Radioactive substance transportation container point cloud processing method, system, equipment and medium

Abstract

The invention discloses a radioactive substance transportation container point cloud processing method, system, equipment and medium, relates to the technical field of three-dimensional point cloud processing and industrial detection, and solves the technical problems of low precision and poor efficiency of three-dimensional point cloud feature extraction. A three-dimensional simplification problem is reduced to a two-dimensional edge detection problem, the processing time is shortened, and large-scale point clouds are efficiently processed. In addition, the unicity of the conventional projection direction can cause the point cloud to lack key features, and the full coverage and integrity can be realized by combining the three-view projection. The method can be directly used for geometric feature fitting, geometric parameter measurement and mechanical simulation optimization of the radioactive substance transportation container, provides reliable data support for container design verification and safety evaluation, and shortens the research and development period of nuclear energy engineering.

Description

Technical Field

[0001] This application relates to the field of three-dimensional point cloud processing and industrial inspection technology, and in particular to a point cloud processing method, system, equipment and medium for radioactive material transport containers. Background Technology

[0002] After the manufacture of radioactive material transport containers, their geometric parameters need to be measured to verify their processing accuracy. Furthermore, reliability assessments following drop tests require high-precision three-dimensional deformation measurement data. While traditional three-dimensional measurement methods (such as laser triangulation) can acquire surface point clouds, the large size of the containers results in excessively large point cloud volumes (hundreds of thousands to millions of points). Therefore, point cloud simplification methods are needed to retain key feature points for measurement. Consequently, an efficient and high-precision three-dimensional feature extraction method is urgently required.

[0003] Existing point cloud feature extraction methods, such as curvature-based local feature extraction and 2D edge detection inverse mapping, are exemplified by patent application CN 119992133 A, which discloses an efficient method for processing measurement point clouds based on similar array features. This method identifies feature regions through clustering, identifies edge shapes based on normal vector analysis, accurately determines feature regions based on edge shapes, and simplifies the point cloud through voxel segmentation of edge shapes. However, this method suffers from drawbacks such as feature redundancy, long computation time, and poor robustness.

[0004] Therefore, this application aims to address the contradiction between large-scale point cloud processing and geometric feature fidelity, and improve the accuracy and efficiency of 3D point cloud feature extraction. Summary of the Invention

[0005] This application provides a method, system, device and medium for point cloud processing of radioactive material transport containers. Its technical purpose is to coordinate the contradiction between large-scale point cloud processing and geometric feature fidelity, and improve the accuracy and efficiency of three-dimensional point cloud feature extraction.

[0006] The above-mentioned technical objective of this application is achieved through the following technical solution: A method for processing point clouds of radioactive material transport containers includes: Construct the minimum bounding cube of the container's 3D point cloud; The safety boundary of the minimum bounding cube is extended, and then the xoy plane, yoz plane and xoz plane in the three views are selected to perform two-dimensional projection on the minimum bounding cube to obtain three projected two-dimensional images. The local coordinate system of the minimum bounding cube in the projected two-dimensional images is obtained. The pixel coordinates of the projected two-dimensional image are calculated based on the local coordinate system, and then the gray value of the corresponding pixel is calculated based on the pixel coordinates. Construct a multi-objective optimization function to optimize the resolution of the projected 2D image and obtain the optimal resolution of the projected 2D image; Edge detection is performed on the projected 2D image based on grayscale values ​​to obtain the edge pixel coordinates; The edge pixel coordinates are normalized according to the optimal resolution; The physical plane coordinates of the projected two-dimensional image are calculated based on the normalized edge pixel coordinates; The validity of the corresponding depth values ​​is checked based on the physical plane coordinates to obtain the final 3D point cloud feature points.

[0007] Preferably, the original point cloud coordinates are set as follows: The coordinates of the eight vertices of the smallest enclosing cube are... x 1~ x 8 is represented as: ; Among them, subscript min and max These represent the minimum and maximum values ​​in the point cloud coordinates, respectively; x i ,y i ,z i ) ∈R 3 , R Represents the set of real numbers. i Describes the th smallest enclosing cube. i One point, ( x i ,y i ,z i ) represents the smallest enclosing cube. i The coordinates of the points o Indicates the origin.

[0008] The expansion of the safety boundary of the minimum bounding cube includes: if the boundary length of the original minimum bounding cube is... L 0 The expanded security boundary is then... .

