Method and system for determining scrap cutting point based on multi-view, electronic device and medium

Abstract

The application provides a scrap steel cutting point position determination method and system based on multiple perspectives, an electronic device and a medium. The method comprises the following steps: acquiring original images of scrap steel to be cut under multiple different spatial perspectives, inputting the original images into a scrap steel contour recognition model to obtain scrap steel contour binary images; performing perspective transformation on the scrap steel contour binary images to obtain bird's-eye view perspective images, determining the cutting point position of the scrap steel to be cut based on scrap steel information of the bird's-eye view perspective images, and mapping the cutting point position from an image coordinate to a spatial cutting coordinate of a cutting device. The scrap steel cutting point position determination scheme provided by the application acquires original images of target scrap steel at different angles, converts the perspective of the original images, determines the spatial cutting coordinate based on the bird's-eye view perspective, and thus realizes accurate judgment of the cutting position of the target scrap steel, realizes high-precision and strong-robustness automatic positioning of the target scrap steel in a complex environment, and provides key technical support for the improvement of the efficiency and precision of a scrap steel cutting device.

Description

Technical Field

[0001] This invention relates to the field of steel cutting technology, and in particular to a method, system, electronic device and medium for determining the cutting point of scrap steel based on multiple perspectives. Background Technology

[0002] Scrap steel recycling and reprocessing is a crucial link in the resource recycling of the steel industry, with high-speed and precise scrap steel cutting being the core step in improving scrap steel processing efficiency and quality. However, in actual cutting, scrap steel itself is characterized by irregular shape, large size differences, easy rusting, and complex textures. Furthermore, it is usually stacked disorderly at the recycling site, leading to mutual obstruction and varying spatial orientations among the scrap steel pieces. This complex physical scenario poses a severe challenge to the autonomous sensing and precise positioning systems of the cutting equipment.

[0003] Currently, traditional scrap steel cutting positioning mainly relies on single-view vision-based positioning methods. Specifically, a single industrial camera is set up at a fixed location to acquire planar images, and the cutting point is determined by manually annotating the single planar image or by a single algorithm recognizing the outline of the scrap steel. However, this method has the following inherent limitations: traditional recognition methods may fail to cut due to partial occlusion, limited viewing angle, or uneven lighting, and the fixed single camera cannot acquire visual information of occluded areas, resulting in "blind spots." If the key cutting point is not within the field of view, it is difficult to achieve the positioning requirements for high-precision cutting, thus directly affecting work efficiency and cutting quality. Summary of the Invention

[0004] This invention provides a method, system, electronic device, and medium for determining the cutting point of scrap steel based on multiple perspectives, in order to solve the technical problems of inaccurate positioning, low cutting precision, low cutting efficiency, and poor cutting quality in scrap steel cutting.

[0005] In a first aspect, the present invention provides a method for determining the cutting point of scrap steel based on multiple perspectives, comprising:

[0006] Acquire original images of the scrap steel to be cut from multiple different spatial perspectives; The original image is input into the scrap steel contour recognition model to obtain a binary image of the scrap steel contour corresponding to each spatial viewpoint. A perspective transformation is performed on multiple binary images of the scrap steel outlines to obtain multiple bird's-eye view images; Based on the scrap steel information in multiple bird's-eye view images, the cutting point of the scrap steel to be cut is determined; The cutting points are mapped from the image coordinate system to the spatial cutting coordinates of the cutting device.

[0007] In one embodiment of the present invention, the original image is input into a scrap steel contour recognition model to obtain a scrap steel contour binary image corresponding to each spatial viewpoint, including: performing distortion correction on the original image based on the camera viewpoint parameters corresponding to each original image to obtain multiple corrected images; inputting the corrected images into the scrap steel contour recognition model to generate a scrap steel mask image; and determining the scrap steel contour binary image based on the confidence level of each pixel value in the scrap steel mask image relative to the scrap steel to be cut.

[0008] In one embodiment of the present invention, determining the binary image of the scrap steel outline based on the confidence level of each pixel value in the scrap steel mask image relative to the scrap steel to be cut includes: when the confidence level is greater than or equal to the preset confidence threshold, determining that the corresponding pixel point is located within the outline of the scrap steel to be cut; when the confidence level is less than the preset confidence threshold, determining that the corresponding pixel point is located outside the outline of the scrap steel to be cut; and constructing the binary image of the scrap steel outline based on multiple pixels within the outline of the scrap steel to be cut.

