An arc length vision detection method, system and device for electric arc additive manufacturing

Abstract

The application discloses an arc length visual detection method, system and device for electric arc additive manufacturing, and relates to the technical field of additive manufacturing detection. The method comprises the following steps: acquiring an original electric arc image; preprocessing the original electric arc image to generate an edge information image; determining a width value at an electric arc peak value in a first dashed line frame in the edge information image; when the width value at the electric arc peak value is greater than or equal to a set threshold value, determining that the edge information image is in an electric arc peak value stage, and deleting the edge information image data in the electric arc peak value stage; and when the width value at the electric arc peak value is less than the set threshold value, determining that the edge information image is in an electric arc base value stage, and generating an arc length change curve according to the edge information image in the electric arc base value stage, thereby solving the problem of difficulty in electric arc arc length detection in electric arc additive manufacturing based on non-melting electrode argon welding.

Description

Technical Field

[0001] This invention relates to the field of additive manufacturing inspection technology, and in particular to a visual inspection method, system and device for arc length in electric arc additive manufacturing. Background Technology

[0002] Additive manufacturing technology can be broadly categorized into metal additive manufacturing and non-metal additive manufacturing. Metal additive manufacturing can be further classified based on the type of raw material and the energy source. According to the type of raw material, it can be divided into powder-lay additive manufacturing, powder-feed additive manufacturing, and wire-feed additive manufacturing. According to the energy source, it can be divided into three types: laser, electron beam, and electric arc. Wire and Arc Additive Manufacturing (WAAM) uses a layer-by-layer welding method to create dense metal solid components. Because it uses an electric arc as the energy beam, it has high heat input and fast forming speed, making it suitable for low-cost, high-efficiency, and rapid near-net-shape forming of large-sized complex components. This technology is mainly based on welding techniques such as Tungsten Inert Gas (TIG), Metal Inert Gas (MIG), and Submerged Arc Welding (SAW). The formed parts are composed entirely of welds, with uniform chemical composition and high density. The open forming environment has no restrictions on the size of the formed parts, and the forming rate can reach several kg / h. However, the surface of WAAM parts fluctuates greatly, the surface quality of the formed parts is low, and the forming accuracy is not particularly high.

[0003] The basic hardware structure of existing arc additive manufacturing should include a forming heat source, a wire feeding system, and a motion actuator. As a motion actuator extending from a point to three dimensions, its displacement and velocity, repeatability, and motion stability are crucial to the dimensional accuracy of the formed part. Currently, commonly used forming systems include TIG + CNC machine tool / worktable, TIG + robot, MIG + CNC machine tool, MIG / Metal Active Gas Arc Welding (MAG) + robot, and Plasma Arc Welding (PAW) + CNC machine tool. Because the energy beam in arc additive manufacturing has characteristics such as low heat flux density, large heating radius, and high heat source intensity, the instantaneous reciprocating heat source interacts strongly with the forming environment during the forming process. The environmental variables caused by heat accumulation are more significant, so the forming accuracy is often lower compared to other heat sources, making it difficult to form some special structures of the part.

[0004] According to the sensing method, electric arc additive manufacturing (WAAM) inspection technology can be divided into digital imaging, thermal sensing, laser, structured light, and electrical signal detection. Digital imaging uses charge-coupled device (CCD) industrial cameras or complementary metal-oxide-semiconductor (CMOS) industrial cameras to acquire images; the signals are all digital. The acquired images are transmitted back to the main control equipment (usually a computer) for monitoring the WAAM status. Thermal sensing constructs a temperature field by collecting temperature information during the WAAM process. The molten pool in the WAAM process is composed of liquid metal with extremely high temperatures, but the surrounding temperature is low, resulting in a rapid cooling rate. The continuous high temperature can cause a secondary heat treatment effect on the formed metal. Thermal sensing technology can observe which parts have greater thermal effects and make corrections in advance. Laser sensing utilizes the high speed of light to emit and receive the returned light to calculate distance and acquire information. The characteristic of laser sensing is that it... One method is active light, where lasers are fast, have strong anti-interference capabilities, and offer higher detection accuracy. Structured light uses one or more cameras combined with algorithms to obtain depth information from two-dimensional images, constructing a three-dimensional model. Structured light technology can obtain real-time three-dimensional data of WAAM forming, but the data volume is large and processing is difficult (structured light applications include facial recognition). Electrical signals directly acquire voltage and current from the equipment. Signals acquired through this method are all analog signals, offering high accuracy and speed. However, there is a conversion between electrical signals and process parameters in the WAAM process, and the accuracy of this conversion is often difficult to guarantee. Therefore, the correlation between electrical signals and the actual WAAM process is low, making it difficult to obtain accurate data. All information acquired through these methods is related to the WAAM process. Modifications to process parameters during WAAM will affect the data obtained by these sensing methods. This information can help improve the quality of WAAM forming. Existing arc length detection technologies in electric arc additive manufacturing mostly focus on the MIG method, while this method is lacking in the TIG field. Summary of the Invention

