Method and device for detecting contractility of cement-based additive manufacturing product based on SFM three-dimensional reconstruction

Abstract

The invention discloses a cement-based additive manufacturing product contractibility detection method and device based on SFM three-dimensional reconstruction. The method comprises the following steps: S1, acquiring a feature image of a cement-based additive manufacturing contractible product based on an epipolar geometric constraint condition; s2, processing the feature images in batches, and only retaining feature factors of the feature images; s3, the three-dimensional structure of the cement-based additive manufacturing shrinkage product is reconstructed through the SFM matching feature factors, and a three-dimensional reconstruction model is obtained; s4, correcting the three-dimensional reconstruction model according to actual parameters of the cement-based additive manufacturing shrinkage product, and exporting the corrected three-dimensional reconstruction model as a point cloud model in a point cloud file format; s5, preprocessing the point cloud model; and S6, comparing the preprocessed point cloud model with the point cloud model of the cement-based additive manufacturing original molding product by adopting an M3C2 algorithm, calculating all dot pitch difference values between the two groups of models, and then determining the contractibility of the cement-based additive manufacturing contractible product according to the dot pitch difference values.

Description

Technical Field

[0001] The present invention relates to the field of shrinkage detection of cement-based additive manufacturing products, and in particular to a method and device for shrinkage detection of cement-based additive manufacturing products based on SFM three-dimensional reconstruction. Background Art

[0002] Cement-based additive manufacturing, also known as concrete 3D printing, is one of the representative achievements of the digital transformation and intelligent construction of the construction industry, and is one of the prominent landing points of the digital transformation of the construction industry. Cement-based additive manufacturing breaks through the template limitations of traditional casting cement-based materials, and can form more complex structures in a short time. It has the advantages of simple curing, high degree of freedom, and green environmental protection. Due to its special molding method, additive manufacturing cement-based materials can shorten the production efficiency of construction from months to days, and from days to hours, and can achieve special customization in design without increasing construction costs. Therefore, the molding effect has become a key evaluation indicator for cement-based additive manufacturing products; and the nature of cement-based materials themselves makes shrinkage significantly affect the molding effect during the additive manufacturing process.

[0003] However, the current shrinkage measurement of cement-based additive manufacturing has not achieved accurate numerical evaluation. The existing method of measuring shrinkage through a ruler and determining the shrinkage value of cement-based additive manufacturing through the scale value has large errors and lacks scientificity and universality. The shrinkage evaluation of complex cement-based additive manufacturing products or special customized shapes is still a difficulty to be solved.

[0004] Therefore, in the face of the current problem that the searchability measurement of cement-based additive manufacturing products cannot achieve accurate numerical evaluation, it is a technical problem that needs to be solved urgently to develop a shrinkage detection method and device for cement-based additive manufacturing products that can provide effective numerical feedback. Summary of the invention

[0005] The purpose of the present invention is to provide a shrinkage detection method for cement-based additive manufacturing products based on SFM three-dimensional reconstruction, so as to solve the problem that the existing search measurement of cement-based additive manufacturing products cannot achieve accurate numerical evaluation.

[0006] In order to solve the above technical problems, in a first aspect, the present invention provides a method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction, comprising the steps of:

[0007] S1: Acquisition of characteristic images of cement-based additive manufacturing shrinkage products based on epipolar geometry constraints;

[0008] S2: Batch process feature images and only retain the feature factors of feature images;

[0009] S3: Reconstruct the three-dimensional structure of the cement-based additive manufacturing shrinkage product through SFM matching characteristic factors to obtain a three-dimensional reconstruction model;

[0010] S4: Correcting the three-dimensional reconstructed model according to actual parameters of the cement-based additive manufacturing shrinkage product, and exporting the corrected three-dimensional reconstructed model as a point cloud model in a point cloud file format;

[0011] S5: preprocessing the point cloud model;

[0012] S6: The M3C2 algorithm is used to compare the preprocessed point cloud model with the point cloud model of the original cement-based additive manufacturing product, and all point distance differences between the two groups of models are calculated. Then, the Gaussian distribution model is used to fit the statistics of all point distance differences, and the shrinkage of the cement-based additive manufacturing shrinkage product is determined based on the statistical results.