[0009] In one embodiment of the present invention, the scrap steel contour recognition model construction step includes: acquiring sample images of scrap steel samples from multiple different spatial perspectives; labeling the multiple sample images to construct a training dataset of the scrap steel samples; and inputting the training dataset into a preset deep learning image segmentation model for data training to form the scrap steel contour recognition model.

[0010] In one embodiment of the present invention, perspective transformation is performed on multiple binary images of scrap steel outlines to obtain multiple bird's-eye view images, including: determining multiple first homography matrices based on multiple camera viewpoint parameters and bird's-eye view projection matrices respectively; converting the pixel points corresponding to the scrap steel to be cut in the binary images of scrap steel outlines into first homogeneous form coordinates; determining bird's-eye view image coordinates according to the first homogeneous form coordinates and the corresponding first homography matrix; and fusing the coordinates of multiple bird's-eye view images related to the same binary image of scrap steel outlines to obtain the bird's-eye view image.

[0011] In one embodiment of the present invention, determining the cutting point of the scrap steel to be cut based on scrap steel information in multiple bird's-eye view images includes: identifying the quantity of scrap steel to be cut based on the binary image of the scrap steel outline; taking the bird's-eye view image with the largest quantity of scrap steel as the main cutting view; determining the edge center of the scrap steel to be cut according to the main cutting view, and taking the edge center as the cutting point.

[0012] In one embodiment of the present invention, mapping the cutting point from the image coordinate system to the spatial cutting coordinates of the cutting device includes: obtaining the height coordinates of the scrap steel to be cut; determining a second homography matrix based on the camera viewpoint parameters and rotation matrix corresponding to the cutting master viewpoint; converting the cutting point into a second homogeneous form coordinate; determining the two-dimensional cutting coordinates of the cutting device according to the second homogeneous form coordinate and the second homography matrix; and determining the spatial cutting coordinates based on the height coordinates and the two-dimensional cutting coordinates.

[0013] This invention also provides a multi-view scrap steel cutting point determination system, the system comprising: an acquisition module for acquiring original images of the scrap steel to be cut from multiple different spatial perspectives; a scrap steel contour recognition module for inputting the original images into a scrap steel contour recognition model to obtain a binary image of the scrap steel contour corresponding to each spatial perspective; a bird's-eye view determination module for performing perspective transformation on multiple binary images of the scrap steel contour to obtain multiple bird's-eye view images; a planar cutting point determination module for determining the cutting point of the scrap steel to be cut based on the scrap steel information in the multiple bird's-eye view images; and a spatial cutting point determination module for mapping the cutting point from the image coordinate system to the spatial cutting coordinates of the cutting equipment.

[0014] The present invention also provides an electronic device, including 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 a computer program stored in the memory to implement the multi-view scrap steel cutting point determination method as described in any of the preceding claims.

[0015] The present invention also provides a computer-readable storage medium storing computer-readable instructions thereon, which, when executed by a computer processor, cause the computer to perform the multi-view scrap steel cutting point determination method as described above.

[0016] The beneficial effects of this invention are as follows: This invention proposes a method, system, electronic device, and medium for determining the cutting point of scrap steel based on multiple perspectives. The method includes: acquiring original images of the scrap steel to be cut from multiple different spatial perspectives; inputting the original images into a scrap steel contour recognition model to obtain a binary image of the scrap steel contour; performing perspective transformation on the binary image of the scrap steel contour to obtain a bird's-eye view image; determining the cutting point of the scrap steel to be cut based on the scrap steel information in the bird's-eye view image; and mapping the cutting point from image coordinates to the spatial cutting coordinates of the cutting equipment. The scrap steel cutting point determination scheme provided by this invention acquires original images of the target scrap steel from different angles, determines the spatial cutting coordinates based on the bird's-eye view by transforming the perspective of the original images, thereby achieving accurate judgment of the cutting position of the target scrap steel. It enables high-precision and robust automatic positioning of the target scrap steel in complex environments, providing key technical support for improving the efficiency, accuracy, and automation level of scrap steel cutting equipment. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0018] In the attached diagram: Figure 1 This is a schematic diagram of a multi-view scrap steel cutting point determination method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the binary outline of scrap steel provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of a bird's-eye view image after the conversion of the binary image of the scrap steel outline provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the two-dimensional cutting coordinates in an actual cutting scenario provided in this embodiment of the invention; Figure 5 This is a schematic diagram of a multi-view scrap steel cutting point determination system provided in one embodiment of the present invention; Figure 6 This is a schematic diagram illustrating a structure suitable for implementing an electronic device according to an exemplary embodiment of the present invention. Detailed Implementation

