A rapid detection and evaluation method for the rolling quality of earth-rock dam filling stones

Abstract

The present invention discloses a method for rapid detection and evaluation of the rolling quality of earth-rock dam filling stone materials, including: obtaining a three-dimensional model of the dam material on the filling bin surface of the earth-rock dam, normalizing the three-dimensional model; converting the format of the three-dimensional model and storing it; converting the obtained three-dimensional model of the dam material into voxel information, and obtaining a voxel octree based on the voxel information and the octree algorithm; obtaining the number of boxes under different box sizes according to the selected box size and the voxel octree as a counting result; drawing a double logarithmic relationship diagram between the box size and the box number based on the counting result, obtaining an estimated value of the fractal dimension through linear fitting, and rapidly detecting and evaluating the rolling quality according to the estimated value. A method for rapid detection and evaluation of the rolling quality of dam material texture information on the filling bin surface of an earth-rock dam based on three-dimensional structured light scanning has high detection efficiency, fast and convenient operation, and is less affected by improper human operation of the operator.

Description

Technical Field

[0001] The invention belongs to the technical field of detecting the rolling effect of filling material on a silo of an earth-rock dam, and in particular relates to a method for quickly detecting and evaluating the rolling quality of filling material of an earth-rock dam. Background Art

[0002] Earth-rock dams mainly use local rockfill as the filling body, so the filling and compaction quality of the filling materials of earth-rock dams is the key to ensuring the quality of the project. The traditional method of testing and evaluating the filling and rolling quality of rockfill materials mainly uses a large number of field tests such as the water injection method to determine the physical indicators of the dam filling stones (such as compaction degree D, porosity, dry density, etc.), which has the disadvantages of being time-consuming and cumbersome operation process. At the same time, since the test is mainly completed by manual operation, it takes up a lot of manpower and time, and the difference between the actual test operation and the standard method will cause the field test results to be inconsistent with the actual. Therefore, how to quickly and conveniently evaluate the rolling effect of the stone at the rolling site of the earth-rock dam filling bin is of great significance to cost control.

[0003] With the continuous development of machine learning, there are also studies on the use of two-dimensional digital images to obtain the surface information of stones at the site of stone rolling in the filling bin of earth-rock dams, obtain their gradation curves after certain processing, and further obtain multiple indicators for evaluating the quality of dam materials to evaluate the rolling effect. However, the use of two-dimensional images to analyze the rolling effect of dam materials has its disadvantages. For example, two-dimensional images cannot provide depth information of objects, and cannot accurately capture the three-dimensional shape and height changes of the stone surface, which may affect the accurate evaluation of the rolling effect.

[0004] In the prior art, a compaction test analysis method based on the settlement rate method has been proposed for evaluating the compaction effect of the fill stone of a concrete face rockfill dam in a reservoir located on a secondary tributary of the Yellow River. This method mainly uses the settlement amount and settlement rate method to measure and evaluate the compaction quality of the rockfill material during the implementation of the project, based on the on-site compaction test of the rockfill material of the face rockfill dam. This method only needs to measure the initial dry density of the rockfill material according to the water irrigation method, and then substitute this data into the analysis formula of the settlement rate method to obtain the dry density at different degrees of compaction, and establish the relationship between the settlement amount and the dry density. After obtaining the settlement at a certain moment, the corresponding density can be obtained, and then the compaction degree of the stone material can be obtained.

[0005] The above method is still completed through experiments. Although the test time is shortened to a certain extent compared with the traditional water filling method, and the results obtained by the water filling method are not much different, the test is still completed manually, which cannot avoid the reduction of the credibility of the experimental results caused by human experimental operation errors.

