Electric arc three-dimensional reconstruction system and method based on binocular vision

Abstract

The invention discloses an electric arc three-dimensional reconstruction system and method based on binocular vision. The system comprises a double-camera image acquisition module, a correction module, a boundary contour point extraction module and a three-dimensional reconstruction module. The dual-camera image acquisition module completes dual-camera parameter calibration and dual-vision arc image shooting operation at different positions; the correction module corrects an optical distortion area and a breakpoint area of the double-vision arc image to obtain a corrected double-vision arc image; the boundary contour point extraction module adopts a third-order Bezier curve to perform segmentation fitting on the boundary contour of the corrected double-vision arc image to obtain continuous double-vision arc boundary points; and the three-dimensional reconstruction module reconstructs a three-dimensional contour line of the electric arc through a triangulation method according to the camera calibration parameters and the double-vision electric arc boundary points, and generates a corresponding three-dimensional point cloud data set. According to the method, the distortion and breakpoint phenomena of the arc image can be repaired, and high-precision reconstruction of three-dimensional data of the arc is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and more particularly to a binocular vision-based three-dimensional arc reconstruction system and method. Background Art

[0002] Arcs, a typical high-energy plasma discharge phenomenon, are widely present in high-voltage electricity transmission, as well as industrial manufacturing processes such as welding, cutting, and electrical discharge machining. Their geometric form, spatial structure, and evolutionary behavior are closely related to process parameters and are important bases for measuring process stability and energy distribution uniformity. However, due to the arc's inherent characteristics of high temperature, high luminous intensity, and rapid structural changes, traditional binocular 3D reconstruction methods are difficult to adapt to its complex imaging environment. Current industrial visual measurement relies on infrared or high-speed monocular cameras to capture arcs in two dimensions, and then indirectly estimates parameters such as arc length and extension based on empirical parameters. However, this method lacks true binocular 3D structural information, and the reconstruction accuracy and stability are poor, making it difficult to meet the needs of precision manufacturing and scientific research.

[0003] At the same time, arc images often experience significant image distortion and edge breakage due to strong light, thermal disturbances, and medium refraction, seriously affecting subsequent boundary recognition and spatial point cloud construction. Existing research has largely ignored this distortion factor, addressing it solely through empirical filtering or simple image enhancement methods. These methods lack systematic modeling and inverse correction mechanisms for optical path distortion.

[0004] Therefore, how to achieve high-precision reconstruction of arc three-dimensional point cloud data is a problem that technical personnel in this field urgently need to solve. Summary of the Invention

[0005] In view of this, the present invention provides a binocular vision-based arc three-dimensional reconstruction system and method, which can repair the distortion and breakpoint phenomena of the arc image and achieve high-precision reconstruction of the arc three-dimensional data.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] In a first aspect, the present invention provides an arc 3D reconstruction system based on binocular vision, comprising: a dual-camera image acquisition module, a correction module, a boundary contour point extraction module and a 3D reconstruction module;

[0008] The dual-camera image acquisition module is installed on a bracket and moves synchronously with the bracket to complete dual-camera parameter calibration and dual-vision arc image shooting operations at different positions;

[0009] The correction module is used to correct the optical distortion area and the breakpoint area of ​​the dual-vision arc image to obtain a corrected dual-vision arc image;

[0010] The boundary contour point extraction module is used to perform segmented fitting on the boundary contour of the corrected dual-vision arc image using a third-order Bezier curve to obtain continuous dual-vision arc boundary points;

[0011] The 3D reconstruction module is used to reconstruct the 3D contour line of the arc through a triangulation method according to the camera calibration parameters and the dual-vision arc boundary points, and generate a corresponding 3D point cloud data set.

[0012] Furthermore, the bracket is an aluminum profile bracket, and the dual-camera acquisition module is two cameras; the bracket is fixed on a slide rail, and the two cameras are installed on the bracket at a fixed angle, and are synchronously translated with the bracket on the slide rail. Every time an image acquisition position point is reached, the two cameras are calibrated once, and the arc is photographed after the camera is calibrated.

[0013] Furthermore, a pre-processing module is included, which is used to perform format conversion, sharpening enhancement, background image extraction and image binarization operations on the preliminarily captured dual-vision arc image.