[0019] The following specific examples illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments. Various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of the present invention. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0020] It should be noted that the illustrations provided in the following embodiments are only schematic representations of the basic concept of the present invention. The drawings only show the components related to the present invention and are not drawn according to the actual number, shape and size of the components in the actual implementation. In the actual implementation, the form, quantity and proportion of each component can be arbitrarily changed, and the layout of the components may also be more complex.

[0021] In the following description, numerous details are explored to provide a more thorough explanation of embodiments of the invention. However, it will be apparent to those skilled in the art that embodiments of the invention may be practiced without these specific details. In other embodiments, well-known structures and devices are shown in block diagram form rather than in detail to avoid obscuring embodiments of the invention.

[0022] To solve the above problems, such as Figure 1 As shown, this application provides a method for determining the cutting point of scrap steel based on multiple perspectives. This method includes at least steps S110 to S150: Step S110: Obtain original images of the scrap steel to be cut from multiple different spatial perspectives.

[0023] Multiple camera devices at different spatial locations are set up around the platform where the scrap steel to be cut is placed. At least four such cameras are set up to take pictures of the scrap steel to be cut from different spatial perspectives, thus obtaining multiple raw images.

[0024] Step S120: Input the original image into the scrap steel contour recognition model to obtain the scrap steel contour binary image corresponding to each spatial viewpoint.

[0025] In one implementation, the original image is input into the scrap steel contour recognition model to obtain a scrap steel contour binary image corresponding to each spatial viewpoint. This includes: performing distortion correction on the original image based on the camera viewpoint parameters corresponding to each original image to obtain multiple corrected images; inputting the corrected images into the scrap steel contour recognition model to generate a scrap steel mask image; and determining the scrap steel contour binary image based on the confidence level of each pixel value in the scrap steel mask image relative to the scrap steel to be cut.

[0026] Specifically, the camera viewpoint parameters of each camera device include the camera intrinsic parameter matrix K.i and extrinsic parameter matrix Where i is the camera device number, 1≤i≤M, and M is the total number of camera devices, M≥4. This is determined using the intrinsic parameter matrix K. i Distortion correction is performed on each original image to obtain multiple corrected images. The correction process for each pixel value in the corrected images is shown in expression (1): (1) in, These are the pixel coordinates within the original image. To complete the correction of pixel coordinates within the image after distortion correction, It is the inverse of the intrinsic parameter matrix.

[0027] Multiple corrected images are input into the scrap steel contour recognition model D to obtain the scrap steel mask image of the area to be cut. The expression for determining the scrap steel mask image is as follows: S i =D(W) i (2) Among them, W i S is the corrected image corresponding to the i-th camera device. i Let D be the scrap steel mask image corresponding to each corrected image, and D be the scrap steel contour recognition model.

[0028] The confidence level of each pixel value in the scrap steel mask image within the area of ​​scrap steel to be cut is obtained. Each confidence level is compared with a preset confidence threshold, and the binary image of the scrap steel outline to be cut is determined based on the comparison results.

[0029] In one implementation, a binary image of the scrap steel outline is determined based on a comparison between the pixel values ​​in the scrap steel mask image and a preset confidence threshold. This includes: when the confidence level is greater than or equal to the preset confidence threshold, determining that the corresponding pixel is located within the outline of the scrap steel to be cut; when the confidence level is less than the preset confidence threshold, determining that the corresponding pixel is located outside the outline of the scrap steel to be cut; and constructing a binary image of the scrap steel outline based on multiple pixels within the outline of the scrap steel to be cut. Specifically, each pixel value in the scrap steel mask image is segmented. This represents the confidence level that a given pixel belongs to the area of ​​scrap steel to be cut, with a value ranging from 0 to 1. When the confidence level is greater than or equal to a preset confidence threshold τ... A confidence level of 1 indicates that the pixel is within the outline of the scrap steel to be cut. When the confidence level is less than the preset confidence threshold τ, A value of 0 indicates that the pixel is outside the outline of the scrap steel to be cut, belonging to the equipment platform area and the environment area. For example... Figure 2 As shown, the white area represents the area of ​​scrap steel to be cut, and the black area represents the equipment platform and the surrounding environment.

