A method for identifying special-shaped columns based on arc segment combination

Abstract

The present invention relates to the technical field of special-shaped column recognition, and in particular, to a special-shaped column recognition method based on a combination of arc segments. The main steps include: acquiring image data of the special-shaped column and extracting the initial outline through denoising and edge detection; employing a geometric feature segmentation method to identify arc segments and straight line segments using the Hough transform; constructing a geometric model using a graph neural network to form a topological structure of feature points; applying a graph matching algorithm to match the geometric model with a standard database, identifying the column type through curvature, length, and spatial distribution analysis; and optimizing the recognition results using a genetic algorithm and outputting the final special-shaped column recognition result. The present invention utilizes techniques such as image processing, geometric feature extraction, graph neural networks, and genetic algorithms to achieve efficient and accurate recognition of special-shaped columns, effectively improving the accuracy and efficiency of special-shaped column recognition.

Description

Technical Field

[0001] The present invention relates to the technical field of special-shaped column recognition, and in particular to a special-shaped column recognition method based on arc segment combination. Background Art

[0002] With the development of modern manufacturing technology, special-shaped column structures have been widely used in many fields such as architecture, aerospace, automobile manufacturing, and industrial design. Special-shaped columns usually have complex geometric shapes, and their cross-sections may present a variety of irregular shapes, such as arcs, combinations of multiple straight lines and arcs, and free curves. Therefore, the detection and recognition of special-shaped columns has become a key step in the precision manufacturing and quality control process. Currently, the following problems still exist: traditional recognition methods are mostly based on feature extraction of straight line segments and regular geometric figures. However, when faced with complex arc segment combination structures, existing technologies have difficulty in achieving high-precision detection and recognition, and lack accurate modeling of complex geometric forms; the shortcomings of existing technologies in shape segmentation and feature extraction lead to low recognition accuracy of special-shaped columns; in existing shape matching technologies, there is a lack of accurate matching methods for complex column features, resulting in large errors in the matching results of special-shaped columns, which can easily lead to recognition errors and misjudgments. Summary of the Invention

[0003] To solve the above problems, the present invention provides a method for identifying special-shaped columns based on a combination of arc segments, which solves the problem of how to achieve high-precision shape segmentation, feature extraction and accurate matching and recognition in special-shaped column structures composed of complex arc segments and multiple straight line segments, thereby improving the accuracy and efficiency of special-shaped column recognition and reducing errors.

[0004] To achieve the above object, the technical solution adopted by the present invention is:

[0005] A method for identifying special-shaped columns based on arc segment combination includes the following steps:

[0006] S1: Obtain image data of the special-shaped column through an image acquisition device, perform preprocessing operations, eliminate noise in the image through a denoising algorithm, and use an edge detection algorithm to obtain the initial contour number of the special-shaped column;

[0007] S2: Based on the initial contour data, the geometric feature segmentation method of arc segments and straight segments is used to decompose the external form of the special-shaped column into several combined line segments, and the arc segment and straight segment features are identified through the Hough transform algorithm;

[0008] S3: Based on the extracted arc and straight line segment features, a graph neural network algorithm is used to construct a geometric model of the special-shaped column. The topological structure diagram of the feature points is established based on the spatial structural relationship of the arc and straight line combination.

[0009] S4: performing shape matching on the geometric model with a known standard special-shaped column database using a graph matching-based shape matching algorithm, and identifying the column type corresponding to the special-shaped line segment combination through curvature matching, length comparison, and spatial distribution analysis;

[0010] S5: Based on the identified column type of the special-shaped column, a genetic algorithm is used to optimize the recognition result, and the final special-shaped column recognition result is output.

[0011] Furthermore, the image data includes the outer contour, surface texture, geometric shape information, color information and depth information of the special-shaped column.

[0012] Furthermore, the initial contour number includes multiple characteristic contour points of the outer contour of the special-shaped column extracted by the edge detection algorithm, and the distribution of the characteristic contour points is divided into several local contour areas according to the geometric characteristics of the special-shaped column, and each area contains one or more contour lines corresponding to arc segments or straight line segments.

[0013] Furthermore, step S2 includes the following steps:

[0014] Based on the initial contour data, key feature points on the contour are selected by analyzing the local curvature change rate, and these feature points are classified into high curvature points, low curvature points and inflection points;

[0015] Applying a dynamic window segmentation algorithm, the size of the segmentation window is adaptively adjusted according to the distribution of feature points, local curvature changes, and geometric shapes. The local curvature and direction of the contour line are analyzed to preliminarily identify candidate areas for arc segments and straight line segments.

[0016] In the candidate area, the weighted Hough transform algorithm is used to extract local line segment features, and the geometric features of arc segments and straight line segments are identified by voting for each contour point.