[0014] Furthermore, the correction module includes an optical image correction submodule and a breakpoint repair submodule;

[0015] The optical image correction submodule is used to construct a BSDF light model containing camera calibration parameters, use the Blender light rendering method to simulate the image imaging path in a virtual rendering environment, and perform segmented light path correction on the optical distortion area caused by the propagation medium;

[0016] The breakpoint repair submodule is used to repair image breakpoints caused by optical refraction.

[0017] Furthermore, the optical image correction submodule includes a virtual environment construction unit, a medium distortion modeling and rendering unit, a correction parameter generation unit, and a distortion correction unit;

[0018] The virtual environment construction unit is used to create virtual dual cameras according to the camera calibration parameters of each image acquisition position point, place a standard calibration plate in the virtual scene, and use the Blender ray rendering method to generate an image of the standard calibration plate at each image acquisition position point without a medium in the virtual rendering scene as a reference image;

[0019] The medium distortion modeling and rendering unit is used to define relevant parameters of the medium model, insert the medium model between the standard calibration plate and the virtual dual cameras for any image acquisition position, and render the distorted image of the calibration plate when the medium is present at the image acquisition position;

[0020] The correction parameter generation unit is used to compare the reference image of the calibration plate at each image acquisition position point with the distorted image, extract the corner point difference, and establish the optical distortion correction parameters of the corresponding image acquisition position point;

[0021] The distortion correction unit is used to perform segmented geometric correction on the real-time dual-vision arc image acquired at the current image acquisition position point based on the optical distortion correction parameters of the current image acquisition position point.

[0022] Furthermore, the breakpoint repair submodule includes a binarization processing unit, a region marking unit, a judgment unit and a breakpoint filling unit;

[0023] The binarization processing unit is used to perform grayscale conversion and binarization processing on the dual-vision arc image;

[0024] The region marking unit is used to mark all independent regions in the binary image and screen out regions that meet the arc characteristics;

[0025] The judgment unit is used to check the continuity of pixels and boundary curvature continuity in the eight-neighborhood region of each pixel point in the arc area using the near point recognition and structural continuity judgment rules, and identify the area that does not meet the near point recognition and structural continuity judgment rules as a breakpoint area;

[0026] The breakpoint filling unit executes an image restoration algorithm based on pixel growth to automatically adhere and fill the breakpoint area.

[0027] Furthermore, the pixel growth-based image restoration algorithm is as follows: starting from the breakpoint position, the marked area attributes of each pixel in its eight neighborhoods are detected, and the neighborhood pixels with the same mark as the breakpoint position are used as the next connection point pixels, and it is necessary to ensure that the grayscale value of the next connection point pixel does not differ from the average grayscale of the current area by more than a threshold, and at the same time, the gradient direction and curvature of the next connection point pixel match the current area.

[0028] Furthermore, the three-dimensional reconstruction module reconstructs the three-dimensional contour line of the arc in the following manner: based on the rotation matrix after dual camera calibration, the arc boundary point photographed by the left camera is used as the origin, and the arc boundary point photographed by the right camera is multiplied by the rotation matrix to obtain the three-dimensional space coordinate point of the arc.

[0029] Furthermore, the system also includes a post-processing module, which is used to perform integration operations on the binocular three-dimensional point cloud data according to the pixel-physical coordinate mapping relationship to obtain the surface area and volume of the arc.

[0030] In a second aspect, the present invention provides a method for 3D reconstruction of electric arcs based on binocular vision, which is applicable to the system as described above, comprising the following steps:

[0031] Complete dual-camera parameter calibration and dual-vision arc image capture operations at different locations;

[0032] Correcting the optical distortion area and the breakpoint area of ​​the dual-vision arc image to obtain a corrected dual-vision arc image;

[0033] The boundary contour of the corrected dual-vision arc image is segmentedly fitted using a third-order Bezier curve to obtain continuous dual-vision arc boundary points.

[0034] According to the camera calibration parameters and the dual-vision arc boundary points, the three-dimensional contour of the arc is reconstructed through the triangulation method, and the corresponding three-dimensional point cloud dataset is generated.

