Steel bundle tag welding method, device, apparatus, and storage medium

Abstract

The application provides a steel bundle label welding method, device, equipment and storage medium, and relates to the technical field of computers. The steel bundle label welding method comprises the following steps: acquiring a two-dimensional image and a three-dimensional point cloud of an end face of a target steel bundle; acquiring a two-dimensional mask image of a single steel bar contained in the target steel bundle; acquiring an end face point cloud corresponding to the single steel bar based on the two-dimensional mask image and the three-dimensional point cloud; and determining a target welding point based on the three-dimensional point cloud, the end face point cloud corresponding to the single steel bar and label information to be welded on the end face, wherein the target welding point is a welding position. The application can more flexibly and accurately determine the label welding point, so that the welded label can better meet the welding requirements.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a method, apparatus, equipment and storage medium for welding steel bundled signs. Background Technology

[0002] In the production and manufacturing process of steel bars, they are delivered to customers in bundles. It is necessary to mark the bundled steel bars with labels, which are welded to the end face of the steel bundle. The labels are used to display information such as the steel type, specifications, production date, and length of a single steel bar.

[0003] Currently, the following method is commonly used to weld signs to the end face of steel bundles: acquiring an image of the end face of the steel bundle to be welded, performing image processing on the end face image (such as image segmentation and contour extraction) to determine the welding points of the sign; and controlling the telescopic welding torch to weld the sign to the steel bundle based on the three-dimensional coordinates of the welding points. However, the welding points determined by the above method are not accurate enough. Summary of the Invention

[0004] This application provides a method, apparatus, equipment, and storage medium for welding steel bundle labels, to solve the problem that the welding points of the labels are not accurately determined when welding labels to the end face of steel bundles in the current method.

[0005] In a first aspect, this application provides a method for welding steel bundle nameplates, including:

[0006] Acquire two-dimensional images and three-dimensional point clouds of the end face of the target steel bundle;

[0007] Obtain a two-dimensional mask image of a single rebar contained in the target steel bundle;

[0008] Based on two-dimensional mask images and three-dimensional point clouds, the end face point cloud corresponding to a single steel bar is obtained;

[0009] Based on the 3D point cloud, the end face point cloud corresponding to a single steel bar, and the label information to be welded on the end face, the target welding point is determined, and the target welding point is the welding position.

[0010] Optionally, the sign information includes the sign size and the position of the welding position relative to the sign. Based on the 3D point cloud, the end face point cloud corresponding to a single rebar, and the sign information to be welded to the end face, the target welding point is determined, including: determining the target welding area based on the 3D point cloud, sign size, and position information; determining the target rebar in the target welding area based on the end face point cloud corresponding to a single rebar; and determining the center point of the target rebar as the target welding point.

[0011] Optionally, the target welding area is determined based on the 3D point cloud, the size of the sign, and the position information, including: obtaining the end face contour of the target steel bundle based on the 3D point cloud; determining the fixed area of ​​the sign on the end face of the target steel bundle based on the size of the sign and the end face contour of the target steel bundle; and determining the target welding area based on the position information and the fixed area.

[0012] Optionally, the target welding area is determined based on the 3D point cloud, the size of the sign, and the position information, including: obtaining the center pose of the end face contour of the target steel bundle based on the 3D point cloud; determining the center position of the target welding area based on the center pose, the size of the sign, and the position information; and determining the target welding area based on the center position and the preset welding range.

[0013] Optionally, the target reinforcing bar in the target welding area is determined based on the end face point cloud corresponding to a single reinforcing bar, including: determining the end face shape of the reinforcing bar in the target welding area based on the end face point cloud corresponding to a single reinforcing bar; and determining the reinforcing bar whose end protrudes relative to the adjacent reinforcing bar as the target reinforcing bar based on the end face shape.

[0014] Optionally, obtaining a two-dimensional mask image of a single rebar contained in the target steel bundle includes: inputting the two-dimensional image into a target detection model for target detection, and obtaining a two-dimensional mask image of a single rebar contained in the target steel bundle.

[0015] Optionally, the target detection model is obtained by: acquiring sample images containing the end face of the steel bundle and its annotation information, which is used to determine the mask of a single rebar in the steel bundle; inputting the sample images into the initial target detection model for fitting processing to obtain a loss function value, which is used to fit the mask of a single rebar; adjusting the parameters of the initial target detection model based on the loss function value, iteratively training the initial target detection model until the calculated loss function value meets the preset evaluation conditions, thus obtaining the target detection model.

[0016] Optionally, after obtaining a two-dimensional mask image of a single rebar contained in the target steel bundle, the steel bundle label welding method further includes: obtaining the number of rebars contained in the target steel bundle based on the two-dimensional mask image of the single rebar contained in the target steel bundle.

[0017] Optionally, the steel bundle label welding method further includes: acquiring a target two-dimensional image of the end face of the steel bundle to which the label has been welded at the target welding point; inputting the target two-dimensional image into a re-inspection model to obtain the re-inspection result; if the re-inspection result is that there is no label, then performing the steps of acquiring the two-dimensional image and three-dimensional point cloud of the end face of the target steel bundle.

