. The smaller the average distance, the higher the similarity between the two. Specifically, the second judgment variable can be calculated by the following formula:

[0095] ;

[0096] in, Characterizing the pending edge The number of pixels, Characterize the fitted edge straight line On pixels, and edges to be determined On pixels, Characterize the fitted edge straight line On and pending edge The corresponding pixel point The distance between Characterization in the flatness evaluation window Fitting edge straight line Second judgment variable: It can be understood that the value of the second judgment variable is in the interval (0, 1), and the larger the value of the second judgment variable is, the higher the similarity between the undetermined edge and the fitted edge straight line is.

[0097] It is understandable that the above steps 3022 and 3023 may be executed synchronously or asynchronously during the specific implementation process, which is not limited here.

[0098] Step 3024: Obtain the concave-convex edge evaluation value corresponding to the pending edge, the concave-convex edge evaluation value is equal to the product of the first judgment variable and the second judgment variable, and the pending edge whose concave-convex edge evaluation value is greater than the second preset threshold is taken as the concave-convex edge. Specifically, the concave-convex edge evaluation value can be calculated by the following formula:

[0099] ;

[0100] Understandably, The concave-convex edge evaluation value can be obtained based on the above related description. The closer the value is to 1, the more the current pending edge The more likely it is a concave-convex edge, the desired concave-convex edge can be screened out by setting a second preset threshold. The second preset threshold can be preset and adjusted according to historical image processing experience and is not limited here.

[0101] In the above embodiment, specifically, in the specific implementation process of the above step 400, Figure 8 FIG. 1 shows a schematic diagram of a process for obtaining a flatness evaluation result. Figure 8 As shown, the following steps may be specifically included:

[0102] Step 401: Obtain the corresponding brightness value of each pixel in the quality control road surface image in the HSV space. It is understandable that when there is a concave or convex edge in the flatness evaluation window, a shadow will be generated at the corresponding position in the corresponding quality control road surface image, and the shadow can be characterized by the V channel in the image HSV space, that is, the brightness value. In the above step 401, the brightness value corresponding to each pixel can be obtained by converting the conventional RGB value of the quality control road surface image to the HSV color space.

[0103] Step 402: Obtain the best neighboring edge corresponding to the concave-convex edge. It is understandable that there may be multiple neighboring edges for the concave-convex edge. In order to improve the accuracy of calculating the flatness, the best neighboring edge corresponding to the current concave-convex edge can be determined based on the distance between the neighboring edge and the current concave-convex edge and the consistency of the extension direction, that is, the neighboring edge that is close to the current concave-convex edge and has a consistent extension direction is used as the best neighboring edge.

[0104] Step 403: construct at least one flatness evaluation sequence between the concave-convex edge and the best neighboring edge, wherein the starting point of the flatness evaluation sequence is the pixel point on the concave-convex edge, the end point of the flatness evaluation sequence is the pixel point on the best neighboring edge, and the flatness evaluation sequence is perpendicular to the fitting edge line corresponding to the concave-convex edge. Exemplarily, Fig. 9 A schematic diagram of a flatness evaluation sequence is shown, such as Fig. 9As shown, a flatness evaluation sequence 007 is constructed between the concave-convex edge 0061 and the corresponding best neighbor edge 0062. The flatness evaluation sequence 007 can be represented in the form of a vector, wherein the starting point is a pixel point located on the concave-convex edge 0061, and the end point is a pixel point located on the best neighbor edge 0062, and the flatness evaluation sequence 007 is perpendicular to the fitting edge straight line 0063 corresponding to the concave-convex edge 0061. It can be understood that the brightness change between the concave-convex edge and the best neighbor edge can be characterized by the brightness change of the flatness evaluation sequence, that is, the corresponding edge height difference change can be obtained by the distribution relationship of the shadow, which is used to characterize the flatness of the road surface.

[0105] Step 404: Obtain the edge flatness corresponding to the concave and convex edges according to the brightness value change of the flatness evaluation sequence and the weight of each flatness evaluation sequence. The flatness evaluation result includes the average of the edge flatness corresponding to all the concave and convex edges in the flatness evaluation window. The brightness value change of each flatness evaluation sequence and the acquisition of the weight will be specifically described below.