[0035] It can be seen from the above technical solutions that compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. This invention acquires arc image data using dual cameras, combines camera calibration parameters with Bezier curve fitting to extract arc boundaries, and achieves binocular 3D point cloud reconstruction and physical parameter calculation of the arc. Simultaneously, by calibrating the dual cameras at different image acquisition locations, high-precision 3D reconstruction of arc data at each location is ensured. This system has a rational structure and high reconstruction accuracy, making it suitable for efficient binocular 3D visual reconstruction and surface parameter quantification of arc-like targets in bright dynamic scenes, providing strong technical support for arc process optimization control, quality assessment, and scientific research analysis.

[0037] 2. The optical image correction submodule of the present invention uses a BSDF-based ray rendering method to simulate the image distortion process. It also uses the corner deviation between the standard calibration image and the image affected by glass to establish an inverse correction model to address image deformation caused by organic glass interference. The image breakpoint repair submodule automatically identifies and repairs edge breaks caused by measured refraction, using pixel connectivity and region growing strategies to improve boundary continuity. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0039] Figure 1 This is a structural framework diagram of the arc 3D reconstruction system based on binocular vision provided by the present invention;

[0040] Figure 2 A schematic diagram of the dual-camera arrangement provided by the present invention;

[0041] Figure 3 The present invention provides a distorted image of a calibration plate captured when organic glass is present;

[0042] Figure 4 This is the corrected calibration plate image provided by the present invention. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0044] like Figure 1 As shown, the embodiment of the present invention discloses an arc 3D reconstruction system based on binocular vision, which is characterized by comprising: a dual-camera image acquisition module, a correction module, a boundary contour point extraction module and a 3D reconstruction module;

[0045] The dual-camera image acquisition module is installed on a bracket and moves synchronously with the bracket to complete dual-camera parameter calibration and dual-vision arc image capture operations at different positions;

[0046] The correction module is used to correct the optical distortion area and the breakpoint area of ​​the dual-vision arc image to obtain a corrected dual-vision arc image;

[0047] The boundary contour point extraction module is used to perform segmented fitting on the boundary contour of the corrected dual-vision arc image using a third-order Bezier curve to obtain continuous dual-vision arc boundary points;

[0048] The 3D reconstruction module is used to reconstruct the 3D contour of the arc through triangulation method according to the camera calibration parameters and the dual-vision arc boundary points, and generate the corresponding 3D point cloud dataset.

[0049] In one embodiment, the bracket is an aluminum profile bracket, and the dual-camera acquisition module is two cameras; the bracket is fixed on a slide rail, such as Figure 2 As shown, the two cameras are mounted on a bracket at a fixed angle and move synchronously with the bracket on the slide rail. Each time an image acquisition position is reached, the two cameras are calibrated and the arc is photographed after the cameras are calibrated.

[0050] When calibrating the camera, the traditional calibration method is used to obtain parameters such as camera intrinsic parameters, extrinsic parameters and rotation matrix.

[0051] By completing the capture and calibration operations at different positions, the binocular distance (baseline length) can be flexibly adjusted, and the baseline can be increased to enhance depth perception when detecting strong arcs. Expanding the spatial range covered by the calibration data allows for more accurate calculation of the camera's lens distortion parameters, avoiding the accumulation of edge errors caused by single-view calibration.

[0052] In one embodiment, the system further includes a preprocessing module, which is used to perform format conversion, sharpening enhancement, background image extraction and image binarization operations on the initially captured dual-vision arc image to improve the accuracy and stability of subsequent boundary recognition and reconstruction.

[0053] In a specific embodiment, the correction module includes an optical image correction submodule and a breakpoint repair submodule;

[0054] The optical image correction submodule is used to construct a BSDF light model containing camera calibration parameters, use the Blender light rendering method to simulate the image imaging path in a virtual rendering environment, and perform segmented light path correction on the optical distortion areas caused by the propagation medium;

[0055] The breakpoint repair submodule is used to repair image breakpoints caused by optical refraction.

[0056] Specifically, the optical image correction submodule includes a virtual environment construction unit, a medium distortion modeling and rendering unit, a correction parameter generation unit and a distortion correction unit.