[0005] The purpose of this invention is to provide a visual inspection method, system, and device for arc length in arc additive manufacturing, which solves the problem of difficult arc length detection in arc additive manufacturing based on non-consumable electrode argon welding.

[0006] To achieve the above objectives, the present invention provides the following solution:

[0007] This invention provides a method for visually inspecting the arc length in electric arc additive manufacturing, comprising:

[0008] Acquire the raw arc image;

[0009] The original electric arc image is preprocessed to generate an edge information image;

[0010] Determine the width value of the arc peak within the first dashed box in the edge information image; the first dashed box is used to detect the arc width region of the edge information image;

[0011] When the width value at the peak of the electric arc is greater than or equal to a set threshold, the edge information image is determined to be in the peak stage of the electric arc, and the edge information image data in the peak stage of the electric arc is deleted.

[0012] When the width value at the peak of the electric arc is less than a set threshold, the edge information image is determined to be in the electric arc base value stage, and an arc length variation curve is generated based on the edge information image in the electric arc base value stage.

[0013] Optionally, an arc length variation curve is generated based on the edge information image at the arc base value stage, specifically including:

[0014] Determine the arc length value within the second dashed box in the target image; the second dashed box is used to detect the arc length region of the target image; the target image is the edge information image in the arc base value stage;

[0015] The arc length value is subjected to abrupt data suppression processing to generate an arc length variation curve.

[0016] Optionally, determining the arc length value within the second dashed box in the target image specifically includes:

[0017] The target image is detected using the second dashed bounding box to determine the target point's ordinate value; the target point's ordinate value is the ordinate value of the non-zero pixel value retrieved within the second dashed bounding box.

[0018] An array is formed by combining the target point's ordinate value and the arc calibration value; the array includes two elements; each element includes the target point's ordinate value and the arc calibration value; the arc calibration value is the ordinate value of the starting point of the arc within the second dashed box;

[0019] The data in the element is subtracted to obtain the difference value, and the difference value is multiplied by the pixel conversion coefficient to obtain the arc length value.

[0020] Optionally, the original arc image is preprocessed, specifically including:

[0021] The original electric arc image is subjected to color conversion processing to generate a grayscale image;

[0022] The grayscale image is subjected to smoothing and noise suppression processing to generate a smoothed and noise-suppressed image;

[0023] Edge detection and pixel binarization are performed on the smoothed noise-suppressed image to obtain an edge information image.

[0024] Optionally, the original arc image is subjected to color conversion processing to generate a grayscale image, specifically including:

[0025] The original electric arc image was processed by color conversion using the color conversion method in the third-party computer vision library OpenCV to generate a grayscale image.

[0026] Optionally, the grayscale image is subjected to smoothing and noise suppression processing to generate a smoothed and noise-suppressed image, specifically including:

[0027] A Gaussian function filter with a window size of 5×5 is used to perform discrete window sliding convolution processing on the grayscale image to generate a smooth noise-suppressed image.

[0028] Optionally, edge detection and pixel binarization are performed on the smoothed noise-suppressed image to obtain an edge information image, specifically including:

[0029] Gradient calculation is performed on the smoothed noise-suppressed image to obtain the gradient calculation result;

[0030] The gradient calculation results are subjected to non-maximum suppression processing, and the data after non-maximum suppression processing is subjected to double threshold edge selection to generate edge selection values;

[0031] Based on the set high and low thresholds, determine whether the edge selection value is a boundary value;

[0032] If the edge selection value is less than the low threshold, then the edge selection value is not a boundary value; if the edge selection value is greater than the high threshold, then the edge selection value is a strong boundary value.

[0033] Based on the determination result of the boundary value, an edge information image is obtained.

[0034] The present invention also provides a visual inspection system for arc length in electric arc additive manufacturing, comprising:

[0035] The image acquisition module is used to acquire the original electric arc image;

[0036] The image preprocessing module is used to preprocess the original electric arc image to generate an edge information image;

[0037] The arc peak width determination module is used to determine the width value of the arc peak within the first dashed box in the edge information image; the first dashed box is used to detect the arc width region of the edge information image.