[0006] The existing intelligent detection technology for the quality of earth-rock dam materials based on digital image processing proposes to segment the digital image of earth-rock dam materials based on the image and gradation data at the dam material detection position, and use the intuitive fuzzy C-means clustering (SIFCM) algorithm that integrates spatial information, and realize the three-dimensional volume reconstruction of earth-rock dam materials using the equivalent ellipsoid volume method. The full gradation characteristic curve of the dam material under real conditions is further obtained after correction by the gradation correction model based on the BP neural network, and then the four major indicators for evaluating the quality of dam materials are obtained. The above technical solution still has several shortcomings:

[0007] (1) The captured images may contain impurities, overlaps, noise, and artifacts, and the images are often affected by inconsistent lighting, resulting in uneven brightness of the images, which makes analysis difficult;

[0008] (2) The image needs to have a fixed focal length or scale during shooting, and there are no obvious boundaries, which will make the analysis of particle size extremely difficult. Summary of the invention

[0009] In order to solve the above technical problems, the present invention proposes a method for quickly detecting and evaluating the rolling quality of earth-rock dam filling stones to solve the problems existing in the above prior art.

[0010] To achieve the above object, the present invention provides a method for quickly detecting and evaluating the rolling quality of earth-rock dam filling stones, comprising:

[0011] Acquire a three-dimensional model of dam materials on the filling bin surface of the earth-rock dam, and perform normalization processing on the three-dimensional model; perform format conversion and storage on the three-dimensional model to obtain a voxel octree;

[0012] Obtaining the number of boxes under different box sizes according to the selected box size and the voxel octree as a counting result;

[0013] A double logarithmic relationship diagram between the box size and the number of boxes is drawn based on the counting results, an estimated value of the fractal dimension is obtained through linear fitting, and the rolling quality is quickly detected and evaluated based on the estimated value.

[0014] Optionally, the process of normalizing the three-dimensional model includes:

[0015] The grid of the three-dimensional model is area-weighted, the number of samples of the surface grid of the three-dimensional model is increased, and the center of gravity of the three-dimensional model is translated to the origin of the standard coordinate system; the maximum distance of the boundary points of the three-dimensional model is obtained, and the size of the three-dimensional model is adjusted according to the maximum distance to achieve normalization of the model.

[0016] Optionally, the process of obtaining the three-dimensional model of the dam material of the earth-rock dam filling bin surface includes:

[0017] Selecting a scanning range, dividing the scanning range to obtain a number of grids, using a sampling method to select a set number of grids as the part to be scanned and performing preprocessing;

[0018] The pre-processed part to be scanned is scanned using three-dimensional structured light to obtain a three-dimensional model of the dam material.

[0019] Optionally, the pretreatment process includes: washing away the loose soil on the surface of the part to be scanned, judging whether the color of the part to be scanned meets the imaging requirements, and if not, evenly and appropriately sprinkling lime powder on the surface of the part to be scanned.

[0020] Optionally, the process of obtaining a voxel octree includes:

[0021] The three-dimensional model is converted into voxelized data, and based on the octree algorithm and the voxelized data, the interface of dam materials with different degrees of rolling is obtained, and the voxelized data is reorganized into an octree structure for storage to obtain a voxel octree.

[0022] Optionally, the process of converting the three-dimensional model into voxelized data includes: selecting the length and shape of the voxel, and using an octree algorithm to recursively divide the three-dimensional model into a number of sub-cubes, wherein during the segmentation process, the length of each edge of the three-dimensional model is compared with the voxel length, and the edges of the three-dimensional model are allocated to voxels based on the comparison results.

[0023] Optionally, the process of obtaining the interface between dam materials and different compaction degrees includes:

[0024] The pixel information of each sub-cube is obtained, and the boundary of dam materials with different compaction degrees is obtained by comparing the pixel information of adjacent sub-cubes; wherein the pixel information includes color and grayscale value.

[0025] Optionally, the process of obtaining the counting result includes:

[0026] The voxel octree is recursively traversed starting from the root node to obtain the number of voxels in each node; several different box sizes are selected, and the number of boxes whose number of voxels contained in all nodes of the voxel octree exceeds the box size is counted respectively to obtain a counting result.