[0013] Furthermore, step S1 specifically includes:

[0014] The CMOS sampling points are deployed based on the cement-based additive manufacturing shrinkage product as the base point, and then the cement-based additive manufacturing shrinkage product is sampled multiple times continuously under the condition of satisfying the epipolar geometry constraint until the sampling points globally cover the cement-based additive manufacturing shrinkage product;

[0015] The epipolar geometry constraints include: the epipolar line angle between adjacent feature images is 150°, and the pole is located on the surface of the product perpendicular to the sampling point.

[0016] Furthermore, batch processing of feature images specifically includes: batch deleting positioning information in all feature images.

[0017] Furthermore, step S3 specifically includes:

[0018] A global SFM is used to optimize the matching of the feature factors in the feature image. The geometric information of the two views is calculated during the matching. View calibration is performed on the geometric verification image and a search tree is established. Then, the feature factors in the search tree are 3D reconstructed and the coordinate information of the feature factors in the current space is restored to obtain a 3D reconstructed model.

[0019] Furthermore, the correction of the three-dimensional reconstructed model in step S4 specifically includes:

[0020] S41: performing horizontal correction on the three-dimensional reconstructed model;

[0021] S42: Perform geometric correction on the three-dimensional reconstructed model.

[0022] Furthermore, step S41 specifically includes:

[0023] Take the corner points of a quadrilateral on the horizontal plane where the cement-based additive manufacturing shrinkage product is placed, take the first point as the rotation center, take the vector formed between the second point and the third point as the new plane dimension, and use the fourth point to determine the plane, so as to perform horizontal calibration on the cement-based additive manufacturing shrinkage product.

[0024] Further, step S42 specifically includes:

[0025] The scaling ratio is calculated according to the characteristic image mark size and the actual size mapping value of the cement-based additive manufacturing shrinkage product, and the XYZ three-dimensional scale is kept at the same scale; the three-dimensional points of the cement-based additive manufacturing shrinkage product are cropped and eliminated, and the cement-based additive manufacturing shrinkage product is geometrically calibrated.

[0026] Furthermore, step S5 specifically includes:

[0027] S51: Establish an octree for the point cloud model, divide the spatial entity into several regions, and define the cell size and the number of octree voxels at the minimum subdivision level;

[0028] S52: Taking the octree as the target to divide the spatial area, the point cloud model is colored, the Z-axis direction is used as the height ramp color scale arrangement, the color band is linearly and gradually applied to the point cloud model, and the point cloud coordinates containing the color gradual information in the Z-axis direction are obtained, the point cloud coordinates are exported to the scalar domain, and are displayed in a "red-yellow-green-blue" color gradual scale;

[0029] S53: Perform position offset on the point cloud model, perform global offset on the spatial area divided by the octree, and move the center of the bounding box to the current coordinate origin.

[0030] In a second aspect, the present invention provides a shrinkage detection device for cement-based additive manufacturing products based on SFM three-dimensional reconstruction, comprising: a processor and a memory; wherein the memory is used to store computer execution instructions, and when the device is running, the processor executes the computer execution instructions stored in the memory to enable the device to perform the detection method provided in the first aspect above.

[0031] The beneficial effects of the present invention are:

[0032] 1. By adopting the epipolar constraint to collect feature images, the view angle change of adjacent feature images is small, and the corresponding relationship of feature image mapping is maintained, which can reduce the number of points to be matched and improve the matching efficiency.

[0033] 2. By using the SFM algorithm to match image feature factors, it is possible to improve robustness and accuracy while maintaining scalability and efficiency; cement-based additive manufacturing products are small in size, the number of feature image sampling times is large, and the number of points to be matched is large; SFM can be more suitable for the three-dimensional reconstruction of products with such features, ensuring that the three-dimensional reconstructed model has fewer noise points and a higher degree of restoration.

