Cutting method and device for steel coil type scrap steel, electronic equipment and storage medium

Abstract

The invention provides a steel coil type scrap steel cutting method and device, electronic equipment and a storage medium. The method comprises the steps that point cloud data of a to-be-cut steel coil in a horizontal posture is obtained; translating and rotating the point cloud data of the steel coil to obtain transformed point cloud data of the steel coil; according to the transformed point cloud data of the steel coil, a cutting path is generated, the cutting path comprises cutting lines located on the two sides of the steel coil, and the cutting line on at least one side is divided into two times to complete cutting; and the steel coil is cut according to the cutting path. Through transformation of the point cloud data of the steel coil, the subsequent path planning calculation process is simpler; the whole steel coil type scrap steel is cylindrical, the steel coil type scrap steel can deform due to internal stress after being cut, the shape of the steel coil can be changed, the shape change of the steel coil after first-time cutting can be avoided by introducing multi-time cutting, and cutting of the steel coil type scrap steel is more reasonable.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent steel manufacturing, and particularly relates to a steel coil type scrap steel cutting method and device, electronic equipment and storage medium. BACKGROUND

[0002] In the steel manufacturing process, scrap steel cutting is one of the key links. The traditional scrap steel cutting method mainly relies on manual operation, among which flame cutting and plasma cutting processes are widely used. However, these methods have many problems, including high labor intensity, serious environmental pollution, and cutting precision being greatly affected by human factors. In addition, scrap steel has a complex shape, and irregular scrap steel is difficult to cut accurately, which increases material waste and energy consumption.

[0003] With the rapid development of industrial automation technology, intelligent and unmanned cutting has become an important research direction in the steel industry. In recent years, advanced technologies such as point cloud processing, computer vision and deep learning have been widely applied in the field of industrial manufacturing, providing the possibility for the intelligentization of scrap steel cutting. However, existing intelligent cutting methods mainly focus on regular-shaped metal plates, and there are still great challenges in size estimation and accurate cutting of steel coil type scrap steel. Therefore, there is an urgent need for an efficient and accurate scrap steel cutting method to improve cutting precision, reduce material waste, and optimize cutting efficiency. SUMMARY

[0004] In view of the defects of the prior art, the present application provides a steel coil type scrap steel cutting method and device, electronic equipment and storage medium to solve the technical problems of low cutting precision and low efficiency of steel coil type scrap steel in the prior art.

[0005] To achieve the above object and other related objects, the present application provides a steel coil type scrap steel cutting method, comprising: obtaining point cloud data of a steel coil in a horizontal posture to be cut; translating and rotating the point cloud data of the steel coil to obtain transformed point cloud data of the steel coil; generating a cutting path according to the transformed point cloud data of the steel coil, the cutting path including cutting lines located on both sides of the steel coil, and the cutting lines on at least one side being divided into two times for cutting; and cutting the steel coil according to the cutting path.

[0006] In an embodiment of the present application, obtaining point cloud data of a steel coil in a horizontal posture to be cut comprises: obtaining initial point cloud data of a plurality of local regions of the scrap steel to be cut, the initial point cloud data of adjacent local regions having overlapping parts; performing splicing processing on the initial point cloud data of the plurality of local regions of the scrap steel to be cut to obtain spliced point cloud data of the scrap steel to be cut; and performing ground point cloud elimination, noise filtering and clustering segmentation processing on the spliced point cloud data of the scrap steel to be cut to obtain point cloud data of each steel coil.

[0007] In an embodiment of the present application, the initial point cloud data of the plurality of local regions of the scrap steel to be cut is spliced to obtain spliced point cloud data of the scrap steel to be processed, including: performing down-sampling processing on each initial point cloud data to obtain a plurality of down-sampled point cloud data; extracting feature points with topological significance and local geometric significance from each down-sampled point cloud data; matching the feature points of adjacent down-sampled point cloud data using a feature point matching algorithm to obtain a feature point mapping relationship; and adjusting the pose of adjacent initial point cloud data or down-sampled point cloud data according to the feature point mapping relationship to obtain the spliced point cloud data of the scrap steel.

[0008] In an embodiment of the present application, the spliced point cloud data of the scrap steel to be cut is subjected to ground point cloud elimination, noise filtering and clustering segmentation processing to obtain point cloud data of each steel coil, including: fitting a ground plane equation for the spliced point cloud data of the scrap steel to be cut using a random sample consensus algorithm; eliminating ground point cloud data in the spliced point cloud data of the scrap steel to be cut according to the ground plane equation to obtain first point cloud data; removing noise in the first point cloud data using a filtering algorithm to obtain second point cloud data; and performing clustering analysis on the second point cloud data using a clustering algorithm to segment out steel coil-related point cloud data to obtain point cloud data of each steel coil.