[0027] Compared with the prior art, the present invention has the following advantages and technical effects:

[0028] The present invention provides a method for quickly detecting and evaluating the rolling quality of earth-rock dam filling stone materials, obtaining a three-dimensional model of the dam material on the filling bin surface of the earth-rock dam, normalizing the three-dimensional model; converting the format of the three-dimensional model and storing it; converting the obtained three-dimensional model of the dam material into voxel information, and obtaining a voxel octree based on the voxel information and the octree algorithm; obtaining the number of boxes under different box sizes according to the selected box size and the voxel octree as a counting result; drawing a double logarithmic relationship diagram between the box size and the box number based on the counting result, obtaining an estimated value of the fractal dimension through linear fitting, and quickly detecting and evaluating the rolling quality according to the estimated value. A method for quickly detecting and evaluating the rolling quality of dam material texture information on the filling bin surface of an earth-rock dam based on three-dimensional structured light scanning has high detection efficiency, fast and convenient operation, and is less affected by improper human operation of the operator. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0030] Figure 1 is a flow chart of a method according to an embodiment of the present invention;

[0031] Figure 2 A diagram of the structure of an articulated arm used for three-dimensional scanning according to an embodiment of the present invention;

[0032] Figure 3 A schematic diagram of a scanner bracket used for three-dimensional scanning according to an embodiment of the present invention. DETAILED DESCRIPTION

[0033] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0034] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0035] Embodiment 1

[0036] like Figure 1-3 As shown, this embodiment provides a method for quickly detecting and evaluating the rolling quality of earth-rock dam filling stones, comprising:

[0037] Acquire a three-dimensional model of dam materials on the filling surface of the earth-rock dam, and perform normalization processing on the three-dimensional model;

[0038] In some specific embodiments, the process of obtaining the three-dimensional model of the dam material of the earth-rock dam filling bin surface includes:

[0039] Selecting a scanning range, dividing the scanning range to obtain a number of grids, using a sampling method to select a set number of grids as the part to be scanned and performing preprocessing;

[0040] The pre-processed part to be scanned is scanned using three-dimensional structured light to obtain a three-dimensional model of the dam material.

[0041] In some specific embodiments, the pretreatment process includes: washing away the loose soil on the surface of the part to be scanned, judging whether the color of the part to be scanned meets the requirements, and if not, evenly and appropriately sprinkling lime powder on the surface of the part to be scanned.

[0042] The process of obtaining a 3D model specifically includes:

[0043] Select the scanning area and size: the scanning range is 700mmx700mm, and the rolling site of the earth-rock dam filling bin surface material is divided into 36 equal-sized grids, which are grouped into 4 grids, resulting in 9 groups. The number of samples is set to 9, and one grid is randomly sampled in a group for inspection;

[0044] Process the sample in the selected grid: Before scanning, wash the loose soil on the surface of the sample to reduce the interference of loose soil on the accuracy of 3D imaging. If the color of the sample is dark and scanning is difficult, sprinkle lime powder evenly and in small amounts on it;

[0045] Use 3D structured light to scan the sample: For 3D structured light, set its technical parameters as follows: wavelength 450-480nm; for scanning camera, set its technical parameters as follows: measurement range 800mmx800mm, depth of field 300mm, spatial resolution 0.05-0.1mm, scanning speed 250,000-480,000 measurements / second, interface mode USB interface; for the bracket, use an extendable two-section articulated arm, the end support is placed with a rotatable accessory, the accessory can fix the scanner, so that the scanning angle is 0°-90°, for the support, a tripod is installed at the bottom for easy fixation. When scanning, move the scanner to the vicinity of the sample, adjust the arm length, make the scanner lens within 150-300mm from the sample, adjust the lens angle, and scan.

[0046] In some specific embodiments, the process of normalizing the three-dimensional model includes:

[0047] The grid of the three-dimensional model is area-weighted, the number of samples of the surface grid of the three-dimensional model is increased, and the center of gravity of the three-dimensional model is translated to the origin of the standard coordinate system; the maximum distance of the boundary points of the three-dimensional model is obtained, and the size of the three-dimensional model is adjusted according to the maximum distance to achieve normalization of the model.

[0048] Specifically, considering that the three-dimensional models have different spatial scales, different spatial positions, rotation angles, topological structures, etc., it is necessary to determine a unified three-dimensional model expression method before voxelizing the dam material after three-dimensional scanning, i.e., S2, to ensure the spatial scale, spatial position, and rotation angle of the model. The specific method is as follows:

[0049] ① Solve the problem of the origin position of the model: Use the method of area weighting of the three-dimensional grid to increase the number of sampling of the surface grid of the three-dimensional model, thereby reducing the deviation and the center of gravity of the model of the triangle grid with different grid areas, and the center of gravity of the three-dimensional model can be translated to the origin of the standard coordinate system. The specific expression is as follows.