[0017] For the identified arc segments and straight line segments, the least squares method is applied to fit and correct their geometric parameters, including the center, radius and curvature parameters of the arc segments and the direction and length of the straight line segments, and finally the optimized arc segment and straight line segment feature parameters are output.

[0018] Furthermore, step S3 includes the following steps:

[0019] Constructing a geometric feature map of the special-shaped column based on the extracted arc segment and straight line segment features, wherein the geometric feature map includes position information, line segment type, curvature, angle, and length of the line segment;

[0020] Based on the geometric feature graph, an adaptive weighting strategy is adopted to encode the connection relationship between arc segments and straight line segments in a graph structure. The weight between each edge is dynamically adjusted according to spatial distance, direction change and curvature change to form an initial spatial topological structure.

[0021] Based on the constructed spatial topological structure, a graph neural network algorithm is used to construct the geometric model of the special-shaped column. By embedding and propagating node features and edge features, the graph neural network learns the spatial relationship and topological characteristics between different line segments, and gradually optimizes the geometric model of the special-shaped column. The final output geometric model contains the global morphological description of the special-shaped column and its local geometric details.

[0022] Furthermore, the geometric model of the special-shaped column is formulated as follows:

[0023]

[0024] Among them, M(υ) represents the geometric model of the special-shaped column, which contains the geometric information of local feature points and their adjacency relationships; h υ Represents the geometric feature representation of node v, which is used to describe the local geometric information of a point on the cylinder; Indicates other feature points on the special-shaped column that are directly connected to the feature point v; c υu and c υw represents the spatial weight between adjacent points; α, β, and γ represent the adjustment parameters in the model, which are used to control the influence weight of different geometric features on the overall model.

[0025] Furthermore, step S4 includes the following steps:

[0026] Based on the geometric model, the graph matching algorithm is used to match the topological structure of the geometric model with the model in the standard special-shaped column database, and the comparison is carried out through the position of feature points, line segment curvature and spatial distribution characteristics;

[0027] In the shape matching process, a multi-scale feature extraction algorithm is used to analyze the different scale information of arc segments and straight line segments in the geometric model;

[0028] The curvature matching algorithm is used to calculate the curvature difference between the matching model and the database model, and the curvature distribution and curve change trend of the arc segment are compared;

[0029] Based on the line segment length comparison algorithm, the length difference between the corresponding arc and straight line segments in the identified geometric model and the standard model is analyzed to compensate and adjust the length error;

[0030] Through spatial distribution analysis, the relative position relationship between the arc segments and straight segments in the geometric model is accurately compared in space. Combined with the changes in the relative spatial positions, the global morphological matching degree of the model is judged, and the recognition result of the special-shaped column is finally determined.

[0031] Furthermore, the graph matching algorithm is formulated as follows:

[0032]

[0033] Among them, E match Indicates the matching degree between the geometric model obtained by the graph matching algorithm and the standard special-shaped column database model; υ i and υ′ i They represent the feature points in the identified special-shaped column geometric model and the corresponding feature points in the standard special-shaped column database; d(υ i ,υ′ i ) represents the characteristic point υ in the geometric model i and the feature point υ′ in the database model i The spatial position difference between i ,υ′ i ) represents the feature point υ i and υ′ i The curvature difference of the arc segment where it is located; Δl(υ i ,υ′ i ) represents the length difference between the corresponding arc or straight line segments in the geometric model and the database model; α1, α2, and α3 represent the weight parameters of position difference, curvature difference, and length difference, respectively; (υ i ,υ j ) represents a feature point pair.

[0034] Furthermore, when the genetic algorithm is used to optimize the recognition result, the fitness function of the column type matching result is calculated to screen out the optimal column matching result.

[0035] The present invention has the following beneficial effects: It uses an image acquisition device to acquire image data of irregularly shaped columns and employs a denoising algorithm to eliminate noise, thereby improving image quality, reducing errors caused by image noise, and making subsequent contour extraction more accurate. Based on edge detection and the Hough transform algorithm, it can effectively identify geometric features such as arc and straight line segments of irregularly shaped columns, enabling the complex irregularly shaped column profile to be decomposed into recognizable basic shape combinations, facilitating subsequent processing. By constructing a geometric model using a graph neural network algorithm, it can fully utilize the spatial structural information of arc and straight line combinations to create a more accurate topological structure graph, enabling precise representation of the shape and structural information of irregularly shaped columns and enhancing the accuracy of shape matching. Based on a graph matching algorithm, the extracted geometric model is matched against a standard database for shape matching. Through methods such as curvature matching, length comparison, and spatial distribution analysis, the corresponding column type can be efficiently and accurately identified, thereby improving recognition accuracy. A genetic algorithm is used to optimize the initial recognition results, enabling a wide range of searches for optimal solutions, further improving the accuracy and reliability of the recognition results and ensuring a high degree of accuracy in the final output of irregularly shaped column recognition results. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a flow chart of a method for identifying special-shaped columns based on arc segment combination according to the present invention.