[0034] 3. Through 3D reconstruction of cement-based additive manufacturing products, volume measurement can be performed globally, and non-destructive testing of product volume parameters can be achieved. Point cloud comparison is used to obtain point distance differences, and the Gaussian distribution model is used to statistically calculate the differences. The measurement accuracy is high and the product shrinkage is quantified, providing data reference for the evaluation of the molding performance of cement-based additive manufacturing products and forming effective numerical feedback. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The same reference numerals are used in these drawings to represent the same or similar parts. The exemplary embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0036] Figure 1 The flowchart of one embodiment of the present invention is shown in FIG.

[0037] Figure 2 Schematic diagram of characteristic image sampling of a spiral tower-shaped cement-based additive manufacturing shrinkage product.

[0038] Figure 3 Schematic diagram of image feature factor matching using SFM.

[0039] Figure 4 It is the numerical result of the point distance difference.

[0040] Figure 5 Numerical statistics for contractility detection. DETAILED DESCRIPTION

[0041] In the first aspect, the present invention discloses a method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction, such as Figure 1 As shown, the steps include:

[0042] S1: Acquisition of characteristic images of cement-based additive manufacturing shrinkage products based on epipolar geometry constraints;

[0043] S2: Batch process feature images and only retain the feature factors of feature images;

[0044] S3: Reconstruct the three-dimensional structure of the cement-based additive manufacturing shrinkage product through SFM matching characteristic factors to obtain a three-dimensional reconstruction model;

[0045] S4: Correcting the three-dimensional reconstructed model according to actual parameters of the cement-based additive manufacturing shrinkage product, and exporting the corrected three-dimensional reconstructed model as a point cloud model in a point cloud file format;

[0046] S5: preprocessing the point cloud model;

[0047] S6: The M3C2 algorithm is used to compare the preprocessed point cloud model with the point cloud model of the original cement-based additive manufacturing product, and all point distance differences between the two groups of models are calculated. Then, the Gaussian distribution model is used to fit the statistics of all point distance differences, and the shrinkage of the cement-based additive manufacturing shrinkage product is determined based on the statistical results.

[0048] According to an embodiment of the present application, step S1 specifically includes:

[0049] The CMOS sampling points are deployed with the cement-based additive manufacturing shrinkage product as the base point. Then, under the condition of satisfying the epipolar geometric constraint, the epipolar angle between adjacent feature images is made to be 150°, and the pole is located on the surface of the product perpendicular to the sampling point. The cement-based additive manufacturing shrinkage product is sampled multiple times continuously until the sampling points globally cover the cement-based additive manufacturing shrinkage product. Figure 2 This embodiment uses epipolar constraints to collect feature images, so that the view angle change of adjacent feature images is small, the corresponding relationship of feature image mapping is maintained, the number of points to be matched can be reduced, and the matching efficiency can be improved.

[0050] According to an embodiment of the present application, batch processing of feature images specifically includes: batch deleting positioning information in all feature images, and geographic information includes longitude, latitude, and altitude information; since the CMOS sampling point is accompanied by a geographic information recording module, this embodiment deletes the geographic information attached to the CMOS sampling point and only retains the feature image parameters, thereby removing interference information in the image and reducing the amount of data transmitted to the next stage.

[0051] According to an embodiment of the present application, step S3 specifically includes:

[0052] The global SFM is used to optimize the matching of the feature factors in the feature image. The geometric information of the two views is calculated during the matching. The view calibration is performed on the geometric verification image and a search tree is established. Then, the feature factors in the search tree are reconstructed in three dimensions to restore the coordinate information of the feature factors in the current space. Figure 3 As shown, a 3D reconstructed model is obtained. This embodiment uses the SFM algorithm to match image feature factors, which can improve robustness and accuracy while maintaining scalability and efficiency; cement-based additive manufacturing products are small in size, feature image sampling times are high, and the number of points to be matched is large; SFM can be more suitable for 3D reconstruction of products with such features, ensuring that the 3D reconstructed model has fewer noise points and a higher degree of restoration.

[0053] According to an embodiment of the present application, the correction of the three-dimensional reconstructed model in step S4 specifically includes:

[0054] S41: performing horizontal correction on the three-dimensional reconstructed model;

[0055] S42: Perform geometric correction on the three-dimensional reconstructed model.