[0050]

[0051] In the above formula, C represents the center of gravity of the three-dimensional mesh model, P i represents the centroid of each 3D grid in the model, S i Represents the area of ​​each triangular face in the three-dimensional mesh model. Through the above method, the center of gravity of the three-dimensional model can be translated to the origin of the coordinate system;

[0052] ②Solve the problem of different scales of different 3D models: The purpose of scaling preprocessing is to scale models of different scales to a unified scale before performing subsequent feature extraction, which can ensure the uniformity and availability of model features. First, obtain the maximum distance of the model boundary points, and then adjust the size of the model according to the maximum distance of the model boundary, as shown in the following formula:

[0053]

[0054] In the above formula, They are the scaling factors in the x, y, and z axes respectively, K represents the overall scaling factor, and P ix Indicates the distance from the centroid of the i-th triangle to the YOZ plane, P iy Indicates the distance from the centroid of the ith triangle to the XOZ plane, P iz Indicates the distance from the centroid of the i-th triangle to the XOY plane, S i Represents the area of ​​each triangular face in the three-dimensional mesh model, and S is the sum of the areas of all triangular networks on the surface of the three-dimensional network model. By dividing the coordinates of each point in the model by the scaling factor K, the relationship between the scaled coordinates and the original coordinates can scale the three-dimensional model to a uniform scale, thereby achieving normalization of the model.

[0055] Converting the three-dimensional model into a new format and storing the converted model to obtain a voxel octree.

[0056] In some specific embodiments, the process of obtaining a voxel octree includes:

[0057] The three-dimensional model is converted into voxelized data, and based on the octree algorithm and the voxelized data, the interface of dam materials with different degrees of rolling is obtained, and the voxelized data is reorganized into an octree structure for storage to obtain a voxel octree.

[0058] Specifically, the acquired three-dimensional image is converted into voxelized data, and the space is divided into uniform voxels (cubic pixels); the octree algorithm and pixel information are used to determine the junction of dam materials with different rolling degrees; the pixel information of the acquired three-dimensional image of the dam material is organized into an octree structure to perform efficient spatial query and analysis;

[0059] In some specific embodiments, the process of converting the three-dimensional model into voxelized data includes: selecting the length and shape of the voxel, and using an octree algorithm to recursively divide the three-dimensional model into a number of sub-cubes, wherein during the segmentation process, the length of each edge of the three-dimensional model is compared with the voxel length, and the edges of the three-dimensional model are allocated to voxels based on the comparison results.

[0060] Specifically, determine the voxelization parameters, including the size of the voxel (side length), the shape of the voxel (cube or cuboid), etc. When selecting the voxel size, the expected rolling effect can be considered, and generally 3% to 7% of the scanned dam material size range is selected;

[0061] The entire three-dimensional space is divided into a root voxel, and the octree algorithm is used to recursively divide the three-dimensional space of the swept dam material into eight sub-cubes; attention should be paid to the voxelization of the edge of the three-dimensional grid model (i.e., the junction of stones). The specific processing methods for different situations are as follows:

[0062] ① The length of the edge l is less than the length of the voxel w. If the two endpoints of the edge are in the same voxel, the edge can be directly assigned to the voxel; if the two endpoints of the edge are in two different voxels, the edge can be assigned to two adjacent voxels according to a certain ratio;

[0063] ② The length of the edge l is greater than the side length w of the voxel. In this case, the edge needs to be divided equally into i new edges l', so that each l' is less than w, and the first case can be used for processing;

[0064] For each voxel, the volume or shape information it represents is calculated.

[0065] In some specific embodiments, the process of obtaining the dam material interface with different rolling degrees includes:

[0066] The pixel information of each sub-cube is obtained, and the boundary of dam materials with different compaction degrees is obtained by comparing the pixel information of adjacent sub-cubes; wherein the pixel information includes color and grayscale value.

[0067] Taking the grayscale value as a reference indicator, when determining the boundaries of different particle sizes, it is first necessary to record the pixel information of each sub-block, such as color, grayscale value, etc., during the construction of the octree; then, by comparing the pixel information of adjacent sub-blocks, identify the transition areas between stones of different particle sizes. These transition areas will appear as a gradient of grayscale; after determining the junction, it can be distinguished by coloring and the data can be stored.