[0037] Figure 2 It is a flowchart of step S3 provided by one embodiment of the present invention.

[0038] Figure 3 It is a flowchart of step S4 provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0039] See also Figure 1-3 As shown, the present invention relates to a method for identifying special-shaped columns based on a combination of arc segments.

[0040] Example

[0041] A method for identifying special-shaped columns based on arc segment combination includes the following steps:

[0042] S1: Obtain image data of the special-shaped column through an image acquisition device, perform preprocessing operations, eliminate noise in the image through a denoising algorithm, and use an edge detection algorithm to obtain the initial contour number of the special-shaped column;

[0043] The image data includes the outer contour, surface texture, geometric shape information, color information and depth information of the special-shaped column; the initial contour data includes multiple characteristic contour points of the outer contour of the special-shaped column extracted by the edge detection algorithm, and the distribution of the characteristic contour points is divided into several local contour areas according to the geometric characteristics of the special-shaped column, and each area contains one or more contour lines corresponding to arc segments or straight line segments.

[0044] It's important to note that image acquisition equipment must possess high resolution, a wide viewing angle, and the ability to capture depth information. Commonly selected devices include high-resolution RGB cameras and RGB-D depth cameras, the latter of which can simultaneously capture both color and depth information. To fully capture the contours and surface features of irregularly shaped columns, the equipment can capture images from multiple angles around the column. This can be achieved by setting up multiple fixed cameras or using a robotic arm for omnidirectional scanning, ensuring that the column's surface features are captured from all angles.

[0045] The captured image data includes color information, depth information, and the surface texture of the irregularly shaped columns. Depth information is particularly important for accurately identifying the 3D geometric structure of irregularly shaped columns. The RGB-D camera data simultaneously outputs two matrices: one representing the column's color (RGB) and the other representing its depth information (D), which is the distance from the camera to each point on the column's surface. If the device uses a laser scanner, point cloud data of the column's surface can be directly acquired. Point cloud data contains higher-precision 3D spatial information, facilitating subsequent contour extraction and shape matching.

[0046] Specifically, the core of edge detection is to extract the contours of an object through grayscale changes in an image. Canny edge detection calculates the gradient information of an image and extracts the parts of the image with large intensity changes as edges. The specific steps include: converting the image into a grayscale image to simplify the calculation; using the Sobel operator or other derivative operators to calculate the gradient of each pixel in the image in the x and y directions. This gradient represents the grayscale change rate of the pixel; to make the edge more refined, the Canny algorithm will eliminate non-maximum points and retain the pixels with the largest local gradient to ensure that the final edge is continuous and precise; to reduce the impact of noise, a dual threshold screening algorithm is used to retain the parts with higher gradient values ​​as strong edges and suppress the parts with lower gradient values.

[0047] In the image after edge detection, a large number of contour segments are usually extracted. However, these segments cannot be used directly to identify special-shaped columns, so it is necessary to further extract representative feature contour points. According to the geometric characteristics of the special-shaped column, the contour line is divided into several local areas. The contour of each local area may be a straight line segment or an arc segment. During segmentation, these local contour areas can be calibrated according to the curvature change points in the image (i.e., turning points or intersections on the edge). Each feature point is divided according to the local area to preliminarily form the overall contour of the special-shaped column. Based on the geometric shape of the image, the distribution area of ​​the edge feature points is divided into arc segments and straight line segments. At this time, the least squares method can be used to fit the local edge and determine the curvature of the curve to segment the straight line and arc areas. The arc area usually shows a continuous curvature change, while the straight line segment corresponds to the part with curvature close to zero.

[0048] S2: Based on the initial contour data, the geometric feature segmentation method of arc segments and straight segments is used to decompose the external form of the special-shaped column into several combined line segments, and the arc segment and straight segment features are identified through the Hough transform algorithm;

[0049] Wherein, the step S2 includes the following steps:

[0050] Based on the initial contour data, key feature points on the contour are selected by analyzing the local curvature change rate, and these feature points are classified into high curvature points, low curvature points and inflection points;

[0051] It should be noted that the feature points are classified into high curvature points, low curvature points and inflection points, as follows:

[0052] High curvature points: These points have large curvature values ​​and are typically located at locations with significant bends, such as the start and end points of arc segments or certain corners. High curvature points indicate significant changes in the morphology of a local area.

[0053] Low curvature points: The curvature value is close to zero. These points are usually located on straight line segments, indicating that there is almost no bending in the local area.