[0057] The virtual environment construction unit is used to create virtual dual cameras according to the camera calibration parameters of each image acquisition position point, place a standard calibration plate in the virtual scene, and use the Blender ray rendering method to generate an image of the standard calibration plate without a medium at each image acquisition position point in the virtual rendering scene as a reference image.

[0058] The medium distortion modeling and rendering unit is used to define the relevant parameters of the medium model, such as the geometric dimensions (thickness, inclination) and optical properties (refractive index, surface roughness) of organic glass. For any image acquisition position, the medium model is inserted between the standard calibration plate and the virtual dual camera, and the calibration plate distorted image when the medium is present at the image acquisition position is rendered, such as Figure 3 Specifically, the optical path can be decomposed into three segments: the propagation path from the camera to the outer surface of the plexiglass, which follows the ideal straight line model; the propagation of light inside the plexiglass, which needs to consider the refractive index and surface scattering characteristics; and the propagation path from the plexiglass to the arc, which needs to compensate for the refraction effect of the medium boundary. The refractive index and scattering characteristics of each medium segment are modeled separately.

[0059] The correction parameter generation unit is used to compare the reference image of the calibration plate at each image acquisition position with the distorted image, extract the corner point difference, and establish the optical distortion correction parameters of the corresponding image acquisition position. Based on the optical distortion correction parameters, the distorted image of the calibration plate is corrected. Figure 4 shown.

[0060] In actual application, the distortion correction unit is used to call the optical distortion correction parameters corresponding to the current image acquisition position based on parameters such as the current position of the slide rail and the thickness of the organic glass, perform pixel-by-pixel reverse mapping on the real-time dual-vision arc image acquired at the current image position point, perform geometric correction in segments, and restore the original spatial imaging consistency.

[0061] Afterwards, adaptive histogram equalization can be performed on the geometrically corrected image to enhance the arc edge contrast.

[0062] The present invention decomposes the optical path and independently models each segment of the BSDF ray model to accurately quantify the nonlinear distortion caused by organic glass. At the same time, the high-precision ray tracing capabilities of Blender are used to generate benchmark data, solving the problem of obtaining an ideal image without media in a real environment. Furthermore, the virtual calibration plate can flexibly cover the entire workspace, avoiding undercorrection of edge areas caused by the limited placement of the physical calibration plate.

[0063] In a specific embodiment, the breakpoint repair submodule includes a binarization processing unit, a region marking unit, a judgment unit, and a breakpoint filling unit;

[0064] The binarization processing unit is used to perform grayscale conversion and binarization processing on the dual-vision arc image; for example, contrast-limited adaptive histogram equalization (CLAHE) preprocessing is used to eliminate local overexposure / underexposure caused by uneven arc light intensity.

[0065] The region marking unit is used to mark all independent regions in the binary image and filter out regions that meet the arc characteristics; for example, morphological filtering is used to remove isolated point noise, and connected domain filtering is used to retain regions that meet certain conditions in area, aspect ratio, and circularity as arc regions.

[0066] The judgment unit is used to check the continuity of pixels and boundary curvature continuity within the eight-neighborhood of each pixel point in the arc area using the near point recognition and structural continuity judgment rules, and identify the area that does not meet the near point recognition and structural continuity judgment rules as a breakpoint area. For example, if the number of valid pixels in the eight-neighborhood is less than or equal to 1, the point is determined to be a breakpoint. The curvature is calculated point by point along the arc boundary. If the curvature of several consecutive points exceeds a threshold, it is marked as a structural fracture area. The gradient direction histogram of the areas on both sides of the breakpoint is calculated. If the direction difference is greater than 30°, it is determined to be structural discontinuity.

[0067] The breakpoint filling unit automatically glues and fills the breakpoint area using a pixel growth-based image restoration algorithm. Specifically, the pixel growth-based image restoration algorithm starts from the breakpoint position and detects the marked area attributes of each pixel in its eight-neighborhood. The neighboring pixels with the same mark as the breakpoint position are used as the next connection point pixels. The grayscale value of the next connection point pixel must not differ from the average grayscale of the current area by more than a threshold. At the same time, the gradient direction and curvature of the next connection point pixel must match the current area to enhance the integrity of the image boundary. This ensures that the restoration result conforms to the actual shape of the arc and provides stable initial image data for subsequent binocular 3D reconstruction.