[0009] Preferably, the local coordinate system is represented as: ; in, d This represents the perpendicular distance from the 3D point cloud to the projected 2D plane, i.e., the projection distance; The pixel coordinates of the projected two-dimensional image are represented as follows: ; in,( i,j ) represents a pixel in a projected two-dimensional image, u i ,v i ) represents the th smallest bounding cube. i The local coordinates of each point in the projected two-dimensional image u max 、v max This represents the maximum value of the local coordinates. u min 、v min This represents the minimum value of the local coordinates. n×n Indicates the resolution of the projected two-dimensional image; Any pixel in the projected two-dimensional image ( i,j The corresponding grayscale value Represented as: .

[0010] Preferably, the objective optimization function is expressed as: ; in, ; This indicates the optimal resolution of the projected two-dimensional image; This represents a quality evaluation function for a projected two-dimensional image. An index representing the sharpness of a projected two-dimensional image. This indicates the noise level of the projected two-dimensional image. This represents the noise penalty coefficient, and λ ∈ (0~1).

[0011] Preferably, the normalization of edge pixel coordinates based on the optimal resolution is expressed as follows: ; in, Represents the edge pixel coordinates.

[0012] Preferably, the physical plane coordinates are represented as: ; in, Represents physical plane coordinates.

[0013] Preferably, the step of performing validity detection on the corresponding depth values ​​based on physical plane coordinates to obtain the final 3D point cloud feature points includes: The effective mask function is defined as follows: ; in, Represents the physical coordinate depth value; The validity of the depth value corresponding to the physical plane coordinates is determined by using an effective mask function. If valid, the depth value is calculated, and the inverse mapping process from 2D edge point features to 3D key feature points is achieved using the valid depth value, resulting in the final 3D point cloud feature points, represented as follows: .

[0014] A point cloud processing system for radioactive material transport containers, comprising: A cube building block is used to construct the minimum bounding cube of a container's 3D point cloud. The projection unit extends the safety boundary of the minimum bounding cube, and then selects the xoy plane, yoz plane and xoz plane in the three views to perform two-dimensional projection on the minimum bounding cube to obtain three projected two-dimensional images, and obtains the local coordinate system corresponding to the minimum bounding cube in the projected two-dimensional images; The first calculation unit calculates the pixel coordinates of the projected two-dimensional image based on the local coordinate system, and then calculates the gray value of the corresponding pixel based on the pixel coordinates. The objective optimization unit constructs a multi-objective optimization function, which optimizes the resolution of the projected 2D image to obtain the optimal resolution of the projected 2D image. The edge detection unit performs edge detection on the projected two-dimensional image based on grayscale values ​​to obtain the edge pixel coordinates. The normalization unit normalizes the edge pixel coordinates according to the optimal resolution; The second calculation unit calculates the physical plane coordinates of the projected two-dimensional image based on the normalized edge pixel coordinates. The validity detection unit performs validity detection on the physical coordinate depth value based on the physical plane coordinates to obtain the final 3D point cloud feature points.

[0015] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of a method for processing point clouds of radioactive material transport containers.

[0016] A computer storage medium storing a computer program, which, when executed by a processor, implements the steps of a point cloud processing method for radioactive material transport containers.

[0017] The above technical solution can achieve at least some of the following technical effects: By combining three-view projection directions to rapidly project 2D images, the simplification problem of 3D modeling is reduced to a 2D edge detection problem, decreasing processing time and enabling efficient processing of large-scale point clouds. Furthermore, the uniformity of conventional projection directions leads to missing key features in the point cloud (due to occlusion), while combining three-view projection achieves full coverage and integrity. The high robustness of 2D edge detection replaces the high error of traditional 3D point cloud feature detection, improving the accuracy of 3D point cloud feature detection and providing a foundation for subsequent high-precision detection of geometric parameters.

[0018] By combining multi-view projection with 2D edge detection and optimizing the resolution of the projected 2D image through a target optimization function, the accuracy of 3D point cloud detection is improved while reducing processing time. This method can be directly applied to geometric feature fitting, geometric parameter measurement, and mechanical simulation optimization of radioactive material transport containers, providing reliable data support for container design verification and safety assessment, and shortening the nuclear energy engineering R&D cycle. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the container 3D point cloud data acquired by the laser 3D scanner in an embodiment of this application; Figure 2 This is a schematic diagram of a two-dimensional projected image of the xoy view in an embodiment of this application; Figure 3 This is a schematic diagram of the edge detection results in an embodiment of this application; Figure 4 This is a simplified point cloud diagram of an embodiment of this application. Detailed Implementation

[0020] The technical solution of this application will be described in detail below with reference to the accompanying drawings.

[0021] The point cloud processing method for radioactive material transport containers described in this application includes: 100: Construct the minimum bounding cube of the container's 3D point cloud and extend the safe boundary of the minimum bounding cube.