[0009] In an embodiment of the present application, the point cloud data of the steel coil is translated and rotated, including: translating and rotating the point cloud data of the steel coil according to the geometric center and length direction of the point cloud data of the steel coil to align the geometric center with the coordinate origin and make the length direction parallel to the coordinate axis.

[0010] In an embodiment of the present application, a cutting path is generated according to the transformed point cloud data of the steel coil, including: fitting a cylinder equation for the transformed point cloud data of the steel coil using a random sample consensus algorithm; calculating size parameters of the steel coil according to the cylinder equation of the steel coil, the size parameters including the center coordinates, radius and height of the steel coil; obtaining cutting lines on both sides of the steel coil according to the size parameters of the steel coil; and generating a cutting path according to the cutting lines on both sides of the steel coil.

[0011] In an embodiment of the present application, the cutting lines on both sides of the steel coil are obtained according to the size parameters of the steel coil, including: obtaining the coordinates of the cutting start point and the cutting end point of a first cutting line on one side of the steel coil according to the size parameters of the steel coil; and obtaining the coordinates of the cutting start point and the cutting end point of a second cutting line on the other side of the steel coil according to the size parameters of the steel coil.

[0012] In an embodiment of the present application, the cutting path is generated according to the cutting lines located on both sides of the steel coil, comprising: obtaining the coordinates of an intermediate cutting point according to the coordinates of the cutting start point and the cutting end point of the first cutting line and a preset cutting ratio; dividing the first cutting line into a third cutting line and a fourth cutting line according to the coordinates of the intermediate cutting point; and generating the cutting path in the order of the third cutting line, the second cutting line and the fourth cutting line.

[0013] In an embodiment of the present application, the preset cutting ratio is 80% to 95%.

[0014] To achieve the above object and other related objects, the present application further provides a steel coil type scrap steel cutting device, comprising: a data acquisition unit configured to acquire point cloud data of a steel coil to be cut lying on the ground; a coordinate transformation unit configured to translate and rotate the point cloud data of the steel coil to obtain transformed point cloud data of the steel coil; a path planning unit configured to generate a cutting path according to the transformed point cloud data of the steel coil, the cutting path comprising cutting lines located on both sides of the steel coil, and the cutting lines on at least one side being divided into two times of cutting; and a cutting unit configured to cut the steel coil according to the cutting path.

[0015] To achieve the above object and other related objects, the present application further provides an electronic device comprising a processor, a memory and a communication bus; the communication bus is configured to connect the processor and the memory; the processor is configured to execute a computer program stored in the memory to implement the method provided in any one of the above embodiments.

[0016] To achieve the above object and other related objects, the present application further provides a computer readable storage medium having a computer program stored thereon, the computer program being configured to cause a computer to execute the method provided in any one of the above embodiments.

[0017] The present application provides a steel coil type scrap steel cutting method and device, electronic device and storage medium, the method transforms the point cloud data of the steel coil, so that the subsequent path planning calculation process is simpler; the steel coil type scrap steel is in a cylindrical shape, and will deform after cutting due to internal stress, which will change the shape of the steel coil, and the introduction of the split cutting can avoid the shape change of the steel coil after the first cutting, so that the cutting of the steel coil type scrap steel is more reasonable. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description only only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained from these drawings without creative labor.

[0019] Figure 1 Flow chart of the steel coil type scrap steel cutting method provided by an embodiment of the present application;

[0020] Figure 2 Detailed flow chart of step S100 provided by an embodiment of the present application;

[0021] Figure 3 Detailed flow chart of step S120 provided by an embodiment of the present application;

[0022] Figure 4 Detailed flow chart of step S130 provided by an embodiment of the present application;

[0023] Figure 5 Detailed flow chart of step S300 provided by an embodiment of the present application;

[0024] Figure 6 Detailed flow chart of step S330 provided by an embodiment of the present application;

[0025] Figure 7 Detailed flow chart of step S340 provided by an embodiment of the present application;

[0026] Figure 8 Schematic diagram of the steel coil type scrap steel cutting device provided by an embodiment of the present application;

[0027] Figure 9 Structural schematic diagram of an electronic device provided by an embodiment of the present application.

[0028] Legend of reference signs: 101, data acquisition unit; 102, coordinate conversion unit; 103, path planning unit; 104, cutting unit; 201, processor; 202, memory. DETAILED DESCRIPTION

[0029] The following describes the embodiments of the present invention through specific embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. It should be noted that the following embodiments and the features in the embodiments can be combined with each other unless they conflict. In addition to the specific methods, equipment, and materials used in the embodiments, based on the understanding of the prior art by those skilled in the art and the description of the present invention, any methods, equipment, and materials of the prior art that are similar or equivalent to the methods, equipment, and materials in the embodiments of the present invention can also be used to implement the present invention.

[0030] It should be understood that the terms used in the examples of the present invention are for describing specific embodiments rather than for limiting the scope of protection of the present invention. Unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as those generally understood by those skilled in the art.