[0056] This implementation corrects the scale and horizontal position of the 3D reconstructed model by performing horizontal correction and geometric correction on the 3D reconstructed model, thereby eliminating interference factors in the 3D reconstructed model and ensuring that there are no interference 3D points in subsequent calculations.

[0057] According to an embodiment of the present application, step S41 specifically includes:

[0058] Take the corner points of a quadrilateral on the horizontal plane where the cement-based additive manufacturing shrinkage product is placed, take the first point as the rotation center, take the vector formed between the second point and the third point as the new plane dimension, and use the fourth point to determine the plane, so as to perform horizontal calibration on the cement-based additive manufacturing shrinkage product.

[0059] According to an embodiment of the present application, step S42 specifically includes:

[0060] The scaling ratio is calculated according to the characteristic image mark size and the actual size mapping value of the cement-based additive manufacturing shrinkage product, and the XYZ three-dimensional scale is kept at the same scale; the three-dimensional points of the cement-based additive manufacturing shrinkage product are cropped and eliminated, and the cement-based additive manufacturing shrinkage product is geometrically calibrated.

[0061] According to an embodiment of the present application, step S5 specifically includes:

[0062] S51: Establish an octree for the point cloud model, divide the spatial entity into several regions, define the cell size (the cell size can be defined as 21) and the number of octree voxels (the number of octree voxels can be defined as 2) at the minimum subdivision level 21 );

[0063] S52: Taking the octree as the target to divide the spatial area, the point cloud model is colored, the Z-axis direction is used as the height ramp color scale arrangement, the color band is linearly and gradually applied to the point cloud model, and the point cloud coordinates containing the color gradual information in the Z-axis direction are obtained, the point cloud coordinates are exported to the scalar domain, and are displayed in a "red-yellow-green-blue" color gradual scale;

[0064] S53: Perform position offset on the point cloud model, perform global offset on the spatial area divided by the octree, move the center of the bounding box to the current coordinate origin, and use normalized matrix transformation instead of point cloud registration to improve the efficiency of point cloud alignment.

[0065] According to one embodiment of the present application, in step S6, the shrinkage of the cement-based additive manufacturing product is specifically manifested as the volume shrinkage produced during the solidification and hardening process of the cement-based material. There is a volume difference between the original formed product of the cement-based additive manufacturing (product A) and the shrunken product of the cement-based additive manufacturing after curing (product B). Therefore, the shrinkage of the cement-based additive manufacturing product is detected by numerically evaluating the above-mentioned volume difference.

[0066] The M3C2 algorithm is used to compare the preprocessed point cloud model with the point cloud model of the original cement-based additive manufacturing product, and all point distance differences between the two groups of models are calculated, including: selecting product A point cloud as the core point cloud, taking product A point cloud as a reference for normal vector fitting, controlling the projection depth to be 1 / 2 of the normal vector of product A point cloud, and measuring the distance between the fitting planes of the field point cloud; the normal calculation uses the scalar domain value of product A point cloud, and defines the measurement accuracy with the scalar domain value, that is, the measurement accuracy is calculated with the Z-axis normal vector scaling factor of 1.0; the output carrier selects product B point cloud, and exports the standard deviation information; Figure 4 shown.

[0067] The Gaussian distribution model is used to count the standard deviation information. The mean of the statistical results represents the overall value of the point distance difference, that is, the shrinkage value of the cement-based additive manufacturing product after curing; the chi-square distance of the statistical results represents the difference between the two groups of point cloud models, that is, the shrinkage rate of the cement-based additive manufacturing product after curing. The shrinkage is reflected by the shrinkage value and the shrinkage rate, such as Figure 5 shown.

[0068] In a second aspect, the present invention provides a shrinkage detection device for cement-based additive manufacturing products based on SFM three-dimensional reconstruction, comprising: a processor and a memory; wherein the memory is used to store computer execution instructions, and when the device is running, the processor executes the computer execution instructions stored in the memory to enable the device to perform the detection method provided in the first aspect above.

[0069] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution of the present invention, which should be included in the scope of the claims of the present invention.