[0022] Preferably, the original point cloud coordinates are set as follows: The coordinates of the eight vertices of the smallest enclosing cube are... x 1~ x 8 is represented as: (1) Among them, subscript min and max These represent the minimum and maximum values ​​in the point cloud coordinates, respectively; x i ,y i ,zi ) ∈R 3 , R Represents the set of real numbers. i Describes the th smallest enclosing cube. i One point, ( x i ,y i ,z i ) represents the smallest enclosing cube. i The coordinates of the points o Indicates the origin.

[0023] 101: Extend the safety boundary of the minimum enclosing cube, and then select the xoy plane, yoz plane and xoz plane in the three views to perform two-dimensional projection on the minimum enclosing cube to obtain three projected two-dimensional images. Obtain the local coordinate system corresponding to the minimum enclosing cube in the projected two-dimensional images.

[0024] Preferably, the expansion of the safety boundary of the minimum enclosing cube includes: if the boundary length of the original minimum enclosing cube is... L 0 The expanded security boundary is then... Extend the safety boundary and design a three-degree-of-freedom orthogonal projection system to generate multi-view depth maps.

[0025] Preferably, if the boundary length of the original minimum enclosing cube is L 0 The expanded security boundary is then... Then, the xoy plane, yoz plane, and xoz plane from the three views are selected to perform two-dimensional projections on the minimum bounding cube, resulting in three projected two-dimensional images. Considering that the aspect ratio of the depth map cannot be determined, the resolution of the depth map is set to [value missing]. n×n Establish three local coordinate systems corresponding to the projection directions, and the local coordinate systems are represented as follows: (2) in, d This represents the vertical distance from the 3D point cloud to the projected 2D plane, i.e., the projection distance.

[0026] 102: Calculate the pixel coordinates of the projected two-dimensional image based on the local coordinate system, and then calculate the gray value of the corresponding pixel based on the pixel coordinates.

[0027] Preferably, the pixel coordinates of the projected two-dimensional image are represented as follows: (3) in,( i,j) represents a pixel in a projected two-dimensional image, u i ,v i ) represents the th smallest bounding cube. i The local coordinates of each point in the projected two-dimensional image u max 、v max This represents the maximum value of the local coordinates. u min 、v min This represents the minimum value of the local coordinates. n×n This indicates the resolution of the projected two-dimensional image.

[0028] Any pixel in the projected two-dimensional image ( i,j The corresponding grayscale value Represented as: (4) 103: Construct a multi-objective optimization function to optimize the resolution of the projected 2D image and obtain the optimal resolution of the projected 2D image.

[0029] Preferably, the objective optimization function is expressed as: (5) in, This indicates the optimal resolution for projecting a two-dimensional image.

[0030] (6) This represents a quality evaluation function for a projected two-dimensional image. This indicates the noise level of the projected two-dimensional image. This represents the noise penalty coefficient, and λ ∈ (0~1).

[0031] An index representing the sharpness of a projected two-dimensional image. The Tenengrad gradient method is used to calculate the image gradient energy, and the sharpness is determined based on the image gradient energy.

[0032] 104: Perform edge detection on the projected two-dimensional image based on the grayscale value to obtain the edge pixel coordinates.

[0033] Preferably, the Canny operator is used to perform edge detection on the projected two-dimensional image, the image is smoothed by Gaussian filtering to reduce noise, the gradient magnitude and direction of the image are calculated using the first derivative of Gaussian, only local maxima in the gradient direction are retained, non-maximum points are suppressed, and edge points are filtered by high and low thresholds to ensure edge continuity.

[0034] 105: Normalize the edge pixel coordinates according to the optimal resolution.

[0035] Preferably, to eliminate the influence of image resolution, edge pixel coordinates are normalized, as follows: (7) in, Represents the edge pixel coordinates.

[0036] 106: Calculate the physical plane coordinates of the projected two-dimensional image based on the normalized edge pixel coordinates.

[0037] Preferably, the physical plane coordinates are represented as: (8) in, Represents physical plane coordinates.

[0038] 107: Perform validity checks on the corresponding depth values ​​based on the physical plane coordinates to obtain the final 3D point cloud feature points.

[0039] Preferably, the step of performing validity detection on the corresponding depth values ​​based on physical plane coordinates to obtain the final 3D point cloud feature points includes: 1071: Employing a predictive boundary detection mechanism, the effective mask function is defined as: (9) in, This represents the physical coordinate depth value.

