Intelligent identification of road surface pit volume and rapid repairing method

Abstract

The application discloses a kind of intelligent identification road surface pit volume and fast repair method, including preparation stage and processing stage;Preparation stage is exposed to the pit needing repair correct multi-angle shooting and data collection, and the standard density of the multiple gradation cold patching material commonly used by cold patching material is obtained by test.Processing stage is divided into modeling part and calculation part;Modeling part imports software and carries out image matching to the picture of shooting, reconstructs model by set area reconstruction, and then exports the model simplification.Use control point to obtain the thickness numerical value of each point of pit, and then calculate the volume of pit.The volume of pit and the standard density of cold patching material used are known, so the specific cold patching material quality required for this pit repair can be obtained, and then the pit is quickly filled.The new process of the application processes pit, which can improve work efficiency, reduce construction difficulty, save manpower, material resources and time cost.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of road pit repair, and particularly relates to a method for intelligently identifying the volume of a road pit and rapidly repairing the road pit. BACKGROUND

[0002] Asphalt pavement has the advantages of low noise, smooth road surface, comfortable driving and the like, and is widely applied to high-grade highways in China. However, the asphalt pavement is prone to cracking, rutting, bumping, pit and the like under the long-term action of climate environment and vehicle load, among which the pit disease is the most common. If not repaired in time, the material at the bottom of the road surface and the edge material will continue to fall off under the action of load, causing the road surface damage to be further increased, and seriously affecting the driving safety. Therefore, it is extremely important to repair the pit of the asphalt pavement in time.

[0003] The pit repair of the asphalt pavement is generally divided into cold and hot repair materials. At present, the hot repair material is mainly used for pit repair in China. However, the construction of the hot repair material is obviously limited by the temperature and the rain and snow season, especially in the early stage of pit formation. The pits are scattered and the engineering quantity is small, and the hot repair material is difficult to produce in small batches, resulting in problems such as resource waste and environmental pollution. The engineering problems of the hot repair asphalt repair material are very prominent. Therefore, a material that is convenient to construct, easy to store and can still repair the pit under extreme climate conditions is needed, and the cold repair material emerges as the times require.

[0004] There are two important reasons why the cold repair asphalt repair material has not been widely applied. One is that the cold repair asphalt repair material has better social benefits, but increases the unit production cost, and mass production is easy to cause resource waste. The other is that the production and construction process of the cold repair asphalt repair material is not well understood, and the construction quality control of the cold repair asphalt repair material for repairing the pit cannot be achieved.

[0005] The cold cutting and digging repair pit repair technology is the most commonly used technology for repairing the pit of the traditional repair material of the highway pavement. The technology needs to cut a rectangular road surface around the pit by 10 cm. The cut road surface area is larger than the original pit area, which not only consumes time and energy, but also wastes materials. Therefore, it is necessary to simplify the production and construction process, effectively control the construction quality of the cold repair asphalt repair material, keep the balance between production and use, and make the pit repair more convenient and fast. SUMMARY

[0006] The purpose of the present application is to provide a method for intelligently identifying the volume of a road pit and rapidly repairing the road pit. The purpose is to solve the problem of time and energy consumption of the traditional cold cutting and digging repair pit repair technology. The new process of the present application is used to process the pit, which can not only improve the work efficiency, but also reduce the construction difficulty, and save the labor, material and time cost.

[0007] TECHNICAL SOLUTION: The method for intelligently identifying the volume of a road pit and rapidly repairing the road pit comprises the following steps:

[0008] Step 1, in the preparation stage, the pit to be repaired is exposed to correct multi-angle shooting and data collection, and the standard density of various gradation cold patch materials commonly used for cold patch repair is obtained through experiments;

[0009] Step 2, the photographed pictures and collected data are imported into software for image matching, a model is reconstructed through set area reconstruction, and the model is simplified and exported;

[0010] Step 3, based on the simplified model, the thickness values of each point of the pit are obtained using control points, the average value and standard deviation of multiple groups of data are taken, the maximum value in the range of average value plus or minus standard deviation is determined as the final thickness, then the model gray image is filtered and gradient algorithm and edge detection are applied to obtain the pit area, the pit volume is calculated through real volume proportion calibration, the pit volume and the standard density of the cold patch material used are known, and the specific cold patch material quality required for this pit repair is obtained;

[0011] Step 4, according to the specific cold patch material quality required for pit repair, the pit is quickly repaired.