[0030] In one implementation, the scrap steel contour recognition model construction steps include: acquiring sample images of scrap steel samples from multiple different spatial perspectives; labeling the multiple sample images to construct a training dataset of scrap steel samples; and inputting the training dataset into a preset deep learning image segmentation model for data training to form a scrap steel contour recognition model. Specifically, sample images of different types of scrap steel samples are acquired on a device platform. These sample images are captured by cameras from multiple different spatial perspectives. The scrap steel samples in the multiple sample images are labeled to construct a high-quality training dataset of scrap steel samples. The training dataset is then input into a preset deep learning image segmentation model to perform contour learning and recognition on the training dataset, thereby obtaining the scrap steel contour recognition model. Deep learning image segmentation models include, but are not limited to, MaskR-CNN (Mask Region-based Convolutional Neural Network), YOLO-Seg (You Only Look Once Segmentation), and SAM (Segment Anything Model).

[0031] Step S130: Perform perspective transformation on multiple binary images of scrap steel outlines to obtain multiple bird's-eye view images.

[0032] In one embodiment, perspective transformation is performed on multiple binary images of scrap steel outlines to obtain multiple bird's-eye view images, including: determining multiple first homography matrices based on multiple camera view parameters and bird's-eye view projection matrices respectively; converting the pixel points corresponding to the scrap steel to be cut in the binary images of scrap steel outlines into first homogeneous form coordinates; determining the bird's-eye view image coordinates according to the first homogeneous form coordinates and the corresponding first homography matrix; and fusing the coordinates of multiple bird's-eye view images related to the same binary image of scrap steel outlines to obtain a bird's-eye view image.

[0033] Specifically, for each camera device, the first homography matrix from the camera image plane to the bird's-eye view plane is calculated based on the camera view parameters and the bird's-eye view projection matrix of each camera device. The method for obtaining the first homography matrix is ​​shown in expression (3): (3) in, Let be the first homography matrix of the i-th camera device. For bird's-eye view planar projection matrix, Let this be the rotation and translation matrix from the camera's coordinate system to the bird's-eye view coordinate system. This is the inverse of the intrinsic parameter matrix of the camera device.

[0034] Convert each pixel (u, v) of the scrap steel region to be cut in the binary image of the scrap steel outline into first homogeneous coordinates (u, v, 1). T The bird's-eye view transformed coordinates (x, y) of the bird's-eye view plane are obtained through the first homography matrix and the first homogeneous form coordinate transformation. bev y bev w) T The expression for determining the coordinate transformation from a bird's-eye view is shown in (4): (4) in, The x-coordinate of the bird's-eye view transformation coordinates. The vertical coordinate is the coordinate of the bird's-eye view transformation, and W is the scale factor, (u, v, 1). T The coordinates are in the first homogeneous form. Let be the first homography matrix of the i-th camera device.

[0035] The obtained bird's-eye view image coordinates are normalized to obtain the bird's-eye view image coordinates. The expression for determining the bird's-eye view image coordinates is shown in (5): , (5) in, The horizontal axis of the bird's-eye view image icon. The vertical coordinate of the bird's-eye view image. The x-coordinate of the bird's-eye view transformation coordinates. The vertical coordinate of the bird's-eye view transformation coordinate system, where W is the scale factor.

[0036] Each binary image of scrap steel outline corresponds to multiple bird's-eye view coordinates. Multiple bird's-eye view coordinates associated with the same binary image of scrap steel outline are fused to obtain the corresponding bird's-eye view image, such as... Figure 3 As shown, Figure 3 To and Figure 2 The corresponding bird's-eye view image, that is: with Figure 2 The corresponding top-down view.

[0037] Step S140: Based on the scrap steel information in multiple bird's-eye view images, determine the cutting point of the scrap steel to be cut.