[0040] 1072: The validity of the depth value corresponding to the physical plane coordinates is determined by an effective mask function. If valid, the depth value is calculated, and a linear mapping relationship is achieved from the pixel space to the three-dimensional coordinate system of the point cloud through the valid depth value, resulting in the final three-dimensional point cloud feature points, represented as: (10) The point cloud processing system for radioactive material transport containers described in this application includes a cube construction unit, a projection unit, a first calculation unit, a target optimization unit, an edge detection unit, a normalization unit, a second calculation unit, and an effectiveness detection unit.

[0041] The cube building block is used to construct the minimum bounding cube of a container's 3D point cloud.

[0042] The projection unit is used to extend the safety boundary of the minimum enclosing cube. Then, the xoy plane, yoz plane and xoz plane in the three views are selected to perform two-dimensional projection on the minimum enclosing cube to obtain three projected two-dimensional images. The local coordinate system corresponding to the minimum enclosing cube in the projected two-dimensional images is obtained.

[0043] The first calculation unit is used to calculate the pixel coordinates of the projected two-dimensional image according to the local coordinate system, and then calculate the gray value of the corresponding pixel according to the pixel coordinates.

[0044] The objective optimization unit is used to construct a multi-objective optimization function, which optimizes the resolution of the projected 2D image to obtain the optimal resolution of the projected 2D image.

[0045] The edge detection unit is used to perform edge detection on the projected two-dimensional image based on grayscale values ​​to obtain the edge pixel coordinates.

[0046] The normalization unit is used to normalize the edge pixel coordinates according to the optimal resolution.

[0047] The second calculation unit is used to calculate the physical plane coordinates of the projected two-dimensional image based on the normalized edge pixel coordinates.

[0048] The validity detection unit is used to perform validity detection on the physical coordinate depth value based on the physical plane coordinates to obtain the final 3D point cloud feature points.

[0049] In this embodiment, a laser 3D scanner is used to acquire point cloud data of the container surface sequentially through wiring, scanner calibration, and parameter settings, such as... Figure 1 As shown, the resolution is set to 2mm.

[0050] Construct the minimum bounding cube of the container's 3D point cloud according to equation (1), extend the safety boundary, and design a three-degree-of-freedom orthogonal projection system.

[0051] Select the xoy plane, yoz plane, and xoz plane from the three views and perform two-dimensional projection to generate a multi-view depth map (i.e., a projected two-dimensional image). Set the depth map resolution to [value missing]. n×n Based on equation (2), establish the corresponding local coordinate systems in the three projection directions. u,v,d ), calculate the pixel index value in the projected two-dimensional image according to equation (3) i,j ), and further calculate the pixel index value according to equation (4). i,j The corresponding image pixel grayscale values. The obtained two-dimensional projection image of the xoy plane is shown below. Figure 2 As shown.

[0052] Further, based on equation (6), define the quality evaluation function of the projected two-dimensional image. Clarity index The Tenengrad gradient method was used to calculate the image gradient energy using the Sobel operator. Furthermore, an adaptive multi-objective optimization function was established based on equation (5), and the optimal image resolution was finally calculated. n=150 .

[0053] The obtained three-view projected 2D image was subjected to edge detection using the Canny edge detection method. The detection results are as follows: Figure 3 As shown.

[0054] Based on the projection parameters, an initial feature point cloud is constructed by inverse mapping to three-dimensional space. The detected edge coordinates are normalized according to equation (7). The physical plane coordinates of the projected image are then calculated according to equation (8). An effective mask function is defined according to equation (9) to determine whether the depth value corresponding to the physical plane coordinates is valid; if valid, the corresponding depth value is calculated. Finally, the inverse mapping process from two-dimensional edge point features to three-dimensional key feature points is realized according to equation (10), and the result is as follows: Figure 4 As shown. The results indicate that the method described in this application can effectively simplify point cloud features while ensuring that key features are not lost.

[0055] The above are exemplary embodiments of this application, and the scope of protection of this application is defined by the claims and their equivalents.

Claims

1. A method for processing point clouds of radioactive material transport containers, characterized in that, include: Construct the minimum bounding cube of the container's 3D point cloud; The safety boundary of the minimum bounding cube is extended, and then the xoy plane, yoz plane and xoz plane in the three views are selected to perform two-dimensional projection on the minimum bounding cube to obtain three projected two-dimensional images. The local coordinate system of the minimum bounding cube in the projected two-dimensional images is obtained. The pixel coordinates of the projected two-dimensional image are calculated based on the local coordinate system, and then the gray value of the corresponding pixel is calculated based on the pixel coordinates. Construct a multi-objective optimization function to optimize the resolution of the projected 2D image and obtain the optimal resolution of the projected 2D image; Edge detection is performed on the projected 2D image based on grayscale values ​​to obtain the edge pixel coordinates; The edge pixel coordinates are normalized according to the optimal resolution; The physical plane coordinates of the projected two-dimensional image are calculated based on the normalized edge pixel coordinates; The validity of the corresponding depth values ​​is checked based on the physical plane coordinates to obtain the final 3D point cloud feature points.