[0012] Further, in step 2, the photographed pictures and collected data are imported into Reality Capture software for image matching, and the align images option in the workflow or alignment is used to match the images.

[0013] In the process of three-dimensional reconstruction using Reality Capture, the software relies on the recognition of color differences between images to determine the corresponding same pixel points in different perspective photos, and then generates a three-dimensional network model by matching the points; when the image quality is low, the shooting coverage is limited, or the target object has high color consistency, which makes it difficult for the automatic matching algorithm to accurately identify the same pixel points, manual intervention of control points is required to optimize the poor control point effect that may occur after computer automatic image matching is accurate, so as to improve the accuracy and reliability of the final three-dimensional model reconstruction;

[0014] Further, in step 2, the reconstruction of the model by the set area specifically includes the following steps:

[0015] Step 2.1, change the perspective in the scene to narrow down the reconstruction area;

[0016] Step 2.2, define the ground plane, click the Define Ground Plane option in the Reconstruction Region, to place the pivot point in the center of the mesh, at the y = 0 position, so that the model will not float above the surface when placed in the AR scene in the future;

[0017] Step 2.3, set the reconstruction region, click the Set Region Automatically option in the Reconstruction Region, to shrink the reconstruction region and remove the excess part, to speed up the reconstruction process;

[0018] Step 2.4, reconstruct the model, click the Calculate Model option in the Workflow, to form the model mesh, and check the normal details of the model;

[0019] Step 2.5, observe whether the mesh is incomplete, if it is incomplete, repeat steps 2.2 to 2.4, otherwise, continue with the following steps;

[0020] Step 2.6, use the Lasso Tool to remove the excess part, first click the Lasso option in the Reconstruction Region, then select the unwanted part on the model, and finally click the Filter Selection option in the Reconstruction Region to remove the selected excess part;

[0021] Step 2.7, simplify the original model to a high polygon model, click the Simplify Tool option in the Reconstruction Region, to reduce the triangle count of the model;

[0022] Step 2.8, click the Smoothing Tool option in the Reconstruction Region, to smooth the model; click the Colorize option in the Workflow, to color the model; click the Unwrap option in the Reconstruction Region, to unfold the model; click the Texturize option in the Reconstruction Region, to add texture;

[0023] Step 2.9, export the model.

[0024] Further, in step 3, the thickness values of the points of the pit groove are selected as the limit value in the Define Distance.

[0025] Further, the rotation model, on the basis of having defined the ground plane, uses the control points to mark the pit groove thickness, clicks the define distance option in the positioning, determines the pit groove thickness specific value, takes the average value and standard deviation of multiple groups of data, and determines the maximum value in the average value plus or minus the standard deviation range as the final calculation thickness.

[0026] The average method is used to average the values of the same pixel position 3 channels RGB, so as to convert the color image of the model established in the above step into a gray scale image, so that the image has only one channel.

[0027] The Gaussian filter operation is performed on the gray scale image, the filter size and the standard deviation are determined, the filter size determines the range of the filter, and the standard deviation determines the weight distribution of the filter; and for each pixel, a matrix of the filter size is taken out with the pixel as the center, and then the new value of each pixel value is calculated, that is, the pixel values around the pixel are multiplied by the corresponding filter weight and summed, and then the new value is assigned to the pixel, the filtering of the pixel is completed, and the above steps are repeated until all pixels are filtered and processed, and finally a smoothed image is obtained, which is used to smooth the image and eliminate noise.