[0031] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In some of the embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring the embodiments of the present invention.

[0032] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functions and operations that may be implemented by the methods and computer program products of various embodiments disclosed in the present invention. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0033] See Figure 1 , Figure 1 A method for cutting coiled scrap steel provided in one embodiment of the present invention includes steps S100 to S400.

[0034] Step S100, obtain the point cloud data of the steel coil in a lying posture to be cut. In order to realize the cutting of the steel coil type scrap steel, the first step is to obtain the point cloud data of the steel coil. For the steel coil type scrap steel, it needs to be in a lying posture, which is beneficial to subsequent cutting path planning. The lying posture specifically refers to that the core of the steel coil is arranged horizontally.

[0035] Please refer to Figure 2 In an embodiment of the present application, step S100 includes steps S110-S130.

[0036] Step S110, obtain the initial point cloud data of a plurality of local regions of the scrap steel to be cut, and the initial point cloud data of adjacent local regions has an overlapping part. The scrap steel is generally placed on the ground when cutting. A truss can be arranged above the stacking area of the scrap steel. A motor and a 3D structured light camera are arranged on the truss. The displacement of the camera can be realized by controlling the motor. The initial point cloud data of the scrap steel at the overhead angle in a local region can be scanned and collected each time the displacement is performed. The initial point cloud data of a plurality of local regions can be obtained at a plurality of positions.

[0037] Step S120, splice the initial point cloud data of a plurality of local regions of the scrap steel to be cut to obtain the spliced point cloud data of the scrap steel to be processed. If the area of the scrap steel to be cut is not large, the complete point cloud data can be obtained by one shooting. Therefore, the point cloud splicing process can not be involved in this step.

[0038] Please refer to Figure 3 In an embodiment of the present application, step S120 includes steps S121-S124.

[0039] Step S121, perform down-sampling processing on each initial point cloud data to obtain a plurality of down-sampled point cloud data. For each initial point cloud data of a local region, it contains a large number of points. In order to improve the splicing speed and accuracy, only the key points need to be found for matching. Therefore, in this step, down-sampling processing is performed first.

[0040] In an embodiment of the present application, step S121 includes steps 1.1-1.4.

[0041] Step 1.1, calculate the curvature of each point in each initial point cloud data. In this step, the curvature as a local geometric feature can significantly improve the discrimination of feature point matching, and the high curvature region usually contains more abundant geometric information. This step 1.1 specifically includes steps 1.1.1-1.1.4.

[0042] Step 1.1.1, obtaining several nearest neighbor points of each point in the initial point cloud data. In this step, for a point P, a K-Dimensional Tree (K-D tree) data structure can be constructed to divide the initial point cloud data into multiple hyper-rectangular regions, and then K nearest neighbor search can be used to find the k nearest neighbors of point P. The K-D tree is essentially a binary tree used for fast retrieval of data in multi-dimensional space (in this application, the spatial coordinates, i.e. three dimensions); K nearest neighbor search starts from the root node of the K-D tree and recursively searches downward to the subspace most likely to contain the nearest neighbors, and finally returns the k points closest to P.

[0043] Step 1.1.2, calculating the nearest neighbor centroid of each point according to the coordinates of the several nearest neighbor points of each point. For point P, its nearest neighbor centroid is the geometric center of the k points found above, i.e. the arithmetic mean of the x, y, z coordinates of the k points.

[0044] Step 1.1.3, obtaining the covariance matrix corresponding to each point according to the coordinates of the several nearest neighbor points and the coordinates of the nearest neighbor centroid. The covariance matrix is used to describe the spatial distribution of other points in the neighborhood of a point (such as being tiled on a plane or being scattered in multiple directions), and by analyzing the eigenvalues of the matrix, the local geometric type can be determined (used for curvature calculation and feature point selection later).

[0045] The definition of the covariance matrix is as follows:

[0046]

[0047] In the formula, each element is calculated as follows:

[0048]

[0049] In the formula, k is the number of points in the neighborhood, m j and n j correspond to the coordinate components of the neighborhood points, for example, σ mn is σ xy , m j and n j correspond to the x and y coordinates, and are the mean values of all m j and n j of the neighborhood points, i.e. the nearest neighbor centroid solved in step S212.

[0050] Step 1.1.4: Obtain the curvature of each point based on the eigenvalues ​​of the covariance matrix corresponding to that point. After obtaining the covariance matrix, numerical methods (such as Newton's method) or analytical methods can be used to calculate the eigenvalues ​​λ1, λ2, and λ3 (usually sorted by size: λ1 ≥ λ2 ≥ λ3 ≥ 0), and the curvature can be solved using the following formula:

[0051]

[0052] Eigenvalues ​​and curvature can be used to represent the distribution of a point cloud. For example, if the eigenvalue λ1≈λ2>>λ3, the point cloud near that point is evenly distributed in all directions and the point is on a smooth plane. If the curvature is large, such as λ1>>λ2≈λ3, the point cloud near that point has significant curvature changes, such as edges, inflection points, or bumps. If λ1≈λ2≈λ3, the point is scattered or noise.