[0054] Inflection point: A transition point between two different shapes (such as an arc segment and a straight line segment). The inflection point has a more dramatic change in curvature, marking a clear transition in geometric shape.

[0055] Applying a dynamic window segmentation algorithm, the size of the segmentation window is adaptively adjusted according to the distribution of feature points, local curvature changes, and geometric shapes. The local curvature and direction of the contour line are analyzed to preliminarily identify candidate areas for arc segments and straight line segments.

[0056] Specifically, after analyzing local curvature features, a dynamic window segmentation algorithm is used to perform preliminary segmentation of the contour lines. The core idea of ​​this algorithm is to dynamically adjust the size of the segmentation window based on the curvature change rate and geometric features to adapt to different geometric forms. The window size is adjusted based on the degree of curvature change. For areas with large curvature (i.e., arc-shaped areas), the window size is reduced to refine the segmentation; for areas with small curvature (i.e., straight-line areas), the window size can be appropriately increased. For each window area, candidate areas for arc segments and straight-line segments are identified based on curvature and direction analysis. Arc regions are characterized by stable and non-zero curvature, while straight-line segments are characterized by near-zero curvature and consistent direction. The overall direction of the area is determined by calculating the angle between contour points within each window. If the direction change is small, the area is a straight-line segment; if the direction change is large and the curvature is stable, the area is an arc segment.

[0057] In the candidate area, the weighted Hough transform algorithm is used to extract local line segment features, and the geometric features of arc segments and straight line segments are identified by voting for each contour point.

[0058] Specifically, the Hough Circle Transform is applied to arc-shaped candidate regions. Each contour point votes for a possible center and radius based on its position within the arc. The center and radius with the highest number of votes are the recognized arc segment features. In the weighted Hough Transform, each contour point is assigned a different weight. The weights are determined based on the curvature and the point's position. Points with higher curvature or closer to the center are assigned higher weights, thereby improving recognition accuracy.

[0059] For line segment candidates, the standard Hough transform is used. Each contour point votes for possible line parameters (slope and intercept) based on its gradient direction. The parameter combination with the highest number of votes is the identified line segment. Similarly, a weighting mechanism is applied, assigning higher weights to points close to the center of the line with more consistent gradient directions, ensuring more accurate line segment identification.

[0060] For the identified arc segments and straight line segments, the least squares method is applied to fit and correct their geometric parameters, including the center, radius and curvature parameters of the arc segments and the direction and length of the straight line segments, and finally the optimized arc segment and straight line segment feature parameters are output.

[0061] S3: Based on the extracted arc and straight line segment features, a graph neural network algorithm is used to construct a geometric model of the special-shaped column. The topological structure diagram of the feature points is established based on the spatial structural relationship of the arc and straight line combination.

[0062] Wherein, the step S3 includes the following steps:

[0063] Constructing a geometric feature map of the special-shaped column based on the extracted arc segment and straight line segment features, wherein the geometric feature map includes position information, line segment type, curvature, angle, and length of the line segment;

[0064] Specifically, the geometric feature graph is a graph structure consisting of nodes and edges, where each node represents a key feature of the special-shaped column profile (such as an arc segment or a straight line segment), and each edge represents the spatial relationship between these features. Each node contains the following geometric feature information:

[0065] Segment type: Identifies whether the node is an arc segment or a straight line segment.

[0066] Position information: including the coordinates of the starting point and end point of the line segment.

[0067] Segment curvature: For arc segments, the reciprocal of the radius is used to represent the curvature; for straight line segments, the curvature is zero.

[0068] Segment angle: For a straight line segment, the direction is expressed by the slope or angle of the segment; for an arc segment, the angle is expressed by the change in the direction of the tangent of the segment.

[0069] Segment length: The physical length of the segment, calculated directly from the start and end point coordinates.

[0070] Based on the geometric feature graph, an adaptive weighting strategy is adopted to encode the connection relationship between arc segments and straight line segments in a graph structure. The weight between each edge is dynamically adjusted according to spatial distance, direction change and curvature change to form an initial spatial topological structure.

[0071] Specifically, after constructing the geometric feature graph, the spatial relationships between different line segments (nodes) need to be connected using edges. To describe this connection, an adaptive weighting strategy can be used to dynamically adjust the weight of each edge based on the spatial distance, direction changes, and curvature changes of the line segments.

[0072] Spatial Distance: Calculates the spatial distance between each pair of adjacent line segments. The distance between two line segments can be calculated using the Euclidean distance between their endpoint coordinates. Closer line segments should have a larger edge weight, indicating a closer spatial connection between them. A threshold can be set to ignore line segments whose distance exceeds a certain range.