[0028] Further, in step 3, the gradient algorithm applied after the filtering processing of the model gray scale image is specifically: the horizontal and vertical gradients around each pixel point are calculated by using the Sobel operator, that is, two convolution kernels of the Sobel operator are applied; the first convolution kernel is used to calculate the horizontal gradient, and the second convolution kernel is used to calculate the vertical gradient; for each pixel, the two convolution kernels are applied to the pixels around the pixel, then the two results are squared and summed, and the gradient size of the pixel is obtained, which is used to detect the edges in the image; for the result obtained by the gradient algorithm, a non-maximum suppression algorithm is used to remove non-maximum points, that is, whether the point is a local maximum value in the gradient direction is judged, if yes, it is retained, otherwise it is suppressed; for the result after non-maximum suppression, a double-threshold algorithm is used to further distinguish edge and non-edge pixels, and the specific implementation is to divide the image pixels into three categories of strong, weak and non-edge pixels, the strong pixels are likely to be edge pixels, the weak pixels need to be further judged, and the non-edge pixels can be excluded.

[0029] Further, for the weak pixels around the strong pixels, a connection operation is performed to obtain complete edge information, a rasterization method is used to convert an irregular closed figure into a set of pixel points, and then a pixel point coverage area algorithm is used to calculate the pit groove top surface area.

[0030] Further, in step 3, the pit groove volume calculated by the real volume proportion calibration specifically includes the following steps:

[0031] Step 3.1, place a cube with a known volume in the pit picture, use the Reality Capture-based intelligent identification pit method to establish a related model, and perform proportional conversion on the real value of the model and the model annotation value;

[0032] Step 3.2, add a square with a known area to the edge detection picture, and perform proportional conversion on the real value of the pit top surface area after edge detection;

[0033] Step 3.3, bring the proportion into the irregular pit volume calculation to obtain the real volume of the pit.

[0034] Further, in step 3, the cold patch material includes AC-5, AC-10, and AC-13, a plurality of Marshall test pieces of different gradation cold patch materials are prepared, and the standard density of the different gradation cold patch materials is obtained by using the average value method.

[0035] Further, step 4 specifically includes the following steps:

[0036] Step 4.1, pit forming, before pit repair, cut the repair area according to the principle of "round pit square repair and inclined pit straight repair";

[0037] Step 4.2, pit cleaning, clean the loose mixture in the pit to avoid affecting the bonding performance between the new and old interfaces;

[0038] Step 4.3, apply the bonding agent, apply the bonding agent at the bottom and around the pit to improve the bonding performance of the new and old interfaces;

[0039] Step 4.4, determine the amount of cold patch material, according to the actual volume of the pit measured by the Reality Capture software, determine the gradation of the repair material, prepare a plurality of Marshall test pieces of the gradation cold patch material, and obtain the standard density of the cold patch material by using the average value method, so as to determine the most suitable repair material quality for the pit;

[0040] Step 4.5, paving and rolling, when the repair depth exceeds 5 cm, layer compaction, ensure that the cold patch material is leveled and 1-2 cm higher than the original road surface during paving, select appropriate compaction tools according to the pit repair depth and area, and use the principle of rolling from the four corners to the middle during compaction, and apply the bonding agent after compaction to seal the edge;

[0041] Step 4.6, pit maintenance, after rolling and forming, maintain the mixture to form a certain strength, and open the traffic, and sprinkle fine sand or mineral powder on the repaired surface to prevent wheel sticking.

[0042] Further, the binder is made of a water-reactive polyurethane material; the compaction tool is a small vibrating flat rammer for small repair area and a small road roller for large repair area; and the curing time is 3-12 hours.