[0053] Step 1.2: Divide the initial point cloud data into several grids according to the preset size.

[0054] Step 1.3: Adjust the grid size according to the point cloud density of each grid to obtain a resized grid.

[0055] These two steps correspond to the adaptive mesh downsampling operation. In step 1.2, the initial point cloud data is first evenly divided into grids of a certain size, with different point cloud densities in each grid. In step 1.3, the grid size is adjusted based on the point cloud density. When adjusting the grid size, for grids with sparse point clouds, you can merge adjacent grids or increase the current grid size to retain more points and avoid losing details. For grids with dense point clouds, you can subdivide the grid or reduce the current grid size to reduce the number of sampling points and computation time.

[0056] Step 1.4: From each resized grid, select the centroid and the point of maximum curvature to obtain the downsampled point cloud data. For each grid, simply selecting the centroid and the point of maximum curvature downsamples the initial point cloud data. This downsampling process preserves detail and provides a basis for subsequent accurate matching.

[0057] Step S122: extracting feature points with both topological significance and local geometric significance from each downsampled point cloud data. In a specific embodiment of the present invention, step S212 includes steps 2.1 to 2.4.

[0058] Step 2.1, construct a simplicial complex of the point cloud according to the down-sampled point cloud data. The point cloud itself is a set of 0-simplices (i.e., a set of points), and to extract topological information from it, high-dimensional simplices need to be constructed, i.e., the adjacency relationship between points needs to be established. The simplicial complex construction algorithm includes but is not limited to Vietoris-Rips complex (VR complex), Alpha complex, The simplicial complex is a basic tool in topological data analysis (TDA) for converting discrete point cloud data into a geometric structure that can calculate topological features. The core idea is to build high-dimensional geometric objects (such as line segments, triangles, tetrahedrons, etc.) through the connection relationship between points, so as to reveal the topological features (such as holes, connectivity) implied in the point cloud.

[0059] Step 2.2, calculate the homology group, Betti number, and boundary matrix according to the simplicial complex of the point cloud to obtain the topological features. In this step, the boundary matrix between 0-simplices (points) is calculated to construct the boundary relationship of 1-simplices (edges), and further to calculate the rank of its homology group to obtain the number of connected components β0 and the number of loops β1; on this basis, the boundary matrix of 2-simplices (triangles) is further calculated to determine the number of closed cavities β2 in the point cloud. Topological features may include, for example, connected components, loop structures, or boundary points of holes, etc.

[0060] Step 2.3, calculate the local geometric features of each point in the down-sampled point cloud data. Local geometric features may include, for example, curvature, normal vector, and neighborhood density, etc.

[0061] Step 2.4, obtain feature points with both topological significance and local geometric significance according to the topological features and local geometric features. In this step, for example, points that satisfy the following conditions can be selected as feature points: (1) located at the boundary of stable topological structure determined by persistent homology analysis (such as high Betti number area); (2) local geometric features are significant (such as curvature greater than a threshold or normal vector mutation).

[0062] Step S123, match the feature points of adjacent down-sampled point cloud data using a feature point matching algorithm to obtain a feature point mapping relationship. In a specific embodiment of the present application, step S123 includes steps 3.1-3.4.

[0063] Step 3.1, generate a feature descriptor that fuses topological-geometric information according to the topological invariants and local geometric features of each feature point. In this way, a stable feature vector can be constructed for subsequent matching. The descriptor includes but is not limited to Betti number feature vector (BNFV), persistent homology features (PHF), local geometric features, etc.

[0064] Step 3.2, according to the distance between the feature descriptors of the adjacent down-sampled point cloud data, the initial matching point pair is obtained by similarity threshold or nearest neighbor search. This step will calculate the adjacency matrix of the graph and store the neighbor relationship, wherein the distance calculation between the feature descriptors includes but is not limited to Wasserstein distance, bottleneck distance, graph edit distance, persistent bar chart difference, etc. The purpose of this step is to find the matching points with high structural consistency through similar adjacency matrix.

[0065] The above topological distance calculation method will combine local geometric features (such as curvature, normal vector, point density, etc.) to further improve the accuracy and robustness of matching. By constructing a topological graph and calculating the topological similarity between feature points, high-precision point cloud matching can be achieved, providing a basis for subsequent error elimination and global optimization.