[0018] Optionally, the re-inspection model is obtained through the following methods: acquiring sample images, which include the end face of the steel bundle and annotation information such as whether a label is welded to the end face of the steel bundle; dividing the sample images into a training set and a validation set according to a certain ratio; inputting the training set into the initial re-inspection model for training, and obtaining the loss function value through the loss function of the initial re-inspection model; stopping training when the loss function value meets the preset evaluation conditions; inputting the validation set into the trained initial re-inspection model, outputting sample images with annotation information; comparing the annotation information of the sample images with annotation information with the annotation information of the corresponding sample images in the validation set; and completing the training of the re-inspection model when the accuracy is higher than a threshold.

[0019] Secondly, this application provides a steel bundle label welding device, comprising:

[0020] The first acquisition module is used to acquire two-dimensional images and three-dimensional point clouds of the end face of the target steel bundle;

[0021] The second acquisition module is used to acquire a two-dimensional mask image of a single steel bar contained in the target steel bundle;

[0022] The third acquisition module is used to acquire the end face point cloud corresponding to a single steel bar based on the two-dimensional mask image and the three-dimensional point cloud;

[0023] The determination module is used to determine the target welding point based on the 3D point cloud, the end face point cloud corresponding to a single steel bar, and the label information to be welded on the end face. The target welding point is the welding position.

[0024] Optionally, the sign information includes the sign size and the position of the welding position relative to the sign. The determination module is specifically used to: determine the target welding area based on the 3D point cloud, sign size and position information; determine the target rebar in the target welding area based on the end face point cloud corresponding to a single rebar; and determine the center point of the target rebar as the target welding point.

[0025] Optionally, when determining the target welding area based on the 3D point cloud, sign size, and position information, the determining module is specifically used to: obtain the end face contour of the target steel bundle based on the 3D point cloud; determine the fixed area of ​​the sign on the end face of the target steel bundle based on the sign size and the end face contour of the target steel bundle; and determine the target welding area based on the position information and the fixed area.

[0026] Optionally, when determining the target welding area based on the 3D point cloud, sign size, and position information, the determining module is specifically used to: obtain the center pose of the end face contour of the target steel bundle based on the 3D point cloud; determine the center position of the target welding area based on the center pose, sign size, and position information; and determine the target welding area based on the center position and the preset welding range.

[0027] Optionally, when the determination module is used to determine the target rebar in the target welding area based on the end face point cloud corresponding to a single rebar, it is specifically used to: determine the end face shape of the rebar in the target welding area based on the end face point cloud corresponding to a single rebar; and determine the rebar whose end protrudes relative to the adjacent rebar as the target rebar based on the end face shape.

[0028] Optionally, the second acquisition module is specifically used to: input the two-dimensional image into the target detection model for target detection, and acquire a two-dimensional mask image of a single steel bar contained in the target steel bundle.

[0029] Optionally, the steel bundle label welding device further includes a fourth acquisition module, used to obtain the target detection model in the following ways: acquiring sample images, which include the end face of the steel bundle and the annotation information of the end face of the steel bundle, the annotation information being used to determine the mask of a single rebar in the steel bundle; inputting the sample images into the initial target detection model for fitting processing to obtain a loss function value, the fitting processing being used to fit the mask of a single rebar; adjusting the parameters of the initial target detection model according to the loss function value, iteratively training the initial target detection model until the calculated loss function value meets the preset evaluation conditions, thus obtaining the target detection model.

[0030] Optionally, the second acquisition module is further configured to: after acquiring a two-dimensional mask image of a single rebar contained in the target steel bundle, obtain the number of rebars contained in the target steel bundle based on the two-dimensional mask image of the single rebar contained in the target steel bundle.

[0031] Optionally, the steel bundle label welding device also includes a processing module for: acquiring a target two-dimensional image of the end face of the steel bundle to which the label has been welded at the target welding point; inputting the target two-dimensional image into a re-inspection model to obtain the re-inspection result; if the re-inspection result is that there is no label, then performing the steps of acquiring the two-dimensional image and three-dimensional point cloud of the end face of the target steel bundle.

[0032] Optionally, the steel bundle label welding device further includes a fifth acquisition module, used to obtain the re-inspection model in the following ways: acquiring sample images, which include the end face of the steel bundle and the labeling information of whether the end face of the steel bundle is welded with a label; dividing the sample images into a training set and a validation set according to a certain ratio; inputting the training set into the initial re-inspection model for training, and obtaining the loss function value through the loss function of the initial re-inspection model; stopping training when the loss function value meets the preset evaluation conditions; inputting the validation set into the trained initial re-inspection model, outputting sample images with labeled information, comparing the labeled information of the sample images with labeled information with the labeled information of the corresponding sample images in the validation set, and completing the training of the re-inspection model when the accuracy is higher than a threshold.

[0033] Thirdly, this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0034] The memory stores instructions that the computer executes;

[0035] The processor executes computer execution instructions stored in memory to implement the steel bundle label welding method as described in the first aspect of this application.

[0036] Fourthly, this application provides a computer-readable storage medium storing computer program instructions, which, when executed by a processor, implement the steel bundle label welding method described in the first aspect of this application.

[0037] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the steel bundle label welding method as described in the first aspect of this application.