[0043] Beneficial effects: Compared with the prior art, the present application has the following remarkable advantages: the method for intelligently identifying the volume of a road pit and for quickly repairing the pit, by processing a large amount of image data through Reality Capture software, converting the image 3D to generate a three-dimensional model of the pit, and by the high-level algorithm capable of optimizing the processing process and improving the accuracy and reliability of the model, intelligently identifying the volume of the pit, and then according to the standard density of different gradation materials obtained through experiments, calculating the accurate mass required for repairing the specific pit, compared with the traditional cold cutting and digging pit repair technology which requires cutting a 10cm wide rectangle of road surface around the pit, this process does not require any cutting and digging of the pit, not only saving raw materials and being green and energy-saving, but also having excellent filling effect. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 Figure for the Reality Capture software photo feature point extraction and alignment of the present application;

[0045] Figure 2 Figure for the initial three-dimensional model construction of the Reality Capture software of the present application;

[0046] Figure 3 Figure for the final model construction by adding color and texture of the Reality Capture software of the present application;

[0047] Figure 4 Figure for the polygon construction in the pit of the Reality Capture software of the present application;

[0048] Figure 5 Figure for the pit edge detection effect of the Python algorithm-based Reality Capture software of the present application;

[0049] Figure 6 Figure for the pit repair example of the present application;

[0050] Figure 7 Figure for the process stage step flow chart of the Reality Capture-based intelligent identification pit method of the present application;

[0051] Figure 8 Figure for the irregular pit volume calculation step flow chart of the present application. DETAILED DESCRIPTION

[0052] For further illustrating the technical means adopted by the present application to achieve the predetermined inventive objectives, the specific embodiments of the present application are described in detail below in conjunction with the drawings. Obviously, the specific embodiments described are only a part of the embodiments of the present application, instead of all the embodiments. Based on the embodiments of the present application, all the other embodiments obtained by those skilled in the art without creative efforts belong to the protection scope of the present application.

[0053] The present application provides a method for intelligently identifying the volume of a pit and groove of an asphalt pavement and rapidly repairing the pit and groove, which comprises a preparation stage and a processing stage.

[0054] In the preparation stage, the pit and groove to be repaired is exposed to correct multi-angle shooting and data collection, and the standard density of three commonly used AC-5, AC-10 and AC-13 graded cold patch materials is obtained through experiments.

[0055] The processing stage comprises a modeling part and a calculation part. In the modeling part, the photos are imported into Reality Capture software for image matching, then the model is reconstructed by region reconstruction, and the model is simplified and exported. In the calculation part, the thickness values of each point of the pit and groove are obtained by using control points based on the built model, the average value and the standard deviation of multiple groups of data are taken, the maximum value within the range of the average value plus or minus the standard deviation is determined as the final thickness, then the model gray image is filtered and edge detection is performed to obtain the area of the pit and groove, the volume of the pit and groove is calculated, the standard density of the cold patch material used for the pit and groove is known, and the specific cold patch material quality required for this pit and groove repair can be obtained.

[0056] Based on the above method, a specific pit and groove is selected for repair, and the pit and groove diagram is shown in Figure 6 , and the operation steps are as follows:

[0057] More than 60 photos of the specified object are taken from different angles, and AC-5 graded cold patch material is selected for repair according to the size and position of the pit and groove, and the density of the graded cold patch material used for this pit and groove repair is measured by experiment to be 6.981 g / cm 3 .

[0058] The picture is imported into Reality Capture software, then the feature points of the picture are extracted and the pictures are aligned, and the specific operation is shown in Figure 1 .

[0059] The control points are added, then the viewing angle in the scene is changed, the ground plane is defined, the region is reconstructed, and the model is reconstructed, and the reconstruction result is shown in Figure 2 .

[0060] Observe whether the network is incomplete, if incomplete, repeat the above steps two and three, otherwise, use the lasso tool to remove the excess, and then simplify the high polygon model, then smooth the model, finally color the model, add texture to the model, get the final model diagram, specific model diagram as shown in Figure 3 .

[0061] Using the control points, take the average value and standard deviation of multiple sets of data 1.334±0.024, take the maximum value within the standard deviation range of the average value as the final calculation thickness, recorded as 1.358.