[0066] Step 3.3, the initial matching points are optimized by using the Hungarian algorithm or the optimal transport algorithm to obtain the optimized matching point pair. This step can calculate the optimal matching point pair based on the topological distance matrix to ensure the global consistency and stability of the matching. First, the topological feature descriptor vector of the feature point set is calculated, and the topological graph between the feature points is constructed. The method for calculating the optimal matching relationship between point sets includes but is not limited to optimal transport (Optimal Transport) or Hungarian algorithm (Hungarian Algorithm) etc. To improve the matching accuracy, in this embodiment, the topological graph matching algorithm (Topological Graph Matching) is combined to use the graph edit distance (Graph Edit Distance, GED) to measure the similarity of the topological structure between the feature points, and the distribution of the matching point pair is optimized.

[0067] Step 3.4, the optimized matching point pair is error-eliminated to obtain the feature point mapping relationship.

[0068] In a specific embodiment of the present application, step 3.4 includes: using a topologically enhanced random sample consensus algorithm to eliminate matching point pairs that are not topologically consistent or do not meet geometric constraints to obtain the eliminated matching point pairs. This step uses a topologically enhanced random sample consensus algorithm for error elimination, combines topological invariants (Betti numbers, persistent homology features) and geometric information (normal vector, curvature, point density, etc.) to screen the matching point set, and eliminates point pairs that are not topologically consistent or do not meet geometric constraints, thereby improving the robustness and accuracy of the matching. After error elimination, the global optimization algorithm includes but is not limited to rigid transformation optimization, local geometric consistency optimization, etc., based on the local curvature and normal vector distribution of the point cloud neighborhood, the overall registration accuracy of the point cloud is optimized to ensure the continuity and smoothness of the spliced point cloud.

[0069] Step S124, according to the feature point mapping relationship, the pose of the adjacent initial point cloud data or the down-sampled point cloud data is adjusted to obtain the spliced point cloud data of the scrap steel. After the feature point matching, the false matching elimination and the global optimization are completed, the mapping relationship determined by the matched feature points is used to adjust the attitude (pose) of the adjacent point cloud, so as to ensure the accurate splicing of the point cloud.

[0070] In a specific embodiment of the present application, step S124 includes steps 4.1-4.4.

[0071] Step 4.1, according to the rigid transformation matrix of the feature point mapping relationship, the optimal rigid transformation is solved, and the point cloud pose is preliminarily adjusted according to the solving result.

[0072] Step 4.2, according to the point cloud data after preliminary adjustment, the point cloud pose is further adjusted by using the iterative closest point algorithm. The iterative closest point algorithm (ICP) is a classical algorithm for three-dimensional point cloud registration, which calculates the optimal rigid body transformation (rotation matrix R and translation vector t) by iterative optimization, so that the distance between the corresponding points of two point clouds is minimized. Based on the fine registration of ICP, the preliminarily adjusted point cloud is input, for example, the standard ICP, the ICP based on Betti number and the weighted ICP are used to further optimize the point cloud alignment.

[0073] Step 4.3, the point cloud pose is further adjusted by using the topological consistency constraint to ensure the continuity of the point cloud boundary and the topological structure. The splicing optimization under the topological consistency constraint, for example, the topological graph optimization combined with the global optimization framework is used to maintain the topological consistency of the feature points in the splicing process, to minimize the global error, to ensure the continuity of the point cloud boundary and the topological structure, and to improve the splicing quality.

[0074] Step 4.4, the point cloud smoothing optimization is performed by using the Gaussian mixture model (GMM), the Poisson reconstruction (Poisson Reconstruction) and the local surface fitting (Moving Least Squares, MLS) to obtain the spliced point cloud data.

[0075] Step S130, the spliced point cloud data of the scrap steel to be cut is subjected to ground point cloud elimination, noise filtering and clustering segmentation processing to obtain the point cloud data of each steel coil.

[0076] Please refer to Figure 4 In a specific embodiment of the present application, step S130 includes steps S131-S134.

[0077] Step S131, using the random sample consensus algorithm, the ground plane fitting is performed on the spliced point cloud data of the waste steel to be cut, and a ground plane equation is obtained. In this step, since the ground point cloud is irrelevant to the steel roll body and may affect the fitting result, the random sample consensus algorithm is first used to perform ground plane fitting on the point cloud data.

[0078] Step S132, according to the ground plane equation, the ground point cloud data in the spliced point cloud data of the waste steel to be cut is removed, and first point cloud data is obtained. After obtaining the ground plane equation in step S131, the ground point cloud is removed by calculating the distance of each point to the plane, so as to accurately retain the first point cloud data of the steel roll.

[0079] Step S133, using a filtering algorithm, remove the noise in the first point cloud data, and obtain the second point cloud data. There may be noise points in the first point cloud data caused by sensor errors, environmental interference or other factors. In order to improve the fitting accuracy, a voxel filter (Voxel Grid Filter) is used for noise reduction processing of the point cloud. The voxel filter divides the point cloud into a plurality of volume units (voxels), and replaces the points in each voxel with the center point of the voxel, thereby effectively reducing the data amount and removing the noise.