[0073] Direction Change: For straight segments, directional differences are determined by analyzing their angle changes. Similar directions (i.e., segments with a small angle between them) indicate that their geometry is spatially consistent and are therefore weighted higher; segments with larger angles receive lower weights. For curved segments, directional changes are analyzed by the direction of their tangents. If the difference in the tangent directions of two curved segments is small, this indicates that they are closer in the overall surface structure and is therefore weighted higher.

[0074] Curvature variation: For curved segments, curvature variation is also an important characteristic. Segments with smaller curvature differences should be given higher weights in the topology, as they are likely to belong to the same surface structure. For connections between straight and curved segments, the weight is calculated based on the transition curvature between the straight and curved segments, with connections with smoother curvature transitions being given higher weights.

[0075] Based on this adaptive weighting strategy, the weight of each edge is calculated, ultimately forming a complete spatial topology of the geometric feature graph. The nodes in this topology represent line segments, the edges represent the spatial relationships between these line segments, and the edge weights reflect the geometric similarity or spatial proximity between them.

[0076] Based on the constructed spatial topological structure, a graph neural network algorithm is used to construct the geometric model of the special-shaped column. By embedding and propagating node features and edge features, the graph neural network learns the spatial relationship and topological characteristics between different line segments, and gradually optimizes the geometric model of the special-shaped column. The final output geometric model contains the global morphological description of the special-shaped column and its local geometric details.

[0077] Specifically, the spatial topology constructed in the previous step is used as the input of the graph neural network. The geometric features of each node (such as position, curvature, direction, etc.) are used as input node features, while the weight of the edge (based on distance, direction, and curvature changes) is used as the edge feature in the graph structure.

[0078] The graph neural network first embeds the geometric features of each node, converting its raw geometric information into a high-dimensional feature vector. Within each layer of the graph neural network, node features propagate between adjacent nodes via edges. Each node transmits its own features to its neighboring nodes through edges (i.e., connections), while also receiving features from them. Edge weights adjust the strength of information propagation. Higher-weight edges strengthen feature propagation between two nodes, thereby better learning the close relationships between these nodes. As the number of graph neural network layers increases, each node gradually receives more information from its neighbors, gradually establishing a connection between local geometric features and the overall structure. For the geometric model of a special-shaped column, the local geometric features of a node (such as the length and curvature of an arc or line segment) are propagated through the graph neural network, gradually establishing a connection with the global shape of the entire special-shaped column. Through multi-layer propagation and feature updates, the graph neural network continuously optimizes the spatial topology of the special-shaped column. The network adjusts the features of each node by learning the relationships between different nodes (line segments), gradually constructing a complete geometric model. For example, for complex special-shaped columns, the network will learn which arc segments belong to the same continuous surface and which straight line segments constitute the main framework of the column, thereby optimizing the accuracy of the overall model.

[0079] Furthermore, the geometric model of the special-shaped column is formulated as follows:

[0080]

[0081] Among them, M(υ) represents the geometric model of the special-shaped column, which contains the geometric information of local feature points and their adjacency relationships; h υ Represents the geometric feature representation of node v, which is used to describe the local geometric information of a point on the cylinder; Indicates other feature points on the special-shaped column that are directly connected to the feature point v; c υu and c uw represents the spatial weight between adjacent points, c υu Represents the spatial geometric relationship between the feature point v and its neighbor feature point u, such as the distance, curvature difference, direction change, etc. uw It represents the relationship between the feature point u and its neighbor w, further capturing the relationship between remote feature points in the spatial topology structure and helping to optimize the global description of the geometric model; α, β, and γ represent the adjustment parameters in the model, which are used to control the influence weight of different geometric features on the overall model.

[0082] S4: performing shape matching on the geometric model with a known standard special-shaped column database using a graph matching-based shape matching algorithm, and identifying the column type corresponding to the special-shaped line segment combination through curvature matching, length comparison, and spatial distribution analysis;

[0083] Wherein, the step S4 includes the following steps:

[0084] Based on the geometric model, the graph matching algorithm is used to match the topological structure of the geometric model with the model in the standard special-shaped column database, and the comparison is carried out through the position of feature points, line segment curvature and spatial distribution characteristics;

[0085] Specifically, the shape matching of the special-shaped column geometric model constructed in step S3 is performed with the model in the standard special-shaped column database. This step adopts a graph matching algorithm to compare the topological structure diagram in the geometric model with the topological structure diagram of the standard column model in the database. The specific steps include: the geometric model of each special-shaped column contains nodes (arcs and straight line segments) and edges (representing the connection relationship between arcs and straight line segments), which constitute the topological graph of the geometric model. The attributes on the nodes include position information, curvature, angle, length, etc. Using graph isomorphism or subgraph isomorphism algorithm, the feature points and line segments in the geometric model are matched one by one with the features in the standard database. Through the graph matching algorithm, the possible candidate models of the target special-shaped column in the standard database can be found. The standard special-shaped column database contains various types of special-shaped column models to ensure that the matching result that best meets the target geometric model can be found.