[0062] The pit and groove top surface plan is obtained by the pit and groove edge detection algorithm of the Python software, and the specific case is shown in Figure 5 ;

[0063] The rasterization method is used: the irregular closed figure is converted into a set of pixel points (rasterization), and then the pixel point coverage area algorithm is used to calculate the pit and groove top surface area, which is 506.41.

[0064] A cube with a known volume is placed in the pit and groove picture, and a related model is established by using Reality Capture software. The real value of the model and the model labeled value are scaled in proportion, and the scaling ratio is 1.14. A square with a known area is added to the edge detection picture, and the real value of the pit and groove top surface area is scaled in proportion after edge detection, and the scaling ratio is 1.12, and the specific volume of the pit and groove is (1.358÷1.14cm)×(506.41÷1.12cm 2 )=537.27cm 3 .

[0065] The actual repair material quality required for repairing the pit and groove is calculated as 6.981g / cm 3 ×537.27cm 3 =3750.68187g≈3.751kg.

[0066] The rapid detection device can be suitable for almost all small and medium pit and groove repair.

[0067] Although the present application has been disclosed as above, it is not intended to limit the present application, and any person skilled in the art can make some changes or modifications to the above disclosed technical content without departing from the scope of the technical solution of the present application, and any simple modification, equivalent change and modification of the above embodiments according to the technical essence of the present application are still within the scope of the technical solution of the present application.

Claims

1. A method of rapid repair of a road surface pit, characterized in that, Comprising the following steps: Step 1, in the preparation stage, exposing the pit slot to be repaired to correct multi-angle shooting and data collection, and obtaining the standard density of the commonly used multiple gradation cold repair materials through experiments; Step 2, importing the photographed pictures and collected data into the software for image matching, reconstructing the model through set area reconstruction, and then simplifying and exporting the model; in Step 2, the photographed pictures and collected data are imported into Reality Capture software for image matching, and the Align Images option in the Alignment or Workflow of the working process is used to match the images; in Step 2, the model reconstruction through set area reconstruction specifically comprises the following steps: Step 2.1, changing the perspective in the scene for reducing the reconstruction area; Step 2.2, defining the ground plane, clicking the Define Ground Plane option in the Reconstruction Region, placing the center point pivot point at the center of the network mesh, and y=0, so that the model will not float on the surface when placed in the AR scene in the future; Step 2.3, setting the region automatically, clicking the Set Region Automatically option in the reconstruction region, reducing the reconstruction area, and removing the redundant part to speed up the reconstruction process; Step 2.4, reconstructing the model, clicking the Calculate Model option in the working process, forming the model network, and checking the normal details of the model; Step 2.5, observing whether the network is incomplete, if it is incomplete, repeating Steps 2.2 to 2.4, otherwise, continuing the following steps; Step 2.6, removing the redundant part by using the Lasso Tool, first clicking the Lasso in the reconstruction region, then selecting the part not wanted on the model, and finally clicking the Filter Selection to delete the selected redundant part; Step 2.7, simplifying the original model into a high polygon model, clicking the Simplify Tool option in the reconstruction region, and reducing the triangle count of the model; Step 2.8, clicking the Smoothing Tool option in the reconstruction region for smoothing the model, clicking the Colorize option in the working process for coloring the model, clicking the Unwrap option in the reconstruction region for unfolding the model, and clicking the Texturize option in the reconstruction region for adding texture; Step 2.9, exporting the model; Step 3, based on the simplified model, the thickness values of each point of the pit are obtained using control points, the average value and standard deviation of multiple groups of data are taken, the maximum value in the range of average value plus or minus standard deviation is determined as the final thickness, then the model gray image is filtered and gradient algorithm and edge detection are applied to obtain the area of the pit, the volume of the pit is calculated through real volume proportion calibration, the pit volume and the standard density of the cold patch material used are known, and the specific cold patch material quality required for this time pit repair is obtained; Step 4, according to the specific cold patch material quality required for pit repair, the pit is quickly repaired.