[0080] Step S134, using a clustering algorithm, the second point cloud data is analyzed by clustering, and the point cloud data related to the steel roll is segmented, and the point cloud data of each steel roll is obtained. The point cloud data may contain other objects or background information unrelated to the steel roll. Through the DBSCAN density clustering algorithm or the Euclidean clustering algorithm, the point cloud data is analyzed by clustering, and the point cloud data related to the steel roll is separated. Specifically, according to the pre-set threshold, the distance between two steel rolls is identified, and different steel roll point clouds are segmented, and other non-steel roll point clouds are excluded according to the size of the point cloud (the number of points in the point cloud).

[0081] Step S200, the point cloud data of the steel roll is translated and rotated to obtain the transformed point cloud data of the steel roll. Specifically, step S200 includes: according to the geometric center and length direction of the point cloud data of the steel roll, the point cloud data of the steel roll is translated and rotated to align the geometric center with the coordinate origin and the length direction with the coordinate axis (for example, X axis). Translation and rotation are performed to make the calculation process more convenient and fast for subsequent generation and planning of the cutting path. In this step, the length direction can be corrected by using the PCA (Principal Component Analysis) method to correct the direction of the point cloud data, calculate the covariance matrix Σ and solve the eigenvectors e1, e2, e3 to determine the main direction (i.e. length direction).

[0082] Step S300, generating a cutting path according to the transformed point cloud data of the steel coil, the cutting path including cutting lines located at two sides of the steel coil, and the cutting line of at least one side being divided into two times of cutting. For the steel coil type scrap steel, due to the existence of stress, special attention needs to be paid to the influence between the first and second cutting when generating the cutting path. In the present application, the side cutting mode is adopted to complete the cutting of the steel coil, and the cutting is completed in two times to eliminate the influence of stress.

[0083] Please refer to Figure 5 In an embodiment of the present application, the cutting path is generated according to the transformed point cloud data of the steel coil, including steps S310-S340.

[0084] Step S310, using the random sample consensus algorithm to perform cylindrical fitting on the transformed point cloud data of the steel coil to obtain a cylindrical equation of the steel coil. The overall shape of the steel coil is similar to a cylinder, and more accurate size of the steel coil can be obtained by cylindrical fitting. The optimal fitting cylinder satisfies: Where P i is a sampling point in the transformed point cloud data, and N is the total number of points.

[0085] Step S320, calculating size parameters of the steel coil according to the cylindrical equation of the steel coil, the size parameters including the center coordinates, radius and height of the steel coil. The radius can also be replaced by the diameter, and the height refers to the height of the cylinder, i.e. the size of the cylinder along the length direction. In addition, the direction of the center line can also be obtained.

[0086] Step S330, obtaining the cutting lines located at two sides of the steel coil according to the size parameters of the steel coil. After obtaining the size parameters of the steel coil, a horizontal plane can be constructed according to the size parameters, and the line segment intersecting with the cylinder is calculated to obtain the cutting lines located at two sides of the steel coil, thereby providing accurate geometric data for subsequent cutting path planning.

[0087] Please refer to Figure 6 In an embodiment of the present application, step S330 includes: S331, obtaining the coordinates of the cutting start point and the cutting end point of the first cutting line located at one side of the steel coil according to the size parameters of the steel coil; and S332, obtaining the coordinates of the cutting start point and the cutting end point of the second cutting line located at the other side of the steel coil according to the size parameters of the steel coil. In this embodiment, the first cutting line and the second cutting line located at two sides of the steel coil are taken as the final cutting lines, and the coordinates of the cutting start point and the cutting end point of the cutting line can be directly calculated by using the size parameters of the steel coil, for example, the coordinates can be calculated according to the following formula:

[0088] x 11 = x C -H / 2, y 11 =y C +r, x12 =x C +H / 2,y 12 =y C +r;

[0089] x 21 =x C -H / 2,y 21 =y C -r,x 22 =x C +H / 2,y 22 =y C -r;

[0090] In the formula, (x 11 ,y 11 ) and (x 12 ,y 12 ) is the cutting starting point P of the first cutting line 11 and cutting end point P 12 Coordinates, (x 21 ,y 21 ) and (x 22 ,y 22 ) is the cutting starting point P of the second cutting line 21 and cutting end point P 12 Coordinates, (x C ,y C ) is the center coordinate of the steel coil, r is the radius of the steel coil, and H is the height of the steel coil.

[0091] Step S340: Generate a cutting path according to the cutting lines on both sides of the steel coil.

[0092] See Figure 7 In a specific embodiment of the present invention, step S340 includes: S341, obtaining the coordinates of the middle cutting point based on the coordinates of the cutting start and end points of the first cutting line and a preset cutting ratio; S342, dividing the first cutting line into a third cutting line and a fourth cutting line based on the coordinates of the middle cutting point; S343, generating a cutting path in the order of the third cutting line, the second cutting line, and the fourth cutting line. In this embodiment, the first cutting line is cut in two steps. First, the first cutting line is divided into the third cutting line and the fourth cutting line using the preset cutting ratio, thereby achieving segmented cutting of the first cutting line. The middle cutting point can be recorded as P0, for example.