2. The point cloud processing method for radioactive material transport containers as described in claim 1, characterized in that, a... The original point cloud coordinates are The coordinates of the eight vertices of the smallest enclosing cube are... x 1~ x 8 is represented as: ； Among them, subscript min and max These represent the minimum and maximum values ​​in the point cloud coordinates, respectively; x i ,y i ,z i ) ∈R 3 , R Represents the set of real numbers. i Describes the th smallest enclosing cube. i One point, ( x i ,y i ,z i ) represents the smallest enclosing cube. i The coordinates of the points o Indicates the origin. The expansion of the safety boundary of the minimum bounding cube includes: if the boundary length of the original minimum bounding cube is... L 0 The expanded security boundary is then... .

3. The point cloud processing method for radioactive material transport containers as described in claim 2, characterized in that, The local coordinate system is represented as follows: ； in, d This represents the perpendicular distance from the 3D point cloud to the projected 2D plane, i.e., the projection distance; The pixel coordinates of the projected two-dimensional image are represented as follows: ； in,( i,j ) represents a pixel in a projected two-dimensional image, u i ,v i ) represents the th smallest bounding cube. i The local coordinates of each point in the projected two-dimensional image u max 、v max This represents the maximum value of the local coordinates. u min 、v min This represents the minimum value of the local coordinates. n×n Indicates the resolution of the projected two-dimensional image; Any pixel in the projected two-dimensional image ( i,j The corresponding grayscale value Represented as: 。 4. The point cloud processing method for radioactive material transport containers as described in claim 3, characterized in that, The objective optimization function is expressed as: ； in, ; This indicates the optimal resolution of the projected two-dimensional image; This represents a quality evaluation function for a projected two-dimensional image. An index representing the sharpness of a projected two-dimensional image. This indicates the noise level of the projected two-dimensional image. This represents the noise penalty coefficient, and λ ∈ (0~1).

5. The point cloud processing method for radioactive material transport containers as described in claim 4, characterized in that, The normalization of edge pixel coordinates based on the optimal resolution is expressed as follows: ； in, Represents the edge pixel coordinates.

6. The point cloud processing method for radioactive material transport containers as described in claim 5, characterized in that, The physical plane coordinates are represented as follows: ； in, Represents physical plane coordinates.

7. The point cloud processing method for radioactive material transport containers as described in claim 6, characterized in that, The step of performing validity checks on the corresponding depth values ​​based on physical plane coordinates to obtain the final 3D point cloud feature points includes: The effective mask function is defined as follows: ； in, Represents the physical coordinate depth value; The validity of the depth value corresponding to the physical plane coordinates is determined by using an effective mask function. If valid, the depth value is calculated, and the inverse mapping process from 2D edge point features to 3D key feature points is achieved using the valid depth value, resulting in the final 3D point cloud feature points, represented as follows: 。 8. A point cloud processing system for a radioactive article transport container, the system being used in any one of the methods described in claims 1-7, characterized in that, The system includes: A cube building block is used to construct the minimum bounding cube of a container's 3D point cloud. The projection unit extends the safety boundary of the minimum bounding cube, and then selects the xoy plane, yoz plane and xoz plane in the three views to perform two-dimensional projection on the minimum bounding cube to obtain three projected two-dimensional images, and obtains the local coordinate system corresponding to the minimum bounding cube in the projected two-dimensional images; The first calculation unit calculates the pixel coordinates of the projected two-dimensional image based on the local coordinate system, and then calculates the gray value of the corresponding pixel based on the pixel coordinates. The objective optimization unit constructs a multi-objective optimization function, which optimizes the resolution of the projected 2D image to obtain the optimal resolution of the projected 2D image. The edge detection unit performs edge detection on the projected two-dimensional image based on grayscale values ​​to obtain the edge pixel coordinates. The normalization unit normalizes the edge pixel coordinates according to the optimal resolution; The second calculation unit calculates the physical plane coordinates of the projected two-dimensional image based on the normalized edge pixel coordinates. The validity detection unit performs validity detection on the physical coordinate depth value based on the physical plane coordinates to obtain the final 3D point cloud feature points.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the point cloud processing method for radioactive material transport containers as described in any one of claims 1 to 7.

10. A computer storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the point cloud processing method for radioactive material transport containers as described in any one of claims 1 to 7.

Images