[0086] In the shape matching process, a multi-scale feature extraction algorithm is used to analyze the different scale information of arc segments and straight line segments in the geometric model;

[0087] The curvature matching algorithm is used to calculate the curvature difference between the matching model and the database model, and the curvature distribution and curve change trend of the arc segment are compared;

[0088] Specifically, the curvature matching algorithm is used to analyze the curvature differences between the models. This step focuses on the curvature distribution of the arc segments and the curve change trend, as follows:

[0089] Curvature distribution analysis: For each arc segment in the geometric model, we first extract its curvature distribution, including the curvature values ​​at the starting and ending points, as well as the change in curvature at the intermediate points. This is then compared with the corresponding arc segment in the standard database, and the curvature difference is calculated.

[0090] Comparison of curve trends: Instead of just comparing curvature values ​​point by point, analyze the overall trend of the curve across the entire arc segment. For example, check whether there is a consistent increasing or decreasing curvature trend, or whether the curvature of a particular segment exhibits a similar fluctuation pattern.

[0091] Through curvature matching, the standard arc segment closest to the target special-shaped column can be accurately found to ensure that the matching arc segment has similar bending characteristics to the standard model.

[0092] Based on the line segment length comparison algorithm, the length difference between the corresponding arc and straight line segments in the identified geometric model and the standard model is analyzed to compensate and adjust the length error;

[0093] Specifically, a line segment length comparison algorithm is used to compare the lengths of arc segments and straight line segments. The specific operation is as follows:

[0094] Arc Segment Length Comparison: The length of an arc segment is determined by both curvature and radian. First, calculate the arc length based on the curvature of the arc segment in the geometric model and the angles between its endpoints. Then compare the arc length with the arc segment in the standard model. If the lengths differ significantly, the arc length may need to be adjusted to match.

[0095] Line Segment Length Comparison: For straight line segments, the length difference between the corresponding straight line segments in the geometric model and the standard model is directly calculated. If there is a large difference in length, the system will automatically correct the length or mark the segment as a possible matching error.

[0096] Length Error Compensation: If the length difference is within a certain range, the system will make subtle adjustments to the length of the line segments in the geometric model based on the error compensation strategy to make it more consistent with the standard model. The compensation process maintains the integrity of curvature and angles, ensuring that the overall structure of the model is not affected.

[0097] Through spatial distribution analysis, the relative position relationship between the arc segments and straight segments in the geometric model is accurately compared in space. Combined with the changes in the relative spatial positions, the global morphological matching degree of the model is judged, and the recognition result of the special-shaped column is finally determined.

[0098] Specifically, by analyzing the spatial distribution of arc segments and straight line segments in the geometric model, their relative positions in three-dimensional space are evaluated, and the global matching degree of the model is further determined. The specific operations are as follows:

[0099] Spatial Position Comparison: For each arc and line segment, the system calculates its relative position in 3D space based on its starting point, endpoint, and direction vector. This is then compared with the corresponding line segment in the standard model to check their relative distance and positional relationship.

[0100] Relative Position Change Analysis: This analyzes the spatial distribution differences between arcs and straight lines. For example, if the distance or angle between two segments conforms to a standard model, the match is high. If the spatial variation between adjacent segments exceeds a preset threshold, the global morphological match is reduced.

[0101] Global morphological matching calculation: The system calculates the global matching degree by comprehensively evaluating the spatial distribution of all line segments. The higher the matching degree, the closer the target geometry model is to the standard model in terms of overall morphology.

[0102] Furthermore, the graph matching algorithm is formulated as follows:

[0103]

[0104] Among them, E match Indicates the matching degree between the geometric model obtained by the graph matching algorithm and the standard special-shaped column database model; υ i and υ′ i They represent the feature points in the identified special-shaped column geometric model and the corresponding feature points in the standard special-shaped column database; d(υ i ,υ′ i ) represents the characteristic point υ in the geometric model i and the feature point υ′ in the database model i The spatial position difference between i ,υ′ i ) represents the feature point υ i and υ′ i The curvature difference of the arc segment where it is located; Δl(υ i ,υ′ i ) represents the length difference between the corresponding arc or straight line segments in the geometric model and the database model; α1, α2, and α3 represent the weight parameters of position difference, curvature difference, and length difference, respectively; (υ i ,υ j ) represents a feature point pair.

[0105] S5: Based on the identified column type of the special-shaped column, a genetic algorithm is used to optimize the recognition result, and the final special-shaped column recognition result is output; when the genetic algorithm is used to optimize the recognition result, the fitness function of the column type matching result is calculated to screen out the optimal column matching result.