[0068] Obtaining the number of boxes under different box sizes according to the selected box size and the voxel octree as a counting result;

[0069] In some embodiments, the process of obtaining the counting result includes:

[0070] The voxel octree is recursively traversed starting from the root node to obtain the number of voxels in each node; several different box sizes are selected, and the number of boxes whose number of voxels contained in all nodes of the voxel octree exceeds the box size is counted respectively to obtain a counting result.

[0071] Specifically, starting from the root node, recursively traverse the octree and count the number of voxels in each node; select different box sizes, and count the number of boxes whose number of voxels in all nodes in the octree exceeds the box size under different box sizes; establish a fractal dimension indicator to evaluate the crushing effect:

[0072] A double logarithmic relationship diagram between the box size and the number of boxes is drawn based on the counting results, an estimated value of the fractal dimension is obtained through linear fitting, and the rolling quality is quickly detected and evaluated based on the estimated value.

[0073] The specific method is: divide the grid into small squares, where the scale of the small squares is set to w, count the total number of boxes Nw that the measuring line passes through, and make log Nw-log w -1 If the double logarithmic graph is a linear relationship, its slope is the fractal dimension D.

[0074]

[0075] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for rapid detection and evaluation of the rolling quality of earth-rock dam filling stones, characterized in that: The following steps are involved: Acquire a three-dimensional model of dam materials on the filling surface of the earth-rock dam, and perform normalization processing on the three-dimensional model; Converting the three-dimensional model into a new format and storing the converted model to obtain a voxel octree. Obtaining the number of boxes under different box sizes according to the selected box size and the voxel octree as a counting result; Based on the counting results, a double logarithmic relationship diagram between the box size and the number of boxes is drawn, an estimated value of the fractal dimension is obtained by linear fitting, and the rolling quality is quickly tested and evaluated according to the estimated value; The process of normalizing the three-dimensional model includes: Performing area weighting on the mesh of the three-dimensional model, increasing the number of sampling on the surface mesh of the three-dimensional model, and translating the center of gravity of the three-dimensional model to the origin of the standard coordinate system; obtaining the maximum distance between the boundary points of the three-dimensional model, and adjusting the size of the three-dimensional model according to the maximum distance to achieve normalization of the model; The process of obtaining the three-dimensional model of the dam material of the earth-rock dam filling bin surface includes: Selecting a scanning range, dividing the scanning range to obtain a number of grids, using a sampling method to select a set number of grids as the part to be scanned and performing preprocessing; The pre-processed part to be scanned is scanned using three-dimensional structured light to obtain a three-dimensional model of the dam material; The pretreatment process includes: washing the surface soil of the part to be scanned, judging whether the color of the part to be scanned meets the imaging requirements, and if not, evenly and appropriately sprinkling lime powder on the surface of the part to be scanned; The process of obtaining a voxel octree includes: The three-dimensional model is converted into voxelized data, and based on the octree algorithm and the voxelized data, the interface of the dam material with different rolling degrees is obtained, and the voxelized data is reorganized into an octree structure for storage to obtain a voxel octree; The process of converting the three-dimensional model into voxelized data includes: selecting the length and shape of the voxel, and selecting an octree algorithm to recursively divide the three-dimensional model into a plurality of sub-cubes, wherein during the division process, the length of each edge of the three-dimensional model is compared with the length of the voxel, and the edge of the three-dimensional model is allocated to the voxel according to the comparison result; The process of obtaining the interface between dam materials with different degrees of rolling includes: The pixel information of each sub-cube is obtained, and the boundary of dam materials with different compaction degrees is obtained by comparing the pixel information of adjacent sub-cubes; wherein the pixel information includes color and grayscale value.

2. The method for rapid detection and evaluation of rolling quality of earth-rock dam filling stones according to claim 1 is characterized in that: The process of obtaining the counting results includes: The voxel octree is recursively traversed starting from the root node to obtain the number of voxels in each node; several different box sizes are selected, and the number of boxes whose number of voxels contained in all nodes of the voxel octree exceeds the box size is counted respectively to obtain a counting result.

Images