[0106] In the above step 404, the brightness value change of each flatness evaluation sequence can be characterized by the average value of the brightness value difference of adjacent pixels in the flatness evaluation sequence. Specifically, the corresponding change value of the brightness value change can be calculated by the following formula:

[0107] ;

[0108] in, Characterizing bumpy edges No. A flatness evaluation sequence, Characterizing the Flatness Evaluation Sequence The The brightness value of each pixel, Characterizing the Flatness Evaluation Sequence Middle The brightness value of each pixel, Characterizing the Flatness Evaluation Sequence Length, Characterizing bumpy edges No. The corresponding change value of the brightness value change of the flatness evaluation sequence, The larger the value is, the greater the change in brightness value corresponding to the flatness evaluation sequence is.

[0109] In the above step 404, the weight of each flatness evaluation sequence can be reflected according to the angle between the edge segment corresponding to the flatness evaluation sequence and the flatness evaluation sequence, that is, the more the flatness evaluation sequence tends to be perpendicular to the concave and convex edge, the more it can characterize the brightness change between the concave and convex edge and the best neighboring edge, and the corresponding weight value is larger. Specifically, Fig.10 A schematic diagram of a process for obtaining the weight of a flatness evaluation sequence is shown, such as Fig.10 As shown, the following steps may be specifically included:

[0110] Step 4041: According to the starting point of the flatness evaluation sequence, an edge segment is intercepted on the concave-convex edge, and the starting point of the flatness evaluation sequence is located at the midpoint of the edge segment.

[0111] Step 4042: Obtain the fitted edge segment straight line corresponding to the edge segment.

[0112] Step 4043: Calculate the weight of the flatness evaluation sequence according to the angle between the fitted edge segment straight line and the flatness evaluation sequence. The closer the angle between the fitted edge segment straight line and the flatness evaluation sequence is to 90°, the greater the weight of the flatness evaluation sequence.

[0113] For example, Fig.11 A schematic diagram showing the angle between a fitting edge segment straight line and a flatness evaluation sequence is shown, as Fig.11 As shown, specifically, the starting point corresponding to the flatness evaluation sequence 007 is 007a, the corresponding edge segment is 0061a, the starting point 007a is the midpoint of the edge segment 0061a, the fitting edge line corresponding to the edge segment 0061a is 0064, and the angle between the flatness evaluation sequence 007 and the fitting edge line 0064 is Correspondingly, the weight of the flatness evaluation sequence can be calculated by the following formula:

[0114] ;

[0115] in, The characterization angle is 90°, Characterizing the Flatness Evaluation Sequence The fitted edge segment straight line corresponding to the corresponding edge segment The angle between Characterizing the Flatness Evaluation Sequence The weight of .

[0116] In the above step 404, the edge flatness corresponding to the concave and convex edges can be calculated by the following formula:

[0117] ;

[0118] in, Characterizes the number of flatness evaluation sequences, Characterizing bumpy edges The corresponding edge flatness, The larger the value, the more concave and convex the edge. The more convex or concave, the more uneven the corresponding road surface area is.

[0119] Furthermore, by evaluating the flatness window All the concave and convex edges in The smoothness of the edge Traverse to get pixel points Corresponding flatness evaluation results ,Right now ,in, Flatness evaluation window The concave and convex edges The number of

[0120] It is understandable that the flatness evaluation results The larger the value, the smaller the pixel Corresponding flatness evaluation window The smoother.

[0121] In the above embodiment, specifically, in the specific implementation process of the aforementioned step 500, based on the relevant description of the aforementioned embodiment, the quality control road surface image can be traversed to obtain the flatness evaluation results corresponding to all pixel points; and then the corresponding quality evaluation is performed according to the flatness evaluation results.