[0038] An arc peak stage determination module is used to determine that the edge information image is in the arc peak stage when the width value at the arc peak is greater than or equal to a set threshold, and to delete the edge information image data in the arc peak stage.

[0039] The arc base value stage determination module is used to determine that the edge information image is in the arc base value stage when the width value at the peak of the arc is less than a set threshold, and to generate an arc length variation curve based on the edge information image in the arc base value stage.

[0040] The present invention also provides a visual inspection device for arc length in electric arc additive manufacturing, comprising: an industrial CCD camera, a three-axis CNC machine tool, and a control computer;

[0041] The industrial CCD camera is located on the support of the three-axis CNC machine tool and is used to acquire raw electric arc images;

[0042] The three-axis CNC machine tool is used for three-way arc additive manufacturing.

[0043] The control computer is connected to the industrial CCD camera and is used to execute the above-described arc length visual inspection method for electric arc additive manufacturing.

[0044] According to specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0045] This invention determines whether the edge information image is in the arc peak stage or the arc base stage by detecting the width value at the arc peak point within the first dashed box. If the edge information image is in the arc peak stage, the data corresponding to that stage is discarded. If the edge information image is in the arc base stage, the data corresponding to that stage needs to be extracted, and the arc length value of the data in the arc base stage is detected using the second dashed box to obtain the arc length data. In addition, abrupt data suppression processing is required for the arc length data to prevent detection errors. Finally, the arc length variation curve is generated by a computer-defined program, thereby realizing the detection of arc length in arc additive manufacturing based on non-consumable electrode argon welding. Attached Figure Description

[0046] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1This is a structural connection diagram of an arc length visual inspection device for electric arc additive manufacturing provided in an embodiment of the present invention;

[0048] Figure 2 A schematic flowchart of a visual inspection method for arc length in electric arc additive manufacturing provided by an embodiment of the present invention;

[0049] Figure 3 The first and second dashed frame diagrams are provided in an embodiment of the present invention for a visual inspection method of arc length in electric arc additive manufacturing.

[0050] Figure 4 This is a schematic diagram of the structure of an arc length visual inspection system for electric arc additive manufacturing provided in an embodiment of the present invention.

[0051] Symbol explanation:

[0052] First dashed box - 1, Second dashed box - 2, Industrial CCD camera - 3, Three-axis CNC machine tool - 4, Control computer - 5, Welding machine - 6, Hot wire machine - 7, Gas cylinder - 8, Welding torch - 9, Wire feeder - 10, Wire material - 11. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] The purpose of this invention is to provide a visual inspection method, system, and device for arc length in arc additive manufacturing, which solves the problem of difficult arc length detection in arc additive manufacturing based on non-consumable electrode argon welding.

[0055] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0056] This invention uses a TIG (Transfer-Injection) + CNC machine tool + digital image vision method to detect the arc length in the WAAM (Wastewater Additive Manufacturing) process. The arc length during the WAAM process is detected using machine vision, and the manufacturing process parameters selected are a 70A current for thin-wall forming. This invention can effectively detect the arc length of single-layer thin-walled parts. Existing arc length detection technologies mostly focus on MIG (Metal Injection) methods, while TIG methods are lacking in research. Generally, arc additive manufacturing, due to its high manufacturing efficiency and large molten pool size, suffers from poor forming accuracy. The rapid changes in the molten droplet transition during manufacturing lead to defects such as misalignment and surface unevenness in the final formed structure, making it impossible to complete a stable structural part manufacturing process at the centimeter scale. The arc length visual detection method of the present invention uses computer vision methods, data acquisition, digital image filtering processing, and analysis of information in the image, including classical image processing methods. This method is designed specifically for arc light interference in the special scenario, achieving the purpose of arc length visual detection in arc additive manufacturing. Facing the characteristics of strong arc light changes and molten pool changes in the TIG method of WAAM, a strategy is designed to remove arc light interference, enabling effective real-time arc length detection in the stable stage of thin-walled single-pass manufacturing.