[0038] The steel bundle label welding method, apparatus, equipment, and storage medium provided in this application obtain a two-dimensional mask image of each individual rebar contained in the target steel bundle from a two-dimensional image of the end face of the target steel bundle. Based on the two-dimensional mask image and the three-dimensional point cloud of the end face of the target steel bundle, the end face point cloud corresponding to each individual rebar is obtained, enabling accurate acquisition of the end face point cloud corresponding to each individual rebar. Based on the three-dimensional point cloud of the end face of the target steel bundle, the end face point cloud corresponding to each individual rebar, and the label information to be welded to the end face, the target welding point is determined, which is the welding position. Because this application combines the three-dimensional point cloud of the end face of the target steel bundle and the end face point cloud corresponding to each individual rebar to determine the target welding point, it can more flexibly and accurately determine the label welding point, thereby enabling the welded label to better meet welding requirements. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application;

[0041] Figure 2 A flowchart illustrating a method for welding steel bundle labels according to an embodiment of this application;

[0042] Figure 3 A schematic diagram of a three-dimensional point cloud provided in an embodiment of this application;

[0043] Figure 4 A schematic diagram of a two-dimensional mask image of a single steel bar contained in a target steel bundle provided in an embodiment of this application;

[0044] Figure 5 A schematic diagram of the end face point cloud corresponding to a single steel bar provided in an embodiment of this application;

[0045] Figure 6 A flowchart illustrating a method for welding steel bundle labels according to another embodiment of this application;

[0046] Figure 7 A schematic diagram of the center pose of the end face contour of a target steel bundle provided in an embodiment of this application;

[0047] Figure 8 A schematic diagram showing the center position of the target welding area according to an embodiment of this application;

[0048] Figure 9 A schematic diagram of the target welding area provided in an embodiment of this application;

[0049] Figure 10 A schematic diagram showing the center position of a single reinforcing bar according to an embodiment of this application;

[0050] Figure 11 A schematic diagram of the target reinforcing bar in the target welding area provided in an embodiment of this application;

[0051] Figure 12 A schematic diagram showing a target steel bundle welded with a label, provided in one embodiment of this application;

[0052] Figure 13 This is a schematic diagram of the structure of a steel bundle label welding device provided in one embodiment of this application;

[0053] Figure 14 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0055] The collection, storage, use, processing, transmission, provision, and disclosure of financial data or user data involved in the technical solution of this application all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0056] Currently, when welding labels to the end face of steel bundles, traditional computer vision methods such as image segmentation and contour extraction are typically used to process 2D or depth images of the end face of the steel bundle to determine the welding points and the number of rebars contained in the bundle. However, the welding points determined by these methods are not accurate enough. The label placement may not be centered, resulting in the label not being at the center of a single rebar, and the selected rebar may be a concave rebar, leading to welding failure. Furthermore, when determining the number of rebars in the bundle using these methods, inaccurate counting may occur when the rebars are tightly fitted, frequently resulting in missed detections.

[0057] To address the aforementioned issues, this application provides a method, apparatus, equipment, and storage medium for welding steel bundle labels. By combining a two-dimensional image and a three-dimensional point cloud of the end face of the target steel bundle, the two-dimensional image is used to obtain a two-dimensional mask image of each individual rebar in the target steel bundle and to determine the number of rebars in the target steel bundle. The three-dimensional point cloud is separated according to each individual rebar, and then the welding position of the label on the target steel bundle is determined based on the overall point cloud of the end face of the target steel bundle and the separate point clouds of each individual rebar. This allows for more flexible and accurate determination of the label welding point, enabling the welded label to better meet welding requirements and more accurately determine the number of rebars in the target steel bundle.

[0058] The following section provides examples illustrating the application scenarios of the solution provided in this application.

[0059] Figure 1 This is a schematic diagram illustrating an application scenario provided by an embodiment of this application. For example... Figure 1 As shown, in this application scenario, server 101 acquires a two-dimensional image and a three-dimensional point cloud of the end face of the target steel bundle using a three-dimensional (3D) structured light camera 102. Based on the two-dimensional image and three-dimensional point cloud of the end face of the target steel bundle, server 101 determines the welding position of the label on the end face of the target steel bundle, and then sends the welding position to robot 103. Robot 103 welds the label to the welding position of the target steel bundle.

[0060] It should be noted that, Figure 1 This is merely a schematic diagram illustrating one application scenario provided by an embodiment of this application. This embodiment does not necessarily represent... Figure 1 The included equipment is not limited, nor is it restricted. Figure 1 The positional relationships between devices are defined. For example, in Figure 1 The application scenario shown may also include a data storage device, which may be an external storage device relative to the server 101 or an internal storage device integrated into the server 101.

[0061] Next, we will introduce the welding method for steel bundled signs through specific examples.

[0062] Figure 2 This is a flowchart illustrating a method for welding steel bundled labels according to an embodiment of this application. The method described in this embodiment can be applied to electronic devices, such as servers or server clusters. Figure 2 As shown, the method in this application embodiment includes:

[0063] S201. Obtain a two-dimensional image and a three-dimensional point cloud of the end face of the target steel bundle.

[0064] In this embodiment of the application, for example, a 3D structured light camera is used to take pictures to obtain a two-dimensional color image and a three-dimensional point cloud of the end face of the target steel bundle. Figure 3 This is a schematic diagram of a three-dimensional point cloud provided in an embodiment of this application, as shown below. Figure 3 As shown, a three-dimensional point cloud of the end face of the target steel bundle is displayed.