[0122] In an optional specific implementation of the embodiment of the present disclosure, it can be configured to generate a first construction quality evaluation when the flatness evaluation results corresponding to all pixel points meet the preset flatness standard. When the flatness evaluation results corresponding to at least a part of the pixel points do not meet the preset flatness standard, the flatness evaluation windows of the pixel points whose flatness evaluation results do not meet the preset flatness standard are further obtained to merge to obtain a set of uneven areas. By judging whether the coverage area of ​​the set of uneven areas exceeds the coverage area of ​​the road construction area of ​​a preset proportion in the quality control road image: if not, a second construction quality evaluation is generated; if so, a third construction quality evaluation is generated. Among them, the road surface flatness corresponding to the first construction quality is better than the road surface flatness corresponding to the second construction quality evaluation, and the road surface flatness corresponding to the second construction quality evaluation is better than the road surface flatness corresponding to the third construction quality, so that the final accurate monitoring of the graded road surface quality after construction can be achieved.

[0123] In summary, the technical solution provided by the present disclosure can realize automatic and precise monitoring and evaluation of the road construction quality through image processing technology after the road construction is completed, based on the acquired original road surface images and quality control road surface images. In the specific image processing process, the embodiment provided by the present disclosure combines the specific construction scene of asphalt pavement flattening, targets the influence of different asphalt material shapes on the road surface image, eliminates the adverse effect of large pieces of asphalt on the flatness evaluation of pixel points, and further optimizes the image edge detection and edge classification screening scheme based on the screened flatness evaluation window, which can accurately and efficiently obtain the road surface flatness evaluation result based on the distribution of concave and convex edges in a specific evaluation window, and has a value that can be promoted.

[0124] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0125] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for monitoring road pavement construction quality, characterized in that: The method comprises: Obtaining original pavement images before pavement construction and quality control pavement images after pavement construction; Based on a preset neighborhood window, a pixel point in the quality control road surface image is obtained, and a neighborhood grayscale change compared with the original road surface image is obtained to generate a flatness evaluation window corresponding to the pixel point; In the flatness evaluation window, the concave and convex edges corresponding to the depressions and / or protrusions of the road surface are obtained, and all the concave and convex edges are taken as a concave and convex edge set; Obtaining a flatness evaluation result corresponding to the pixel point according to the concave-convex edge set; Based on the flatness evaluation results corresponding to all the pixel points in the quality-controlled road surface image, obtaining the construction quality evaluation result of the road surface area corresponding to the quality-controlled road surface image; The process of obtaining the flatness evaluation result includes the following steps: Obtaining the corresponding brightness value of each pixel in the quality control road surface image in the HSV space; Obtaining the best neighborhood edge corresponding to the concave and convex edge; At least one flatness evaluation sequence is constructed between the concave-convex edge and the best neighboring edge, wherein the starting point of the flatness evaluation sequence is a pixel point on the concave-convex edge, the end point of the flatness evaluation sequence is a pixel point on the best neighboring edge, and the flatness evaluation sequence is perpendicular to the fitting edge straight line corresponding to the concave-convex edge; Obtaining the edge flatness corresponding to the concave and convex edge according to the brightness value change of the flatness evaluation sequence and the weight of each of the flatness evaluation sequences; The flatness evaluation result includes an average value of edge flatness corresponding to all the concave and convex edges in the flatness evaluation window.

2. The road pavement construction quality monitoring method according to claim 1, characterized in that: The process of acquiring the original road surface image and the quality control road surface image includes the following steps: Based on a preset image acquisition device, before the pavement construction begins, a captured image including at least a portion of the pavement construction area is acquired as the original pavement image; Based on the preset sensing equipment, continuously sense whether the road construction equipment has completely passed through the road construction area; When the sensing device senses that the road construction equipment has completely passed through the road construction area, based on the image acquisition device, acquiring a captured image as the quality control road image; The image acquisition areas corresponding to the original road surface image and the quality control road surface image are the same.

3. The road pavement construction quality monitoring method according to claim 1, characterized in that: The process of obtaining the grayscale change of the neighborhood of the pixel point includes the following steps: Taking the pixel point as the center and based on a preset neighborhood extension direction, dividing the neighborhood window into a plurality of neighborhood blocks, each of which has the same area; Obtaining the grayscale change between each pixel point in the neighborhood block and the corresponding position of the original road surface image to generate a corresponding pixel grayscale change value; The pixel grayscale change values ​​of all pixels in the neighborhood block are summed to obtain a neighborhood grayscale change value corresponding to the neighborhood block, and the neighborhood grayscale change situation includes a neighborhood grayscale change value corresponding to each of the neighborhood blocks.