[0038] In one implementation, determining the cutting point of the scrap steel to be cut based on scrap steel information from multiple bird's-eye view images includes: identifying the quantity of scrap steel to be cut based on a binary image of the scrap steel contour; selecting the bird's-eye view image with the largest quantity of scrap steel as the main cutting view; and determining the edge center of the scrap steel to be cut based on the main cutting view, and using the edge center as the cutting point. Specifically, quantity identification is performed on each binary image of the scrap steel contour to obtain the quantity of scrap steel to be cut in each image; based on the quantity of scrap steel, the image with the largest quantity of scrap steel is selected from the bird's-eye view images corresponding to each binary image of the scrap steel contour, and this image is used as the main cutting view; the cutting edge length of the scrap steel to be cut is calculated based on the main cutting view, and half of the cutting edge length is used as the edge center of the scrap steel to be cut, and the edge center is used as the cutting point of the bird's-eye view.

[0039] Step S150: Map the cutting points from the image coordinate system to the spatial cutting coordinates of the cutting device.

[0040] In one embodiment, mapping the cutting point from the image coordinate system to the spatial cutting coordinates of the cutting device includes: obtaining the height coordinates of the scrap steel to be cut; determining a second homography matrix based on the camera viewpoint parameters and rotation matrix corresponding to the main cutting viewpoint; converting the cutting point into a second homogeneous form coordinate; determining the two-dimensional cutting coordinates of the cutting device based on the second homogeneous form coordinate and the second homography matrix; and determining the spatial cutting coordinates based on the height coordinates and the two-dimensional cutting coordinates.

[0041] Specifically, because the thickness and shape of each piece of scrap steel to be cut vary, each cutting operation typically only cuts a single-digit number of pieces. The height coordinates Z of the scrap steel to be cut can be obtained through laser ranging or 3D matching based on the type of scrap steel. d The intrinsic parameter matrix K of the camera device corresponding to the main viewpoint is obtained by using Zhang Zhengyou's calibration method or other calibration methods. d and extrinsic parameter matrix [R d |t d ], where R d Let t be a rotation matrix. d The translation vector. The intrinsic parameter matrix K of the camera device corresponding to the main viewpoint. d As shown in expression (6): (6) Among them, K d To cut the intrinsic parameter matrix of the camera device corresponding to the main viewpoint, , The focal length (in pixels) of the camera device in the x and y directions corresponding to the main viewpoint. , The coordinates of the main point.

[0042] The second homography matrix H from the image plane to the cutting plane is obtained through camera calibration. p The second homography matrix H p The definite expression is as follows (7): (7) Among them, H p Let K represent the second homography matrix. d To cut the intrinsic parameter matrix of the camera device corresponding to the main viewpoint, Let t be the first two columns of the rotation matrix R. d It is a translation vector.

[0043] Convert the cutting point to second homogeneous form coordinates P. img =[x, y, 1] T Calculate the two-dimensional transformed coordinates based on the second homography matrix and the second homogeneous form coordinates. =[ , , ] T The expression for determining the two-dimensional transformed coordinates is shown in (8): (8) in, For two-dimensional coordinate transformation, It is the inverse of the second homography matrix. The coordinates are in the second homogeneous form.

[0044] The two-dimensional transformed coordinates are normalized to obtain the two-dimensional cutting coordinates of the cutting equipment. The expression for determining the two-dimensional cutting coordinates is shown in (9): (9) Among them, X d The X-axis coordinate of the two-dimensional cutting coordinate system, Y... d The Y-axis coordinate of the two-dimensional cutting coordinate system. The X-axis coordinate of the two-dimensional transformation coordinate. The Y-axis coordinate of the two-dimensional transformation coordinates. This is the second conversion factor.

[0045] like Figure 4 As shown, the point (1734.0, 598.8) is a two-dimensional cutting coordinate; by fusing the two-dimensional cutting coordinates and the height coordinates, the spatial cutting coordinates P are obtained. 3D = (X d Y d Z d This allows the cutting equipment to cut the scrap steel according to the spatial cutting coordinates, ensuring accurate spatial cutting point positioning and improving the cutting efficiency of the scrap steel.