[0065] S202. Obtain a two-dimensional mask image of the single steel bar contained in the target steel bundle.

[0066] In this step, after obtaining a two-dimensional image of the end face of the target steel bundle, a two-dimensional mask image of the individual steel bars contained in the target steel bundle can be obtained based on the two-dimensional image of the end face of the target steel bundle. Specifically, for example, target detection can be performed on the two-dimensional image of the end face of the target steel bundle to obtain a two-dimensional mask image of the individual steel bars contained in the target steel bundle. Figure 4 This is a schematic diagram of a two-dimensional mask image of a single steel bar contained in a target steel bundle, provided in an embodiment of this application, as shown below. Figure 4 As shown, each rebar corresponds to a small rectangular frame 401, and each small rectangular frame 401 is a two-dimensional mask image of each rebar. For details on how to obtain the two-dimensional mask image of a single rebar contained in the target steel bundle, please refer to the subsequent embodiments, which will not be repeated here.

[0067] S203. Based on the two-dimensional mask image and the three-dimensional point cloud, obtain the end face point cloud corresponding to a single steel bar.

[0068] In this step, after obtaining a two-dimensional mask image of the individual rebars contained in the target steel bundle, the two-dimensional mask image of the individual rebar can be combined with the three-dimensional point cloud of the end face of the target steel bundle. Based on the two-dimensional mask image of the individual rebar, the three-dimensional point cloud of the end face of the target steel bundle can be segmented to extract the end face point cloud corresponding to the individual rebar. For example, Figure 5 This is a schematic diagram of the end face point cloud corresponding to a single reinforcing bar provided in an embodiment of this application, as shown below. Figure 5 As shown, it illustrates the basis Figure 3 The three-dimensional point cloud of the end face of the target steel bundle shown is obtained, and the end face point cloud of the single steel bar marked with different colors is obtained.

[0069] S204. Based on the 3D point cloud, the end face point cloud corresponding to a single steel bar, and the label information to be welded on the end face, determine the target welding point, which is the welding position.

[0070] In this step, after obtaining the end-face point cloud corresponding to a single rebar, the target welding point can be determined based on the 3D point cloud, the end-face point cloud corresponding to the single rebar, and the information of the label to be welded to the end face. It can be understood that the target welding point is the welding position where the label is welded to the end face of the target steel bundle. For details on how to determine the target welding point based on the 3D point cloud, the end-face point cloud corresponding to the single rebar, and the information of the label to be welded to the end face, please refer to subsequent embodiments; they will not be repeated here.

[0071] Once the target welding point is determined, the label can be welded to the target welding point on the end face of the target steel bundle.

[0072] The steel bundle label welding method provided in this application embodiment obtains a two-dimensional mask image of the individual rebars contained in the target steel bundle from a two-dimensional image of the end face of the target steel bundle. Based on the two-dimensional mask image and the three-dimensional point cloud of the end face of the target steel bundle, the end face point cloud corresponding to the individual rebar is obtained, which can accurately obtain the end face point cloud corresponding to the individual rebar. Based on the three-dimensional point cloud of the end face of the target steel bundle, the end face point cloud corresponding to the individual rebar, and the label information to be welded to the end face, the target welding point is determined, which is the welding position. Since this application embodiment combines the three-dimensional point cloud of the end face of the target steel bundle and the end face point cloud corresponding to the individual rebar to determine the target welding point, it can more flexibly and accurately determine the label welding point, thereby enabling the welded label to better meet the welding requirements.

[0073] Figure 6 A flowchart illustrating a method for welding steel bundle labels according to another embodiment of this application. Based on the above embodiments, this application further describes the method for welding steel bundle labels. Figure 6 As shown, the method in this application embodiment may include:

[0074] S601. Obtain a two-dimensional image and a three-dimensional point cloud of the end face of the target steel bundle.

[0075] For a detailed description of this step, please refer to [link / reference]. Figure 2 The relevant description of S201 in the illustrated embodiment will not be repeated here.

[0076] In the embodiments of this application, Figure 2 Step S202 may further include step S602 as follows:

[0077] S602. Input the two-dimensional image into the target detection model to perform target detection and obtain a two-dimensional mask image of the single steel bar contained in the target steel bundle.

[0078] In this step, the target detection model is pre-trained and used to perform target detection on the two-dimensional image of the end face of the target steel bundle. By inputting the two-dimensional image of the end face of the target steel bundle into the target detection model for target detection, two-dimensional mask images of the individual steel bars contained in the target steel bundle can be directly obtained.

[0079] Optionally, the target detection model is obtained by: acquiring sample images containing the end face of the steel bundle and its annotation information, which is used to determine the mask of a single rebar in the steel bundle; inputting the sample images into the initial target detection model for fitting processing to obtain a loss function value, which is used to fit the mask of a single rebar; adjusting the parameters of the initial target detection model based on the loss function value, iteratively training the initial target detection model until the calculated loss function value meets the preset evaluation conditions, thus obtaining the target detection model.