[0057] In the process of arc additive manufacturing, because the arc is generated by breaking down the air, and the wire current interferes with the molten pool, the connection process between the molten metal droplet and the molten pool is random and not in a stable state. Every moment is unreproducible. When a deviation occurs at a certain moment without the operator's knowledge, this defect will accumulate over time and eventually lead to collapse, uneven surface, or failure to form. The changes in the molten pool and the generation of small defects are related to the arc length in the arc additive manufacturing process. When the arc length is long, the transition of the molten droplet will have a free fall stage, and it cannot connect well with the molten pool due to the influence of gravity. When the arc length is short, the molten droplet cannot contact the molten pool due to mechanical structure and cannot participate in the connection. Both of these situations will greatly affect the stability and forming quality of WAAM manufacturing.

[0058] In summary, arc length, as a key data point, influences the entire WAAM process. The most effective method for obtaining arc length data is through visual observation (the electric arc light is too intense for the human eye to see directly; it needs to be captured by a camera, processed, and displayed). Therefore, this invention is based on... Figure 1 The visual inspection device built with an industrial CCD camera 3 shown uses computer vision methods to perform filtering, gradient calculation, color conversion, and feature extraction, effectively obtaining data on the key variable arc length in the WAAM process. This data serves as a basis to guide the operator in indirectly controlling the process, reducing defects and improving manufacturing quality.

[0059] Example 1

[0060] This embodiment discloses a visual inspection method for arc length in electric arc additive manufacturing, such as... Figure 1 and Figure 2 As shown, this method uses an industrial CCD camera 3 to take real-time photos during manufacturing and transmits the acquired raw arc digital images back to the computer. The computer processes the images according to a pre-set program to extract the arc length data. At the start of the experiment, the camera needs to be turned on for data calibration and the calibration data needs to be written. All experimental equipment needs to be turned on, and a single-layer reciprocating melt channel with a length of 100mm needs to be set. During the WAAM manufacturing process, the raw images will be transmitted back to the computer in the control cabinet. The computer communicates with the camera via the GigE network protocol to obtain the raw images and trigger the image processing program. The specific method includes the following steps:

[0061] Step 101: Obtain the original arc image; the original arc image is: an image actually captured by the industrial CCD camera 3 during the arc additive manufacturing process.

[0062] Step 102: Preprocess the original electric arc image to generate an edge information image.

[0063] Step 103: Determine the width value of the arc peak within the first dashed box 1 in the edge information image; the first dashed box 1 is used to detect the arc width region of the edge information image.

[0064] Step 104: When the width value at the peak of the electric arc is greater than or equal to a set threshold, it is determined that the edge information image is in the peak stage of the electric arc, and the edge information image data in the peak stage of the electric arc is deleted.

[0065] Step 105: When the width value at the peak of the electric arc is less than a set threshold, the edge information image is determined to be in the electric arc base value stage, and an arc length variation curve is generated based on the edge information image in the electric arc base value stage.

[0066] Step 102 specifically includes: performing color conversion processing on the original electric arc image to generate a grayscale image; performing smoothing and noise suppression processing on the grayscale image to generate a smoothed and noise-suppressed image; and performing edge detection and pixel binarization processing on the smoothed and noise-suppressed image to obtain an edge information image.

[0067] The purpose of grayscale processing is to simplify the color information of the image, making it easier to perform image gradient calculations later. There are various methods for grayscale processing. The simplest method is to directly average the three channels of the RGB (red, green, blue) image to obtain a grayscale image. More complex methods include weighting the RGB three colors according to the photosensitivity of the human retina. This method uses the color conversion method in the third-party computer vision library OpenCV to perform color conversion.

[0068] The purpose of smoothing noise suppression is to remove outlier pixel values ​​from an image, making image gradient changes smoother and preventing the appearance of boundaries near noise points in subsequent gradient calculations. Smoothing noise suppression methods use a Gaussian function as a filter, which is a linear smoothing filter. A 5×5 region is selected as the filter window, and a discrete window sliding convolution operation is performed on the entire image to generate the processed image. Furthermore, noise in an image is similar to noise in music; it can be attributed to interference signals outside of useful features or the main signal. In an image, this manifests as interfering pixels outside of useful features. These pixels are outliers, useless, and their surrounding image gradients change drastically.

[0069] The edge gradient detection method consists of four steps: filtering, image gradient calculation, non-maximum suppression (NMS) on the gradient results, and double-threshold edge selection. The third step retains only the most probable edge information at a specific point, where the gradient maxima around that point are considered the most likely edges. The fourth step uses high and low thresholds; values ​​below the low threshold are considered not boundaries, while values ​​above the high threshold are considered strong boundaries. Weak boundaries are determined by their connection to strong boundaries. The image gradient can be understood mathematically as the direction of the gradient, where the gradient reaches its maximum value (the magnitude of the directional derivative). The image gradient represents the direction of the fastest change in pixel value around a given pixel. There are eight directions around a pixel, and the direction with the largest gradient indicates the edge information in the image, i.e., the image feature boundary.