[0046] like Figure 5 As shown, the present invention also provides a multi-view scrap steel cutting point determination system, the system comprising: Acquisition module 510, the acquisition module, is used to acquire original images of the scrap steel to be cut from multiple different spatial perspectives; The scrap steel contour recognition module 520 is used to input the original image into the scrap steel contour recognition model to obtain a binary image of the scrap steel contour corresponding to each spatial viewpoint. The bird's-eye view determination module 530 is used to perform perspective transformation on multiple binary images of scrap steel outlines to obtain multiple bird's-eye view images. The planar cutting point determination module 540 is used to determine the cutting point of the scrap steel to be cut based on scrap steel information in multiple bird's-eye view images; The spatial cutting point determination module 550 is used to map the cutting point from the image coordinate system to the spatial cutting coordinates of the cutting device.

[0047] It should be noted that the multi-view scrap steel cutting point determination system and the multi-view scrap steel cutting point determination method provided in the above embodiments belong to the same concept. The specific methods of performing each step have been described in detail in the system embodiments and will not be repeated here.

[0048] This invention proposes a method and system for determining the cutting point of scrap steel based on multiple perspectives. The method includes: acquiring original images of the scrap steel to be cut from multiple different spatial perspectives; inputting the original images into a scrap steel contour recognition model to obtain a binary image of the scrap steel contour; performing a perspective transformation on the binary image of the scrap steel contour to obtain a bird's-eye view image; determining the cutting point of the scrap steel to be cut based on the scrap steel information in the bird's-eye view image; and mapping the cutting point from image coordinates to the spatial cutting coordinates of the cutting equipment. The scrap steel cutting point determination scheme provided by this invention acquires original images of the target scrap steel from different angles, determines the spatial cutting coordinates based on the bird's-eye view by transforming the perspective of the original images, thereby achieving accurate judgment of the cutting position of the target scrap steel. It enables high-precision and robust automatic positioning of the target scrap steel in complex environments, providing key technical support for improving the efficiency, accuracy, and automation level of scrap steel cutting equipment.

[0049] In some embodiments, an electronic device is also provided, which may be a server, and its internal structure diagram is shown below. Figure 6As shown, the electronic device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used to communicate with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of the server-side method described above.

[0050] In some embodiments, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it performs the following steps: acquiring original images of scrap steel to be cut from multiple different spatial perspectives; inputting the original images into a scrap steel contour recognition model to obtain a binary image of the scrap steel contour corresponding to each spatial perspective; performing perspective transformation on the multiple binary images of the scrap steel contour to obtain multiple bird's-eye view images; determining the cutting point of the scrap steel to be cut based on the scrap steel information in the multiple bird's-eye view images; and mapping the cutting point from the image coordinate system to the spatial cutting coordinates of the cutting device.

[0051] Embodiments of the present invention also provide a computer-readable storage medium storing a computer program thereon, which, when executed by a computer processor, causes the computer to perform the aforementioned multi-view scrap steel cutting point determination method. This computer-readable storage medium may be included in the electronic device described in the above embodiments, or it may exist independently and not assembled into the electronic device.

[0052] It should be noted that the computer-readable medium shown in the embodiments of the present invention can be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. Computer-readable signal media can also be any computer-readable medium other than computer-readable storage media, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0053] It should be noted that the functions or steps that can be implemented by the computer-readable storage medium or electronic device described above can be referred to the relevant descriptions on the server side and client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0054] The above embodiments are merely illustrative of the principles and effects of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or alter the above embodiments without departing from the spirit and scope of the present invention. Therefore, all equivalent modifications or alterations made by those skilled in the art without departing from the spirit and technical concept disclosed in the present invention should still be covered by the claims of the present invention.

Claims

1. A method for determining the cutting point of scrap steel based on multiple perspectives, characterized in that, include: Acquire original images of the scrap steel to be cut from multiple different spatial perspectives; The original image is input into the scrap steel contour recognition model to obtain a binary image of the scrap steel contour corresponding to each spatial viewpoint. A perspective transformation is performed on multiple binary images of the scrap steel outlines to obtain multiple bird's-eye view images; Based on the scrap steel information in multiple bird's-eye view images, the cutting point of the scrap steel to be cut is determined; The cutting points are mapped from the image coordinate system to the spatial cutting coordinates of the cutting device.