[0080] For example, the object detection model is implemented using a neural network. The input to the object detection model is a sample image containing the end face of the steel bundle and its annotation information, such as small rectangular boxes corresponding to individual rebars in the bundle to be detected. During the training phase, the sample image is input into the initial object detection model for fitting, fitting a mask (e.g., the position, length, and width of the mask) for each individual rebar in the bundle, and obtaining a loss function value. The loss function is used to measure the accuracy of the mask's position, length, and width. The loss function value is then passed back to the convolutional kernels of each layer of the object detection model to update their weights. That is, based on the loss function value, the parameters of the initial object detection model are adjusted, and the initial object detection model is iteratively trained until the calculated loss function value meets the preset evaluation conditions, thus obtaining the object detection model. Compared with current object detection algorithms, this approach eliminates the need for manually setting complex rules, has good generalization performance, does not require setting anchor points (if the anchor points are not set properly, it will theoretically lead to missed detections), and removes time-consuming operations such as fully connected layers, allowing the object detection model to focus only on local features.

[0081] S603. Based on the two-dimensional mask image of the single steel bar contained in the target steel bundle, obtain the number of steel bars contained in the target steel bundle.

[0082] In this step, after obtaining a two-dimensional mask image of each individual rebar contained in the target steel bundle, the number of rebars contained in the target steel bundle can be determined based on this mask image. It can be understood that the number of masks corresponding to each individual rebar in the target steel bundle represents the total number of rebars contained in the target steel bundle.

[0083] S604. Based on the two-dimensional mask image and the three-dimensional point cloud, obtain the end face point cloud corresponding to a single steel bar.

[0084] For a detailed description of this step, please refer to [link / reference]. Figure 2 The relevant description of S203 in the illustrated embodiment will not be repeated here.

[0085] The signage information includes the signage size and the position of the welding point relative to the signage itself. In this embodiment, Figure 2 Step S204 can further include the following three steps S605 to S607:

[0086] S605. Determine the target welding area based on the 3D point cloud, sign size, and location information.

[0087] In this step, the size of the sign and the position of the welding location relative to the sign can be predetermined. After obtaining the 3D point cloud of the end face of the target steel bundle, the target welding area can be determined based on the 3D point cloud, the sign size, and the position information. Optionally, the target welding area is determined so that the sign, after being welded to the end face of the target steel bundle, will not exceed the end face contour of the target steel bundle; or, based on the requirement of welding the sign to the end face of the target steel bundle, the target welding area is determined so that the degree to which the sign, after being welded to the end face of the target steel bundle, exceeds the end face contour of the target steel bundle is within a set standard.

[0088] Further, optionally, determining the target welding area based on the 3D point cloud, the size of the sign, and the position information may include: obtaining the end face contour of the target steel bundle based on the 3D point cloud; determining the fixed area of ​​the sign on the end face of the target steel bundle based on the size of the sign and the end face contour of the target steel bundle; and determining the target welding area based on the position information and the fixed area.

[0089] For example, refer to Figure 3 The 3D point cloud of the end face of the target steel bundle shown can be used to obtain its end face contour. The size of the label is predetermined, and a fixed area for the label on the end face of the target steel bundle can be determined based on the label size and the end face contour. This fixed area ensures that the label, after being welded to the end face of the target steel bundle, will not exceed the end face contour. Based on this fixed area and the position information of the label's welding location relative to the label, the target welding area can be determined.

[0090] Furthermore, considering ensuring that the marking position is as centered as possible on the end face of the target steel bundle, as a possible implementation method, the target welding area is determined based on the three-dimensional point cloud, the size of the label, and the position information. This may include: obtaining the center pose of the end face contour of the target steel bundle based on the three-dimensional point cloud; determining the center position of the target welding area based on the center pose, the size of the label, and the position information; and determining the target welding area based on the center position and the preset welding range.

[0091] For example, Figure 7A schematic diagram of the center pose of the end face contour of the target steel bundle provided in an embodiment of this application, as shown below. Figure 7 As shown, the center pose of the end face contour of the target steel bundle, obtained from the 3D point cloud of the target steel bundle's end face, is illustrated. Figure 7 The gray circle located at the center of the 3D point cloud. Figure 8 A schematic diagram showing the center position of the target welding area provided in an embodiment of this application, as shown below. Figure 8 As shown, the center position of the target welding area is determined based on the center pose, sign size, and position information. Figure 8 The grayscale circle is located slightly above the center of the 3D point cloud. The preset welding range is, for example, the welding range centered on the center of the target welding area and including at least two steel bars. Figure 9 This is a schematic diagram of the target welding area provided in an embodiment of this application, as shown below. Figure 9 As shown, the area within the rectangle is the target welding area determined based on the center position of the target welding area and the preset welding range. The target welding area contains the poses of three steel bars (corresponding to the three gray circles within the rectangle).

[0092] S606. Determine the target reinforcing bar in the target welding area based on the end face point cloud corresponding to a single reinforcing bar.

[0093] In this step, after determining the target welding area, the target rebar in the target welding area can be determined based on the end face point cloud corresponding to a single rebar.

[0094] Further, optionally, determining the target rebar in the target welding area based on the end face point cloud corresponding to a single rebar may include: determining the end face shape of the rebar in the target welding area based on the end face point cloud corresponding to a single rebar; and determining the rebar whose end protrudes relative to the adjacent rebar as the target rebar based on the end face shape.