2. The method of claim 1, wherein In step 3, the thickness values of each point of the pit are selected as the limit values in the define distance.

3. The method of claim 1, wherein In step 3, the gradient algorithm applied after filtering the model gray image is specifically: the horizontal and vertical gradients around each pixel point are calculated using the Sobel operator, that is, two convolution kernels of the Sobel operator are applied; the first convolution kernel is used to calculate the horizontal gradient, and the second convolution kernel is used to calculate the vertical gradient; for each pixel, the two convolution kernels are applied to the surrounding pixels respectively, then the squares of the two results are summed to obtain the gradient size of the pixel, which is used to detect the edges in the image; for the result obtained by the gradient algorithm, a non-maximum suppression algorithm is used to remove non-maximum points, that is, whether the point is a local maximum value in the gradient direction is judged, if yes, it is retained, otherwise it is suppressed; for the result after non-maximum suppression, a double threshold algorithm is used to further distinguish edge and non-edge pixels, and the specific implementation is to divide the image pixels into three categories: strong, weak and non-edge pixels, the strong pixels are likely to be edge pixels, the weak pixels need to be further judged, and the non-edge pixels can be excluded.

4. The method of claim 3, wherein the step of applying the patch is performed by a robot. For the weak pixels around the strong pixels, a connection operation is performed to obtain complete edge information, a rasterization method is used to convert the irregular closed figure into a set of pixel points, and then a pixel point coverage area algorithm is used to calculate the pit top surface area.

5. The method of claim 1, wherein the method further comprises the step of: In step 3, the pit volume calculated through real volume proportion calibration specifically includes the following steps: ​ Step 3.1, place a cube with a known volume in the pit picture, use the intelligent identification pit method based on Reality Capture to establish a related model, and perform proportional conversion on the real value of the model and the model annotation value; Step 3.2, add a square with a known area to the edge detection picture, and perform proportional conversion on the real value of the pit top surface area after edge detection; Step 3.3, bring the proportion into the irregular pit volume calculation to obtain the real volume of the pit.

6. The method of claim 1, wherein In step 3, the cold patch material includes AC-5, AC-10 and AC-13, a plurality of Marshall test pieces of different gradation cold patch materials are prepared, and the standard density of different gradation cold patch materials is obtained by using the average value method.

7. The method of claim 1, wherein the method further comprises the step of: Step 4 specifically includes the following steps: ​ Step 4.1, pit forming, before pit repair, the repair area is cut according to the principle of "round pit square repair and inclined pit straight repair"; Step 4.2, pit cleaning, cleaning the loose mixture in the pit to avoid affecting the bonding performance between the new and old interfaces; Step 4.3, applying adhesive, applying adhesive to the bottom and around the pit to improve the bonding performance of the new and old interfaces; Step 4.4, determining the amount of cold patch material, determining the grading of the repair material according to the actual volume of the pit measured by the Reality Capture software, obtaining the standard density of the cold patch material by making a plurality of Marshall test pieces of the grading of the cold patch material, and determining the most suitable repair material quality for the pit; Step 4.5, paving and rolling, when the repair depth exceeds 5 cm, layering and compacting, ensuring that the cold patch material is 1-2 cm higher than the original road surface after paving and leveling, selecting appropriate compaction tools according to the repair depth and area of the pit, and rolling from the periphery to the center during compaction, and applying adhesive to seal the edges after compaction; Step 4.6, pit maintenance, after rolling and forming, curing until the mixture forms a certain strength, then opening traffic, and spreading fine sand or mineral powder on the repaired surface to prevent wheel sticking.

8. The method of claim 7, wherein the method further comprises the step of: The adhesive is made of water-reactive polyurethane material; the compaction tool is a small vibrating plate tamper for small repair areas and a small road roller for large repair areas; the curing time is 3-12 hours.

Images