4. The road pavement construction quality monitoring method according to claim 3 is characterized in that: The process of generating the flatness evaluation window includes the following steps: Normalizing the neighborhood grayscale change value corresponding to each of the neighborhood blocks; Determine whether the grayscale change values ​​of the neighborhood after normalization are all greater than or equal to a first preset threshold: If not, the neighborhood block with the smallest neighborhood grayscale change value is removed from the neighborhood window, and the remaining set of the neighborhood blocks is used as the flatness evaluation window.

5. The road pavement construction quality monitoring method according to claim 1, characterized in that: The process of obtaining the concave-convex edge set includes the following steps: Based on the edge detection algorithm, the edge distribution in the flatness evaluation window is obtained, wherein the edge distribution includes the shape, position and extension direction of each detected pending edge; In the flatness evaluation window, according to the relative position relationship between the pending edge and the neighboring edge and the shape of the pending edge, the concave and convex edges are screened out from all the pending edges to obtain the concave and convex edge set.

6. The road pavement construction quality monitoring method according to claim 5, characterized in that: In the process of screening the concave and convex edges according to the relative positional relationship between the edge to be determined and the neighboring edge and the shape of the edge to be determined, the following steps are included: Performing linear fitting on the undetermined edge to obtain a fitting edge straight line corresponding to the undetermined edge; Obtaining an average angle between the fitted edge straight line of the undetermined edge and the corresponding fitted edge straight line of the neighborhood edge, and generating a corresponding first judgment variable according to the average angle, wherein the value of the first judgment variable is in the interval (0, 1), and the larger the value of the first judgment variable is, the closer the average angle is to 180°; Obtaining the similarity between the fitted edge straight line and the corresponding undetermined edge, and generating a corresponding second judgment variable according to the similarity, wherein the value of the second judgment variable is in the interval (0, 1), and the larger the value of the second judgment variable is, the higher the similarity between the undetermined edge and the fitted edge straight line is; A concave-convex edge evaluation value corresponding to the pending edge is obtained, where the concave-convex edge evaluation value is equal to the product of the first judgment variable and the second judgment variable, and the pending edge whose concave-convex edge evaluation value is greater than a second preset threshold is taken as the concave-convex edge.

7. The road pavement construction quality monitoring method according to claim 1, characterized in that: The process of obtaining the weight of the flatness evaluation sequence includes the following steps: According to the starting point of the flatness evaluation sequence, an edge segment is intercepted on the concave-convex edge, and the starting point of the flatness evaluation sequence is located at the midpoint of the edge segment; Obtain a fitted edge segment straight line corresponding to the edge segment; The weight of the flatness evaluation sequence is calculated according to the included angle between the fitted edge segment straight line and the flatness evaluation sequence. The closer the included angle between the fitted edge segment straight line and the flatness evaluation sequence is to 90°, the greater the weight of the flatness evaluation sequence.

8. The road pavement construction quality monitoring method according to claim 1, characterized in that: The process of obtaining the construction quality assessment result includes the following steps: Traversing the quality-controlled road surface image to obtain the flatness evaluation results corresponding to all the pixel points; When the flatness evaluation results corresponding to all the pixel points meet the preset flatness standard, a first construction quality evaluation is generated.

9. The road pavement construction quality monitoring method according to claim 8, characterized in that: The process of obtaining the construction quality assessment result also includes the following steps: When the flatness evaluation results corresponding to at least a portion of the pixel points do not meet the preset flatness standard, obtain flatness evaluation windows of all pixel points whose flatness evaluation results do not meet the preset flatness standard, so as to merge them to obtain an uneven area set; Determine whether the coverage area of ​​the set of uneven areas exceeds the coverage area of ​​the road construction area of ​​a preset proportion in the quality control road image: If not, a second construction quality evaluation is generated; If so, a third construction quality evaluation is generated; Among them, the road surface flatness corresponding to the first construction quality is better than the road surface flatness corresponding to the second construction quality evaluation, and the road surface flatness corresponding to the second construction quality evaluation is better than the road surface flatness corresponding to the third construction quality.

Images