[0095] For example, Figure 10 This is a schematic diagram of the center pose of a single reinforcing bar provided in an embodiment of this application, as shown below. Figure 10 As shown, the center pose of a single rebar can be calculated based on the point cloud of its corresponding end face, where 1001 is the X-axis, 1002 is the Y-axis, and 1003 is the Z-axis. The target welding area contains the center poses of multiple rebars, allowing the determination of the end face shape of the rebars within the target welding area. This end face shape includes, for example, whether the end of the rebar protrudes relative to adjacent rebars. Based on the end face shape, rebars whose ends protrude relative to adjacent rebars can be identified as target rebars. It is understood that target rebars are easier to weld signs onto. For example, Figure 11 A schematic diagram of the target reinforcing bar in the target welding area provided in an embodiment of this application, as shown below. Figure 11As shown, the steel bar protruding from the end of the target welding area relative to the adjacent steel bar is taken as the target steel bar, corresponding to... Figure 11 The grayscale circle is located slightly above the center of the 3D point cloud.

[0096] S607. The center point of the target reinforcing bar is determined as the target welding point.

[0097] In this step, after identifying the target reinforcing bar within the target welding area, the center point of the target reinforcing bar can be determined as the target welding point based on its center orientation. After determining the target welding point, the label can be welded to the target welding point on the end face of the target steel bundle. For example, Figure 12 A schematic diagram showing a target steel bundle welded with a label, as provided in one embodiment of this application, is shown below. Figure 12 As shown, the sign is centered at the welding position of the target steel bundle, and the four corners do not exceed the end face outline of the target steel bundle.

[0098] S608. Obtain a target two-dimensional image of the end face of the steel bundle to which the nameplate has been welded at the target welding point.

[0099] For example, a 3D structured light camera can be used to take a picture to obtain a target two-dimensional image of the end face of a steel bundle with a sign welded to the target welding point.

[0100] S609. Input the target two-dimensional image into the re-inspection model to obtain the re-inspection result.

[0101] In this step, the re-inspection model is pre-trained and used to identify whether the label welded at the target welding point has fallen off based on the target 2D image. If it has fallen off, it needs to be re-labeled. After obtaining the target 2D image, the target 2D image can be input into the re-inspection model to directly obtain the re-inspection result.

[0102] Optionally, the re-inspection model is obtained through the following methods: acquiring sample images, which include the end face of the steel bundle and annotation information such as whether a label is welded to the end face of the steel bundle; dividing the sample images into a training set and a validation set according to a certain ratio; inputting the training set into the initial re-inspection model for training, and obtaining the loss function value through the loss function of the initial re-inspection model; stopping training when the loss function value meets the preset evaluation conditions; inputting the validation set into the trained initial re-inspection model, outputting sample images with annotation information; comparing the annotation information of the sample images with annotation information with the annotation information of the corresponding sample images in the validation set; and completing the training of the re-inspection model when the accuracy is higher than a threshold.

[0103] For example, the input to the re-inspection model is sample images, which include information such as the end face of the steel bundle and whether a label is welded to the end face of the steel bundle. For example, there are 100 sample images; this application does not limit this. The sample images are divided into a training set and a validation set proportionally. The training set is input into the initial re-inspection model for training, and the loss function value is obtained through the loss function of the initial re-inspection model. Training stops when the loss function value approaches 0. The trained initial re-inspection model is validated using the validation set. When the accuracy is higher than a threshold, the training of the re-inspection model is complete.

[0104] S610. Determine whether the re-inspection result is that there is no label.

[0105] In this step, assuming the re-inspection result is 1, it indicates that there is a sign; assuming the re-inspection result is 2, it indicates that there is no sign. Based on the re-inspection result, it can be determined whether the re-inspection result indicates no sign. If the re-inspection result is 2, indicating no sign, then step S601 is executed; if the re-inspection result is 1, indicating that there is a sign, then the process ends.

[0106] The steel bundle sign welding method provided in this application embodiment involves inputting a two-dimensional image into a target detection model for target detection, thereby obtaining a two-dimensional mask image of the individual rebars contained in the target steel bundle. Based on this mask image, the number of rebars in the target steel bundle can be determined more accurately. Using the two-dimensional mask image and the three-dimensional point cloud of the end face of the target steel bundle, the end face point cloud corresponding to each individual rebar is obtained, ensuring accurate acquisition of the end face point cloud for each rebar. Based on the three-dimensional point cloud of the end face of the target steel bundle, the end face point cloud corresponding to each individual rebar, and the sign information to be welded to the end face, the target welding point is determined, which is the welding location. Because this application embodiment combines the three-dimensional point cloud of the end face of the target steel bundle and the end face point cloud corresponding to each individual rebar to determine the target welding point, the welding point of the sign can be determined more flexibly and accurately, thus enabling the welded sign to better meet welding requirements.

[0107] The following are embodiments of the apparatus described in this application, which can be used to execute the embodiments of the method described in this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in this application.

[0108] Figure 13 This is a schematic diagram of the structure of a steel bundle label welding device provided in one embodiment of this application, as shown below. Figure 13 As shown, the steel bundle label welding device 1300 of this application embodiment includes: a first acquisition module 1301, a second acquisition module 1302, a third acquisition module 1303, and a determination module 1304. Wherein:

[0109] The first acquisition module 1301 is used to acquire a two-dimensional image and a three-dimensional point cloud of the end face of the target steel bundle.