[0068] In one embodiment, the three-dimensional reconstruction module reconstructs the three-dimensional contour line of the arc by matching the left and right cameras point by point, taking the arc boundary point photographed by the left camera as the origin, and multiplying the arc boundary point photographed by the right camera by the marked rotation matrix to obtain the three-dimensional spatial coordinate point of the arc, and then mapping the two-dimensional boundary into a binocular three-dimensional point cloud to restore the arc spatial structure.

[0069] More advantageously, the system also includes a post-processing module, which is used to perform integration operations on the binocular 3D point cloud data according to the pixel-physical coordinate mapping relationship to obtain the surface area and volume of the arc. Specifically, based on the obtained 3D point cloud data, the area corresponding to each cross section is layered and integrated along the Z axis to calculate the overall surface area and volume of the arc, thereby achieving structural quantification and feature parameter extraction.

[0070] The arc surface area calculation formula is:

[0071]

[0072] Among them, A b is the arc surface area, N represents the number of layers along the Z axis, D j represents the cross-sectional perimeter of the jth layer, and ΔD represents the physical thickness of each layer;

[0073] The arc volume calculation formula is:

[0074]

[0075] Among them, Vb represents the volume of the arc, A b is the arc surface area, N represents the number of layers along the Z axis, D j represents the cross-sectional diameter of the jth layer, ΔD represents the physical thickness of each layer, Dt represents the pixel diameter of the tth layer, and Dpixel represents the conversion factor from pixels to physical units.

[0076] In other embodiments, the present invention further provides a binocular vision-based three-dimensional reconstruction method for electric arcs, which is applicable to the above system and includes the following steps:

[0077] Complete dual-camera parameter calibration and dual-vision arc image capture operations at different locations;

[0078] Correcting the optical distortion area and the breakpoint area of ​​the dual-vision arc image to obtain a corrected dual-vision arc image;

[0079] The boundary contour of the corrected dual-vision arc image is segmentedly fitted using a third-order Bezier curve to obtain continuous dual-vision arc boundary points.

[0080] According to the camera calibration parameters and the dual-vision arc boundary points, the three-dimensional contour of the arc is reconstructed through the triangulation method, and the corresponding three-dimensional point cloud dataset is generated.

[0081] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Reference can be made to the common and similar parts between the various embodiments. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple, and the relevant parts can be referred to the method description.

[0082] The above description of the disclosed embodiments is intended to enable one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be readily apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention is not limited to the embodiments shown herein but is intended to conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An arc 3D reconstruction system based on binocular vision, characterized in that: include: Dual-camera image acquisition module, correction module, boundary contour point extraction module and 3D reconstruction module; The dual-camera image acquisition module is installed on a bracket and moves synchronously with the bracket to complete dual-camera parameter calibration and dual-vision arc image shooting operations at different positions; The correction module is used to correct the optical distortion area and the breakpoint area of ​​the dual-vision arc image to obtain a corrected dual-vision arc image; The boundary contour point extraction module is used to perform segmented fitting on the boundary contour of the corrected dual-vision arc image using a third-order Bezier curve to obtain continuous dual-vision arc boundary points; The 3D reconstruction module is used to reconstruct the 3D contour line of the arc through a triangulation method according to the camera calibration parameters and the dual-vision arc boundary points, and generate a corresponding 3D point cloud data set.

2. The arc 3D reconstruction system based on binocular vision according to claim 1, characterized in that: The bracket is an aluminum profile bracket, and the dual-camera acquisition module is two cameras; the bracket is fixed on a slide rail, and the two cameras are installed on the bracket at a fixed angle and move synchronously with the bracket on the slide rail. Every time an image acquisition position point is reached, the two cameras are calibrated, and the arc is photographed after the camera is calibrated.

3. The arc 3D reconstruction system based on binocular vision according to claim 1, characterized in that: The system also includes a pre-processing module, which is used to perform format conversion, sharpening enhancement, background image extraction and image binarization operations on the preliminarily captured dual-vision arc image.