[0070] Step 105 specifically includes: such as Figure 3As shown, firstly, the arc width is determined based on the first dashed box 1 as the retrieval area. If the arc width reaches a threshold, it is considered the peak arc stage, and the data is discarded (occupying approximately 5-10% of the time in one arc pulse cycle). The arc base value stage is used for data extraction. If the first step determines it to be the base value stage, the target image is detected using the second dashed box 2 to determine the target point's ordinate value. The ordinate value of the target point is the ordinate value of the non-zero pixel value retrieved within the second dashed box 2 (at this time, the binarized image only contains 0 and 255). The ordinate value of the target point is combined with the arc calibration value to form an array. The array includes two elements: the ordinate value of the target point and the arc calibration value. The arc calibration value is the ordinate value of the starting point of the arc within the second dashed box. The data in the element is subtracted to obtain the difference value, and the difference value is multiplied by the pixel conversion coefficient to obtain the arc length value.

[0071] The arc base stage and arc peak stage are inevitable processes in gas tungsten inert gas welding (GTAW) with pulse mode in arc additive manufacturing. It is a pulsed arc form with two positions: the high position is the arc peak, and the low position is the arc base. Because the overall brightness of the image is too high during the arc peak stage, the arc and the molten pool are indistinguishable. Therefore, the top of the molten pool can only be seen under the low brightness of the arc base stage, which is used for arc length measurement. Arc length cannot be detected during the arc peak stage.

[0072] Furthermore, in actual operation, the industrial CCD camera 3 needs to be turned on for data calibration. After writing the calibration data, all experimental equipment is turned on, and the experiment is set as a single-layer reciprocating melting channel experiment with a length of 100mm. The data calibration is implemented through software. In actual operation, the arc calibration value is the ordinate value of the arc's starting point. Since the position of the arc's starting point in the image is constant, it can be calibrated in one go. At the same time, a ruler is placed at the same distance from the industrial CCD camera 3 and the arc. The number of pixels in the image is determined based on the 10mm length of the ruler. 10mm divided by the corresponding number of pixels yields the pixel conversion coefficient.

[0073] Furthermore, the WAAM process is continuous (based on actual physical processes), making abrupt changes in arc length impossible. Therefore, the final step involves limiting and suppressing these abrupt changes to prevent detection errors. Data is invalidated and deleted when the arc length increment exceeds 25%, allowing for detection at the next time step. After the experiment concludes, an Excel spreadsheet displaying the arc length change curve is automatically exported according to the predefined program. The spreadsheet generation utilizes the QtAxContainer component (a dedicated table processing tool within Qt), a third-party library of the Qt cross-platform development framework. At the start of the experiment, an existing Excel spreadsheet is pre-opened using this tool. Each time an arc length is calculated, the storage function is activated. The software writes the arc length data by accessing a specific cell address in the spreadsheet, incrementing the cell number after each write. The data writing module is only activated after the arc length is calculated, activating storage once per arc length. At the end of the experiment, the detection module is closed, and the spreadsheet export is automatically triggered, directly exporting the Excel file containing the data.

[0074] Example 2

[0075] This embodiment discloses a visual inspection system for arc length in electric arc additive manufacturing, such as... Figure 4 As shown, the system includes the following modules:

[0076] Image acquisition module 201 is used to acquire raw electric arc images.

[0077] The image preprocessing module 202 is used to preprocess the original electric arc image to generate an edge information image.

[0078] The arc peak width determination module 203 is used to determine the width value of the arc peak within the first dashed box in the edge information image; the first dashed box is used to detect the arc width region of the edge information image.

[0079] The arc peak stage determination module 204 is used to determine that the edge information image is in the arc peak stage when the width value at the arc peak is greater than or equal to a set threshold, and to delete the edge information image data in the arc peak stage.

[0080] The arc base value stage determination module 205 is used to determine that the edge information image is in the arc base value stage when the width value at the peak of the arc is less than a set threshold, and to generate an arc length variation curve based on the edge information image in the arc base value stage.