[0093] In a specific embodiment of the present application, the preset cutting ratio can be, for example, 80% to 95%, and preferably can be set to 90%. This cutting ratio can also be understood as the ratio of the distance between the cutting start point and the intermediate cutting point to the distance between the cutting start point and the cutting end point, that is, the first cutting ratio when cutting the first cutting line. After completing the cutting of the preset ratio (i.e., the third cutting line), the cutting along the second cutting line is performed, so that even if the second cutting line is completely cut, the steel coil will not be deformed. Finally, the cutting of the remaining part of the first cutting line (i.e., the fourth cutting line) is completed.

[0094] It can be understood that, for a single steel coil, it corresponds to three cutting lines, which are connected by five cutting points (P 11 , P 12 , P 21 , P 22 , P0) according to the above requirements. In order to facilitate the generation of the cutting path, the coordinates of these cutting points can be inversely transformed into the world coordinate system, and the inverse transformation process corresponds to the process in step S200. For multiple steel coils, all cutting points and cutting lines can be integrated to plan a more reasonable cutting path.

[0095] When planning the path, for example, the A* search algorithm or the Dijkstra algorithm can be combined to plan the robot path, to ensure that the cutting path is optimal, to optimize the cutting path according to the shape of the steel coil, and to ensure that the cutting process is efficient and saves materials. Reinforcement learning (Reinforcement Learning) method can also be used to optimize the path, to reduce the cutting time and improve the cutting efficiency, to combine the cutting energy consumption model, to optimize the cutting sequence through the energy minimization principle, to reduce the steel coil thermal deformation, and to improve the cutting quality.

[0096] Step S400, cutting the steel coil according to the cutting path. After obtaining the cutting path, the cutting head can be driven along the cutting path by the mechanical hand to cut the steel coil, so that the cutting of the steel coil type scrap steel can be realized.

[0097] It should be noted that the step division of the above methods is only for the purpose of clear description, and when implemented, one step can be combined or some steps can be split and decomposed into multiple steps, as long as the same logical relationship is included, and all are within the protection scope of the present application; adding irrelevant modifications or introducing irrelevant designs in the algorithm or process, but not changing the core design of the algorithm and process are within the protection scope of the patent.

[0098] Please refer to Figure 8 , Figure 8A steel coil type scrap steel cutting device provided by an embodiment of the present application comprises a data acquisition unit 101, a coordinate transformation unit 102, a path planning unit 103 and a cutting unit 104. The data acquisition unit 101 is configured to acquire point cloud data of a steel coil to be cut lying on the ground. The coordinate transformation unit 102 is configured to perform translation and rotation on the point cloud data of the steel coil to obtain transformed point cloud data of the steel coil. The path planning unit 103 is configured to generate a cutting path according to the transformed point cloud data of the steel coil, the cutting path comprising cutting lines located on both sides of the steel coil, and the cutting lines on at least one side being divided into two times of cutting. The cutting unit 104 is configured to cut the steel coil according to the cutting path.

[0099] It should be noted that the cutting device of the embodiment corresponds to the cutting method described above, and the functional modules in the cutting device correspond to the respective steps in the cutting method. The cutting device of the embodiment can be implemented in cooperation with the cutting method, that is, the related technical details mentioned in the cutting method of the above embodiment can also be applied to the cutting device of the embodiment without conflict.

[0100] Please refer to Figure 9 , Figure 9 An electronic device provided by an embodiment of the present application comprises a processor 201, a memory 202 and a communication bus, the communication bus being configured to connect the processor 201 and the memory 202, and the processor 201 being configured to execute a computer program stored in the memory 202 to implement the cutting method described above.

[0101] The electronic device described above is a device capable of automatically performing numerical calculation and / or information processing according to pre-set or stored instructions, and the hardware thereof includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a digital signal processor (DSP), an embedded device, etc.

[0102] The electronic device described above can be any electronic product capable of human-computer interaction with a user, such as a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an interactive Internet Protocol Television (IPTV), a smart wearable device, etc.

[0103] The electronic device can also include a network device and / or a user device. The network device includes, but is not limited to, a single network server, a server group composed of multiple network servers, or a cloud composed of a large number of hosts or network servers based on cloud computing.

[0104] The network in which the electronic device is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (VPN), and the like.

[0105] The processor can be, for example, a general-purpose processor, including a central processing unit (CPU), a network processor (NP), and the like; a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and the like. The memory can include a random access memory (RAM) and can also include a non-volatile memory, such as at least one disk storage.

[0106] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon, the computer program being used to cause a computer to execute the cutting method.