4. The arc 3D reconstruction system based on binocular vision according to claim 1, characterized in that: The correction module includes an optical image correction submodule and a breakpoint repair submodule; The optical image correction submodule is used to construct a BSDF light model containing camera calibration parameters, use the Blender light rendering method to simulate the image imaging path in a virtual rendering environment, and perform segmented light path correction on the optical distortion area caused by the propagation medium; The breakpoint repair submodule is used to repair image breakpoints caused by optical refraction.

5. The arc 3D reconstruction system based on binocular vision according to claim 4, characterized in that: The optical image correction submodule includes a virtual environment construction unit, a medium distortion modeling and rendering unit, a correction parameter generation unit and a distortion correction unit; The virtual environment construction unit is used to create virtual dual cameras according to the camera calibration parameters of each image acquisition position point, place a standard calibration plate in the virtual scene, and use the Blender ray rendering method to generate an image of the standard calibration plate at each image acquisition position point without a medium in the virtual rendering scene as a reference image; The medium distortion modeling and rendering unit is used to define relevant parameters of the medium model, insert the medium model between the standard calibration plate and the virtual dual cameras for any image acquisition position, and render the distorted image of the calibration plate when the medium is present at the image acquisition position; The correction parameter generation unit is used to compare the reference image of the calibration plate at each image acquisition position point with the distorted image, extract the corner point difference, and establish the optical distortion correction parameters of the corresponding image acquisition position point; The distortion correction unit is used to perform segmented geometric correction on the real-time dual-vision arc image acquired at the current image acquisition position point based on the optical distortion correction parameters of the current image acquisition position point.

6. The arc 3D reconstruction system based on binocular vision according to claim 4, characterized in that: The breakpoint repair submodule includes a binarization processing unit, a region marking unit, a judgment unit and a breakpoint filling unit; The binarization processing unit is used to perform grayscale conversion and binarization processing on the dual-vision arc image; The region marking unit is used to mark all independent regions in the binary image and screen out regions that meet the arc characteristics; The judgment unit is used to check the continuity of pixels and boundary curvature continuity in the eight-neighborhood region of each pixel point in the arc area using the near point recognition and structural continuity judgment rules, and identify the area that does not meet the near point recognition and structural continuity judgment rules as a breakpoint area; The breakpoint filling unit executes an image restoration algorithm based on pixel growth to automatically adhere and fill the breakpoint area.

7. The arc 3D reconstruction system based on binocular vision according to claim 6, characterized in that: The pixel growth-based image restoration algorithm is as follows: starting from the breakpoint position, the marked area attributes of each pixel in its eight neighborhoods are detected, and the neighboring pixels with the same mark as the breakpoint position are used as the next connection point pixels. It is necessary to ensure that the grayscale value of the next connection point pixel does not exceed the average grayscale difference of the current area by more than a threshold, and at the same time, the gradient direction and curvature of the next connection point pixel match those of the current area.

8. The arc 3D reconstruction system based on binocular vision according to claim 1, characterized in that: The 3D reconstruction module reconstructs the 3D contour line of the arc by: according to the rotation matrix after dual camera calibration, taking the arc boundary point photographed by the left camera as the origin, multiplying the arc boundary point photographed by the right camera by the rotation matrix to obtain the 3D spatial coordinate point of the arc.

9. The arc 3D reconstruction system based on binocular vision according to claim 1, characterized in that: The system also includes a post-processing module, which is used to perform integration operations on binocular three-dimensional point cloud data according to a pixel-physical coordinate mapping relationship to obtain the surface area and volume of the arc.

10. A three-dimensional reconstruction method of arc based on binocular vision, characterized in that: The system is applicable to any one of claims 1 to 9, comprising the following steps: Complete dual-camera parameter calibration and dual-vision arc image capture operations at different locations; Correcting the optical distortion area and the breakpoint area of ​​the dual-vision arc image to obtain a corrected dual-vision arc image; The boundary contour of the corrected dual-vision arc image is segmentedly fitted using a third-order Bezier curve to obtain continuous dual-vision arc boundary points. According to the camera calibration parameters and the dual-vision arc boundary points, the three-dimensional contour of the arc is reconstructed through the triangulation method, and the corresponding three-dimensional point cloud dataset is generated.

Images