[0110] The second acquisition module 1302 is used to acquire a two-dimensional mask image of a single steel bar contained in the target steel bundle.

[0111] The third acquisition module 1303 is used to acquire the end face point cloud corresponding to a single steel bar based on a two-dimensional mask image and a three-dimensional point cloud.

[0112] The determination module 1304 is used to determine the target welding point based on the three-dimensional point cloud, the end face point cloud corresponding to a single steel bar, and the label information to be welded on the end face. The target welding point is the welding position.

[0113] Optionally, the sign information includes the sign size and the position of the welding position relative to the sign. The determination module 1304 can be specifically used to: determine the target welding area based on the three-dimensional point cloud, the sign size and position information; determine the target rebar in the target welding area based on the end face point cloud corresponding to a single rebar; and determine the center point of the target rebar as the target welding point.

[0114] Optionally, when determining the target welding area based on the 3D point cloud, the size of the sign, and the position information, the determining module 1304 can be specifically used to: obtain the end face contour of the target steel bundle based on the 3D point cloud; determine the fixed area of ​​the sign on the end face of the target steel bundle based on the size of the sign and the end face contour of the target steel bundle; and determine the target welding area based on the position information and the fixed area.

[0115] Optionally, when determining the target welding area based on the 3D point cloud, sign size, and position information, the determining module 1304 can be specifically used to: obtain the center pose of the end face contour of the target steel bundle based on the 3D point cloud; determine the center position of the target welding area based on the center pose, sign size, and position information; and determine the target welding area based on the center position and the preset welding range.

[0116] Optionally, when determining the target rebar in the target welding area based on the end face point cloud corresponding to a single rebar, the determining module 1304 can specifically be used to: determine the end face shape of the rebar in the target welding area based on the end face point cloud corresponding to a single rebar; and determine the rebar whose end protrudes relative to the adjacent rebar as the target rebar based on the end face shape.

[0117] Optionally, the second acquisition module 1302 can be specifically used to: input the two-dimensional image into the target detection model for target detection, and acquire a two-dimensional mask image of a single steel bar contained in the target steel bundle.

[0118] Optionally, the steel bundle label welding device further includes a fourth acquisition module 1305, used to obtain a target detection model in the following ways: acquiring a sample image, the sample image containing the end face of the steel bundle and the annotation information of the end face of the steel bundle, the annotation information being used to determine the mask of a single rebar in the steel bundle; inputting the sample image into an initial target detection model for fitting processing to obtain a loss function value, the fitting processing being used to fit the mask of a single rebar; adjusting the parameters of the initial target detection model according to the loss function value, iteratively training the initial target detection model until the calculated loss function value meets the preset evaluation conditions, thus obtaining the target detection model.

[0119] Optionally, the second acquisition module 1302 can also be used to: after acquiring a two-dimensional mask image of a single rebar contained in the target steel bundle, obtain the number of rebars contained in the target steel bundle based on the two-dimensional mask image of the single rebar contained in the target steel bundle.

[0120] Optionally, the steel bundle label welding device 1300 also includes a processing module 1306, used to: acquire a target two-dimensional image of the end face of the steel bundle to which the label has been welded at the target welding point; input the target two-dimensional image into the re-inspection model to obtain the re-inspection result; if the re-inspection result is that there is no label, then perform the steps of acquiring the two-dimensional image and three-dimensional point cloud of the end face of the target steel bundle.

[0121] Optionally, the steel bundle label welding device further includes a fifth acquisition module 1307, used to obtain the re-inspection model in the following ways: acquiring sample images, the sample images containing the end face of the steel bundle and the annotation information of whether the end face of the steel bundle is welded with a label; dividing the sample images into a training set and a validation set according to a ratio; inputting the training set into the initial re-inspection model for training, and obtaining the loss function value through the loss function of the initial re-inspection model; stopping training when the loss function value meets the preset evaluation conditions; inputting the validation set into the trained initial re-inspection model, outputting sample images with annotation information, comparing the annotation information of the sample images with annotation information with the annotation information of the corresponding sample images in the validation set, and completing the training of the re-inspection model when the accuracy is higher than a threshold.

[0122] The apparatus of this embodiment can be used to execute the technical solutions of any of the method embodiments shown above. Its implementation principle and technical effect are similar, and will not be repeated here.

[0123] Figure 14 This is a schematic diagram of the structure of an electronic device provided according to an embodiment of this application. Exemplarily, the electronic device may be provided as a server or a computer. (Refer to...) Figure 14The electronic device 1400 includes a processing component 1401, which further includes one or more processors, and memory resources represented by memory 1402 for storing instructions, such as application programs, that can be executed by the processing component 1401. The application programs stored in memory 1402 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 1401 is configured to execute instructions to perform any of the method embodiments described above.

[0124] Electronic device 1400 may also include a power supply component 1403 configured to perform power management of electronic device 1400, a wired or wireless network interface 1404 configured to connect electronic device 1400 to a network, and an input / output (I / O) interface 1405. Electronic device 1400 may operate on an operating system stored in memory 1402, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or similar.

[0125] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the above-mentioned steel bundle label welding method.

[0126] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described steel bundle label welding method.