[0107] The above embodiments are only illustrative of the principles and effects of the present application, and are not intended to limit the present application. Any person skilled in the art can modify or change the above embodiments without departing from the spirit and scope of the present application. Therefore, all equivalent modifications or changes made by those skilled in the art without departing from the spirit and technical ideas disclosed by the present application should be covered by the claims of the present application.

Claims

1. A method for cutting steel coil scrap, characterized in that: include: Acquire point cloud data of the steel coil to be cut in a horizontal position; translating and rotating the point cloud data of the steel coil to obtain transformed point cloud data of the steel coil; generating a cutting path based on the transformed point cloud data of the steel coil, wherein the cutting path includes cutting lines located on both sides of the steel coil, and the cutting line on at least one side is divided into two steps to complete the cutting; The steel coil is cut according to the cutting path.

2. The method for cutting coil-shaped scrap steel according to claim 1, characterized in that: include: Obtain point cloud data of the steel coil to be cut in a horizontal position, including: Acquiring initial point cloud data of a plurality of local areas of the scrap steel to be cut, wherein the initial point cloud data of adjacent local areas have overlapping portions; performing splicing processing on the initial point cloud data of multiple local areas of the scrap steel to be cut to obtain spliced ​​point cloud data of the scrap steel to be processed; The spliced ​​point cloud data of the scrap steel to be cut are subjected to ground point cloud elimination, noise filtering and cluster segmentation processing to obtain point cloud data of each steel coil.

3. The method for cutting coil-shaped scrap steel according to claim 1, characterized in that: include: The spliced ​​point cloud data of the scrap steel to be cut is subjected to ground point cloud removal, noise filtering, and cluster segmentation processing to obtain point cloud data of each steel coil, including: Using a random sampling consensus algorithm, performing ground plane fitting on the spliced ​​point cloud data of the scrap steel to be cut to obtain a ground plane equation; Eliminating ground point cloud data from the spliced ​​point cloud data of the scrap steel to be cut according to the ground plane equation to obtain first point cloud data; Using a filtering algorithm to remove noise from the first point cloud data to obtain second point cloud data; A clustering algorithm is used to perform cluster analysis on the second point cloud data, segment the point cloud data related to the steel coil, and obtain the point cloud data of each steel coil.

4. The method for cutting coil-shaped scrap steel according to claim 1, characterized in that: include: Generating a cutting path according to the transformed point cloud data of the steel coil includes: Performing cylindrical fitting on the transformed point cloud data of the steel coil using a random sampling consensus algorithm to obtain a cylindrical equation of the steel coil; Calculating the dimensional parameters of the steel coil according to the cylinder equation of the steel coil, wherein the dimensional parameters include the center coordinates, radius, and height of the steel coil; According to the size parameters of the steel coil, cutting lines are obtained on both sides of the steel coil; A cutting path is generated according to the cutting lines located on both sides of the steel coil.

5. The method for cutting coil-shaped scrap steel according to claim 4, characterized in that: include: According to the size parameters of the steel coil, cutting lines located on both sides of the steel coil are obtained, including: According to the size parameters of the steel coil, coordinates of a cutting start point and a cutting end point of a first cutting line located on one side of the steel coil are obtained; According to the size parameters of the steel coil, the coordinates of the cutting start point and the cutting end point of the second cutting line located on the other side of the steel coil are obtained.

6. The method for cutting coil-shaped scrap steel according to claim 5, characterized in that: include: According to the cutting lines on both sides of the steel coil, a cutting path is generated, including: Obtaining the coordinates of the middle cutting point according to the coordinates of the cutting start point and the cutting end point of the first cutting line and a preset cutting ratio; dividing the first cutting line into a third cutting line and a fourth cutting line according to the coordinates of the middle cutting point; The cutting path is generated in the order of the third cutting line, the second cutting line, and the fourth cutting line.

7. The method for cutting coil-shaped scrap steel according to claim 6, characterized in that: include: The preset cutting ratio is 80% to 95%.

8. A device for cutting steel coil scrap, characterized in that: include: A data acquisition unit, used to acquire point cloud data of the steel coil to be cut lying on the ground; a coordinate transformation unit, configured to translate and rotate the point cloud data of the steel coil to obtain transformed point cloud data of the steel coil; a path planning unit, configured to generate a cutting path based on the transformed point cloud data of the steel coil, wherein the cutting path includes cutting lines on both sides of the steel coil, and the cutting of at least one side of the cutting line is completed in two steps; A cutting unit is used to cut the steel coil according to the cutting path.

9. An electronic device, characterized in that: The system comprises a processor, a memory and a communication bus; the communication bus is used to connect the processor and the memory; the processor is used to execute the computer program stored in the memory to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that A computer program is stored thereon, and the computer program is used to enable a computer to execute the method according to any one of claims 1 to 7.

Images