[0106] Specifically, in the previous steps, a shape matching algorithm was used to preliminarily generate multiple candidate column type matching results. These candidate matching results serve as the initial population, and each candidate result can be considered an "individual," representing a matching solution for a special-shaped column. For each individual, a binary or real number encoding method is used to represent the combination of column features. For example, the characteristics of a column may include the curvature, length, and relative position of arc and straight line segments. The diversity of the initial population is very important, ensuring that different individuals have different combinations of geometric features, thereby increasing the exploration capability during the optimization process.

[0107] The fitness function is a criterion for evaluating the quality of each individual (matching result). In this method, the definition of the fitness function is closely related to the degree of matching of the geometric features of the special-shaped columns, and mainly considers the following aspects:

[0108] Curvature Matching: Calculates the difference in curvature between the target column and the arc segments of candidate columns in the database. The smaller the curvature difference, the more consistent the arc features of the two, and the higher the fitness value. Normalizing the curvature ensures that the curvature matching of each individual is comparable.

[0109] Length Matching: Compares the lengths of recognized arc and straight line segments to the lengths of known standard cylinder segments. The smaller the length difference, the higher the fitness value. This can be calculated using the relative error of the length difference.

[0110] Spatial Distribution Similarity: This evaluates whether the relative positions and distributions of arc and straight line segments in 3D space are consistent with known cylinders. Spatial distribution similarity is quantified by calculating the relative distances and angles between feature points.

[0111] Overall shape fit: Calculates the overall contour fit based on the combined relationships of all arc segments and straight line segments. The overall contour of the target column is superimposed and compared with the overall contour of the standard model to quantify the shape similarity between the two.

[0112] Through selection, crossover, and mutation operations, the genetic algorithm generates multiple candidate matching results. Each candidate is evaluated using a fitness function, gradually selecting the optimal match. As the number of iterations increases, the genetic algorithm ultimately outputs the optimal column matching result. The optimal identification result contains the complete geometric characteristics of the special-shaped column, including information such as the curvature, length, and spatial distribution of its curved and straight segments, and matches it to a specific column type in the standard special-shaped column database.

[0113] The final recognition results are output, including the following: the type of special-shaped columns identified; the matching degree between the recognition results and the standard model (such as the matching score of curvature, length, and topological structure); error analysis (such as spatial distribution error, matching geometric error); optimization steps of the recognition process and final convergence information.

[0114] In summary, the present invention utilizes high-resolution image acquisition equipment combined with denoising and edge detection algorithms to accurately capture the contour information of irregularly shaped columns. In particular, the inclusion of depth information allows for a comprehensive display of the three-dimensional geometric structure, improving recognition accuracy. By analyzing local curvature and dynamically adjusting the segmentation window, it effectively distinguishes between straight and curved segments, enabling more accurate local segmentation of complex geometric forms. Furthermore, the weighted Hough transform algorithm further precisely identifies these characteristic segments, ensuring the accuracy and reliability of the recognition results.

[0115] This paper employs a graph neural network to construct a geometric model of special-shaped columns through feature propagation and learning between nodes and edges. This model comprehensively describes the global morphology and local geometric details of special-shaped columns, providing a solid foundation for subsequent shape matching. Through curvature matching, length comparison, and spatial distribution analysis, the system is able to perform high-precision shape matching of special-shaped columns, ensuring the final identification of special-shaped columns is consistent with the standard column model in the database. Based on the matching results, a genetic algorithm is introduced for further optimization. Through multiple rounds of iteration, the optimal column matching result is screened, improving the accuracy of the recognition results and outputting detailed geometric parameters and error analysis information.

[0116] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary engineering technicians in this field should fall within the scope of protection determined by the claims of the present invention.