[0127] The aforementioned computer-readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0128] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an application-specific integrated circuit (ASIC). Alternatively, the processor and the readable storage medium can exist as discrete components in a steel bundle label welding apparatus.

[0129] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0130] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for welding steel bundle nameplates, characterized in that, include: Acquire two-dimensional images and three-dimensional point clouds of the end face of the target steel bundle; Obtain a two-dimensional mask image of a single steel bar contained in the target steel bundle; Based on the two-dimensional mask image, the three-dimensional point cloud is segmented to obtain the end face point cloud corresponding to the single steel bar; Based on the three-dimensional point cloud, the size of the sign to be welded on the end face, and the position of the welding position of the sign relative to the sign, the target welding area is determined; The target reinforcing bar in the target welding area is determined based on the end face point cloud corresponding to the single reinforcing bar. The center point of the target reinforcing bar is determined as the target welding point, which is the welding position.

2. The method for welding steel bundle labels according to claim 1, characterized in that, Determining the target welding area based on the three-dimensional point cloud, the sign size, and the position information includes: Based on the three-dimensional point cloud, obtain the end face contour of the target steel bundle; Based on the size of the sign and the end face contour of the target steel bundle, determine the fixed area of ​​the sign on the end face of the target steel bundle; The target welding area is determined based on the location information and the fixed area.

3. The method for welding steel bundle labels according to claim 1, characterized in that, Determining the target welding area based on the three-dimensional point cloud, the sign size, and the position information includes: Based on the three-dimensional point cloud, obtain the center pose of the end face contour of the target steel bundle; The center position of the target welding area is determined based on the center pose, the size of the sign, and the position information. The target welding area is determined based on the center position and the preset welding range.

4. The method for welding steel bundle labels according to claim 1, characterized in that, The step of determining the target rebar in the target welding area based on the end face point cloud corresponding to the single rebar includes: Based on the end face point cloud corresponding to the single rebar, the end face shape of the rebar in the target welding area is determined; Based on the end face shape, the steel bar whose end protrudes relative to the adjacent steel bar is identified as the target steel bar.

5. The method for welding steel bundle labels according to claim 1, characterized in that, The step of obtaining a two-dimensional mask image of a single rebar contained in the target steel bundle includes: The two-dimensional image is input into the target detection model for target detection to obtain a two-dimensional mask image of the single steel bar contained in the target steel bundle.

6. The method for welding steel bundle labels according to claim 5, characterized in that, The target detection model was obtained through the following method: Acquire a sample image, the sample image containing the end face of the steel bundle and the annotation information of the end face of the steel bundle, the annotation information being used to determine the mask of a single steel bar in the steel bundle; The sample image is input into the initial target detection model for fitting processing to obtain the loss function value. The fitting processing is used to fit the mask of the single steel bar. Based on the loss function value, the parameters of the initial target detection model are adjusted, and the initial target detection model is iteratively trained until the calculated loss function value meets the preset evaluation conditions, thus obtaining the target detection model.

7. The method for welding steel bundle labels according to any one of claims 1 to 6, characterized in that, After obtaining the two-dimensional mask image of the single rebar contained in the target steel bundle, the method further includes: The number of steel bars contained in the target steel bundle is obtained based on a two-dimensional mask image of a single steel bar contained in the target steel bundle.

8. The method for welding steel bundle labels according to any one of claims 1 to 6, characterized in that, Also includes: Obtain a target two-dimensional image of the end face of the steel bundle to which the sign has been welded at the target welding point; The target two-dimensional image is input into the re-inspection model to obtain the re-inspection result; If the re-inspection result is that there is no label, then the step of obtaining the two-dimensional image and three-dimensional point cloud of the end face of the target steel bundle is performed.

9. The method for welding steel bundle nameplates according to claim 8, characterized in that, The re-inspection model was obtained through the following method: Acquire sample images, which include the end face of the steel bundle and labeling information such as whether the end face of the steel bundle has a label welded on it. Divide the sample images into training set and validation set according to the proportion. The training set is input into the initial re-examination model for training, and the loss function value is obtained through the loss function of the initial re-examination model. Training is stopped when the loss function value meets the preset evaluation conditions. The validation set is input into the trained initial re-examination model, which outputs sample images with labeled information. The labeled information of the sample images with labeled information is compared with the labeled information of the corresponding sample images in the validation set. When the accuracy is higher than the threshold, the training of the re-examination model is completed.

10. A welding device for steel bundle labels, characterized in that, include: The first acquisition module is used to acquire two-dimensional images and three-dimensional point clouds of the end face of the target steel bundle; The second acquisition module is used to acquire a two-dimensional mask image of a single steel bar contained in the target steel bundle; The third acquisition module is used to segment the three-dimensional point cloud based on the two-dimensional mask image to obtain the end face point cloud corresponding to the single steel bar. The determination module is used to determine the target welding area based on the three-dimensional point cloud, the size of the sign to be welded on the end face, and the position information of the welding position of the sign relative to the sign; determine the target rebar in the target welding area based on the end face point cloud corresponding to the single rebar; and determine the center point of the target rebar as the target welding point, which is the welding position.

11. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the steel bundle label welding method as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions, which, when executed by a processor, implement the steel bundle label welding method as described in any one of claims 1 to 9.

Images