Claims

1. A method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction, characterized in that: Includes steps: S1: Acquisition of characteristic images of cement-based additive manufacturing shrinkage products based on epipolar geometry constraints; S2: batch processing the feature images, and only retaining the feature factors of the feature images; S3: Reconstruct the three-dimensional structure of the cement-based additive manufacturing shrinkage product through SFM matching characteristic factors to obtain a three-dimensional reconstruction model; S4: correcting the three-dimensional reconstructed model according to actual parameters of the cement-based additive manufacturing shrinkage product, and exporting the corrected three-dimensional reconstructed model as a point cloud model in a point cloud file format; S5: preprocessing the point cloud model; S6: The M3C2 algorithm is used to compare the preprocessed point cloud model with the point cloud model of the original cement-based additive manufacturing product, and all point distance differences between the two groups of models are calculated. Then, the Gaussian distribution model is used to fit the statistics of all point distance differences, and the shrinkage of the cement-based additive manufacturing shrinkage product is determined based on the statistical results.

2. The method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction according to claim 1 is characterized in that: The step S1 specifically includes: The CMOS sampling points are deployed based on the cement-based additive manufacturing shrinkage product as the base point, and then the cement-based additive manufacturing shrinkage product is sampled multiple times continuously under the condition of satisfying the epipolar geometry constraint until the sampling points globally cover the cement-based additive manufacturing shrinkage product; The epipolar geometry constraint conditions include: the epipolar line angle between adjacent feature images is 150°, and the pole is located on the surface of the product perpendicular to the sampling point.

3. The method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction according to claim 2 is characterized in that: The batch processing of the feature images specifically includes: batch deleting the positioning information in all feature images.

4. The method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction according to claim 3 is characterized in that: The step S3 specifically includes: A global SFM is used to optimize the matching of the characteristic factors in the characteristic image, calculate the geometric information of the two views during the matching, perform view calibration on the geometric verification image and establish a search tree, then perform three-dimensional reconstruction of the characteristic factors in the search tree, restore the coordinate information of the characteristic factors in the current space, and obtain a three-dimensional reconstructed model.

5. The method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction according to claim 4 is characterized in that: In step S4, correcting the three-dimensional reconstructed model specifically includes: S41: performing horizontal correction on the three-dimensional reconstructed model; S42: Perform geometric correction on the three-dimensional reconstructed model.

6. The method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction according to claim 5 is characterized in that: The step S41 specifically includes: Take the corner points of a quadrilateral on the horizontal plane where the cement-based additive manufacturing shrinkage product is placed, take the first point as the rotation center, take the vector formed between the second point and the third point as the new plane dimension, and use the fourth point to determine the plane, so as to perform horizontal calibration on the cement-based additive manufacturing shrinkage product.

7. The method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction according to claim 6, characterized in that: The step S42 specifically includes: The scaling ratio is calculated according to the characteristic image mark size and the actual size mapping value of the cement-based additive manufacturing shrinkage product, and the XYZ three-dimensional scale is kept at the same scale; the three-dimensional points of the cement-based additive manufacturing shrinkage product are cropped and eliminated, and the cement-based additive manufacturing shrinkage product is geometrically calibrated.

8. The method for detecting shrinkage of cement-based additive manufacturing products based on SFM three-dimensional reconstruction according to claim 7, characterized in that: The step S5 specifically includes: S51: establishing an octree for the point cloud model, dividing the spatial entity into a number of regions, and defining a cell size and a number of octree voxels at a minimum subdivision level; S52: Taking the octree as the target, the point cloud model is colorized, the Z-axis direction is used as the height ramp color scale arrangement, and the color band is linearly and gradually applied to the point cloud model to obtain the point cloud coordinates containing the color progressive information in the Z-axis direction, and the point cloud coordinates are exported to the scalar domain, and are displayed in a "red-yellow-green-blue" color scale progressive manner; S53: Perform position offset on the point cloud model, perform global offset on the spatial area divided by the octree, and move the center of the bounding box to the current coordinate origin.

9. A shrinkage detection device for cement-based additive manufacturing products based on SFM three-dimensional reconstruction, characterized in that: include: A processor and a memory; wherein the memory is used to store computer-executable instructions, and when the device is running, the processor executes the computer-executable instructions stored in the memory so that the device performs the detection method described in any one of claims 1-8.

Images