Claims

1. A method for identifying special-shaped columns based on arc segment combination, characterized in that: The following steps are involved: S1: Obtain image data of the special-shaped column through an image acquisition device, perform preprocessing operations, eliminate noise in the image through a denoising algorithm, and use an edge detection algorithm to obtain the initial contour number of the special-shaped column; S2: Based on the initial contour data, the geometric feature segmentation method of arc segments and straight segments is used to decompose the external form of the special-shaped column into several combined line segments, and the arc segment and straight segment features are identified through the Hough transform algorithm; S3: Based on the extracted arc and straight line segment features, a graph neural network algorithm is used to construct a geometric model of the special-shaped column. The topological structure diagram of the feature points is established based on the spatial structural relationship of the arc and straight line combination. S4: performing shape matching on the geometric model with a known standard special-shaped column database using a graph matching-based shape matching algorithm, and identifying the column type corresponding to the special-shaped line segment combination through curvature matching, length comparison, and spatial distribution analysis; S5: Based on the identified column type of the special-shaped column, a genetic algorithm is used to optimize the recognition result, and the final special-shaped column recognition result is output; Step S3 includes the following steps: Constructing a geometric feature map of the special-shaped column based on the extracted arc segment and straight line segment features, wherein the geometric feature map includes position information, line segment type, curvature, angle, and length of the line segment; Based on the geometric feature graph, an adaptive weighting strategy is adopted to encode the connection relationship between arc segments and straight line segments in a graph structure. The weight between each edge is dynamically adjusted according to spatial distance, direction change and curvature change to form an initial spatial topological structure. Based on the constructed spatial topology, a graph neural network algorithm is used to construct a geometric model of the special-shaped column. By embedding and propagating node and edge features, the graph neural network learns the spatial relationships and topological properties between different line segments, gradually optimizing the geometric model of the special-shaped column. The final output geometric model includes a global morphological description of the special-shaped column and its local geometric details. The formula of the geometric model of the special-shaped column is as follows: ； in, Represents the geometric model of special-shaped columns, including the geometric information of local feature points and their adjacency relationships; The geometric feature representation of the feature point v is used to describe the local geometric information of a point on the cylinder; Indicates other feature points on the special-shaped column that are directly connected to the feature point v; and Represents the spatial weight between adjacent points; 、 、 Represents the adjustment parameters in the model, which are used to control the influence weight of different geometric features on the overall model.

2. The method for identifying special-shaped columns based on arc segment combination according to claim 1, characterized in that: The image data includes the outer contour, surface texture, geometric shape information, color information and depth information of the special-shaped column.

3. The method for identifying special-shaped columns based on arc segment combination according to claim 1, characterized in that: The initial contour count includes multiple characteristic contour points of the outer contour of the special-shaped column extracted by an edge detection algorithm. The distribution of the characteristic contour points is divided into several local contour areas according to the geometric characteristics of the special-shaped column. Each area contains one or more contour lines corresponding to arc segments or straight line segments.

4. The method for identifying special-shaped columns based on arc segment combination according to claim 1, characterized in that: Step S2 includes the following steps: Based on the initial contour data, key feature points on the contour are selected by analyzing the local curvature change rate, and these feature points are classified into high curvature points, low curvature points and inflection points; Applying a dynamic window segmentation algorithm, the size of the segmentation window is adaptively adjusted according to the distribution of feature points, local curvature changes, and geometric shapes. The local curvature and direction of the contour line are analyzed to preliminarily identify candidate areas for arc segments and straight line segments. In the candidate area, the weighted Hough transform algorithm is used to extract local line segment features, and the geometric features of arc segments and straight line segments are identified by voting for each contour point. For the identified arc segments and straight line segments, the least squares method is applied to fit and correct their geometric parameters, including the center, radius and curvature parameters of the arc segments and the direction and length of the straight line segments, and finally the optimized arc segment and straight line segment feature parameters are output.

5. The method for identifying special-shaped columns based on arc segment combination according to claim 1, characterized in that: The S4 comprises the following steps: Based on the geometric model, the graph matching algorithm is used to match the topological structure of the geometric model with the model in the standard special-shaped column database, and the comparison is carried out through the position of feature points, line segment curvature and spatial distribution characteristics; In the shape matching process, a multi-scale feature extraction algorithm is used to analyze the different scale information of arc segments and straight line segments in the geometric model; The curvature matching algorithm is used to calculate the curvature difference between the matching model and the database model, and the curvature distribution and curve change trend of the arc segment are compared; Based on the line segment length comparison algorithm, the length difference between the corresponding arc and straight line segments in the identified geometric model and the standard model is analyzed to compensate and adjust the length error; Through spatial distribution analysis, the relative position relationship between the arc segments and straight segments in the geometric model is accurately compared in space. Combined with the changes in the relative spatial positions, the global morphological matching degree of the model is judged, and the recognition result of the special-shaped column is finally determined.

6. The method for identifying special-shaped columns based on arc segment combination according to claim 5, characterized in that: The formula of the graph matching algorithm is as follows: ； in, Indicates the matching degree between the geometric model obtained by the graph matching algorithm and the standard special-shaped column database model; and They represent the feature points in the identified special-shaped column geometric model and the corresponding feature points in the standard special-shaped column database respectively; Represents feature points in geometric models and feature points in the database model The spatial position difference between Representing feature points and The curvature difference of the arc segment; Indicates the difference in length between the corresponding arc or straight line segments in the geometric model and the database model; 、 and The weight parameters representing position difference, curvature difference and length difference respectively; Represents a pair of feature points.

7. The method for identifying special-shaped columns based on arc segment combination according to claim 1, characterized in that: When the genetic algorithm is used to optimize the recognition result, the fitness function of the column type matching result is calculated to screen out the optimal column matching result.

Images