2. The method for determining the cutting point of scrap steel based on multiple perspectives according to claim 1, characterized in that, The original image is input into the scrap steel contour recognition model to obtain binary images of the scrap steel contours corresponding to each spatial viewpoint, including: Based on the camera viewpoint parameters corresponding to each of the original images, distortion correction is performed on the original images to obtain multiple corrected images; The corrected image is input into the scrap steel contour recognition model to generate a scrap steel mask image; Based on the confidence level of each pixel value in the scrap steel mask image relative to the scrap steel to be cut, the binary image of the scrap steel outline is determined.

3. The method for determining the cutting point of scrap steel based on multiple perspectives according to claim 2, characterized in that, Based on the confidence level of each pixel value in the scrap steel mask image relative to the scrap steel to be cut, the binary image of the scrap steel outline is determined, including: When the confidence level is greater than or equal to the preset confidence threshold, it is determined that the corresponding pixel is located within the outline of the scrap steel to be cut. When the confidence level is less than the preset confidence threshold, it is determined that the corresponding pixel is located outside the outline of the scrap steel to be cut. The binary image of the scrap steel outline is constructed based on multiple pixels within the outline of the scrap steel to be cut.

4. The method for determining the cutting point of scrap steel based on multiple perspectives according to claim 2, characterized in that, The steps for constructing the scrap steel contour recognition model include: Acquire sample images of scrap steel from multiple different spatial perspectives; The sample images are labeled to construct a training dataset of the scrap steel samples; The training dataset is input into a preset deep learning image segmentation model for data training to form the scrap steel contour recognition model.

5. The method for determining the cutting point of scrap steel based on multiple perspectives according to claim 2, characterized in that, Perspective transformation was performed on multiple binary images of the scrap steel outlines to obtain multiple bird's-eye view images, including: Multiple first homography matrices are determined based on multiple camera viewpoint parameters and the bird's-eye view projection matrix, respectively; Convert the pixel points corresponding to the scrap steel to be cut in the binary image of the scrap steel outline into first homogeneous coordinates. The bird's-eye view coordinates are determined based on the first homogeneous form coordinates and the corresponding first homography matrix. The coordinates of multiple bird's-eye view images associated with the same binary image of the scrap steel outline are fused to obtain the bird's-eye view image.

6. The method for determining the cutting point of scrap steel based on multiple perspectives according to claim 1, characterized in that, Based on scrap steel information from multiple bird's-eye view images, the cutting point of the scrap steel to be cut is determined, including: Based on the binary image of the scrap steel outline, the quantity of scrap steel to be cut is identified; The bird's-eye view image with the largest quantity of scrap steel is used as the main cutting perspective; Based on the cutting perspective, the edge center of the scrap steel to be cut is determined, and the edge center is used as the cutting point.

7. The method for determining the cutting point of scrap steel based on multiple perspectives according to claim 6, characterized in that, Mapping the cutting points from the image coordinate system to the spatial cutting coordinates of the cutting device includes: Obtain the height coordinates of the scrap steel to be cut; The second homography matrix is ​​determined based on the camera viewpoint parameters and rotation matrix corresponding to the cutting main viewpoint; Convert the cutting points into second homogeneous coordinates; The two-dimensional cutting coordinates of the cutting device are determined based on the second homogeneous form coordinates and the second homography matrix. The spatial cutting coordinates are determined based on the height coordinates and the two-dimensional cutting coordinates.

8. A multi-view scrap steel cutting point determination system, characterized in that, The system includes: The acquisition module is used to acquire original images of the scrap steel to be cut from multiple different spatial perspectives; The scrap steel contour recognition module is used to input the original image into the scrap steel contour recognition model to obtain a binary image of the scrap steel contour corresponding to each spatial viewpoint; The bird's-eye view determination module is used to perform perspective transformation on multiple binary images of the scrap steel outline to obtain multiple bird's-eye view images. The planar cutting point determination module is used to determine the cutting point of the scrap steel to be cut based on scrap steel information in multiple bird's-eye view images; The spatial cutting point determination module is used to map the cutting point from the image coordinate system to the spatial cutting coordinates of the cutting device.

9. An electronic device, characterized in that, It includes 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 a computer program stored in the memory to implement the multi-view scrap steel cutting point determination method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, It stores a computer program that, when executed by the computer's processor, causes the computer to perform the multi-view scrap steel cutting point determination method as described in any one of claims 1 to 7.

Images