[0081] Example 3

[0082] This embodiment discloses a visual inspection device for arc length in electric arc additive manufacturing, such as... Figure 1As shown, the device includes: an industrial CCD camera 3, a three-axis CNC machine tool 4, a control computer 5, a welding machine 6, a hot wire machine 7, a gas cylinder 8, a welding torch 9, a wire feeder 10, and wire 11.

[0083] An industrial CCD camera 3, located on the support of the three-axis CNC machine tool 4, is used to acquire the original arc image; the three-axis CNC machine tool 4 is used for three-dimensional arc additive manufacturing; a control computer 5, connected to the industrial CCD camera 3, is used to execute the arc length visual detection method for arc additive manufacturing described in Embodiment 1; the welding torch 9 is connected to the welding machine 6 and the gas cylinder 8 respectively; the wire feeder 10 is connected to the hot wire machine 7 and the wire 11 respectively; the wire feeder 10 feeds the wire laterally, and the wire feeder 10 is fixedly connected to the welding torch 9 through a support.

[0084] Arc additive manufacturing is based on non-consumable electrode argon welding technology. While using an electric arc as the heat input, a hot wire power supply is added as a preheating power source for the wire. The hot wire power supply works by using resistance heating. The entire device is mounted on a three-axis CNC machine tool, with the spindle replaced by a welding torch. The CNC control system controls the wire feeder and the machine tool spindle, while the hot wire power supply is independently controlled. The hot wire power supply is manually switched on and off before the experiment and operates continuously during deposition. In the experiment, the wire is fed laterally, with the wire feed nozzle fixed to the welding torch, exhibiting no relative movement during deposition. In different scanning paths, the direction of the wire feed speed and the angle between the wire feed speed and the tungsten electrode remain constant, but the directions of the wire feed speed and the scanning speed change.

[0085] As a technical method under WAAM, TIG currently has no other effective alternative. Based on the technical solution of this invention, different sensors such as laser and structured light can be used to replace the original data acquisition method, but the process remains the same. Different levels of processing are performed according to the type of original data (digital image processing for data image signals, and point cloud data processing for laser and structured light).

[0086] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the systems disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the descriptions are relatively simple; relevant parts can be referred to the method section.

[0087] This document uses specific examples to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. Furthermore, those skilled in the art will recognize that, based on the ideas of the present invention, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for visually inspecting the arc length in electric arc additive manufacturing, characterized in that, include: Acquire the raw arc image; The original electric arc image was acquired by an industrial CCD camera located on a three-axis CNC machine tool support. The original electric arc image is preprocessed to generate an edge information image, specifically including: using the color conversion method in the third-party computer vision library OpenCV to perform color conversion processing on the original electric arc image to generate a grayscale image; performing smoothing and noise suppression processing on the grayscale image to generate a smoothed and noise-suppressed image; and performing edge detection and pixel binarization processing on the smoothed and noise-suppressed image to obtain the edge information image. Determine the width value of the arc peak within the first dashed box in the edge information image; the first dashed box is used to detect the arc width region of the edge information image; When the width value at the peak of the electric arc is greater than or equal to a set threshold, the edge information image is determined to be in the peak stage of the electric arc, and the edge information image data in the peak stage of the electric arc is deleted. When the width value at the peak of the electric arc is less than a set threshold, the edge information image is determined to be in the arc base value stage, and an arc length variation curve is generated based on the edge information image in the arc base value stage, specifically including: The target image is detected using a second dashed bounding box to determine the ordinate value of the target point; the second dashed bounding box is used to detect the arc length region of the target image; the target image is the edge information image in the arc base value stage; the ordinate value of the target point is the ordinate value of the non-zero pixel value retrieved within the second dashed bounding box; An array is formed by combining the target point's ordinate value and the arc calibration value; the array includes two elements; each element includes the target point's ordinate value and the arc calibration value; the arc calibration value is the ordinate value of the starting point of the arc within the second dashed box; The data in the element is subtracted to obtain the difference value, and the difference value is multiplied by the pixel conversion coefficient to obtain the arc length value; The arc length value is subjected to abrupt data suppression processing to generate an arc length variation curve; the abrupt data suppression method is amplitude limiting.

2. The method for visual inspection of arc length in electric arc additive manufacturing according to claim 1, characterized in that, The grayscale image is subjected to smoothing and noise suppression processing to generate a smoothed and noise-suppressed image, specifically including: A Gaussian function filter with a window size of 5×5 is used to perform discrete window sliding convolution processing on the grayscale image to generate a smooth noise-suppressed image.

3. The method for visual inspection of arc length in electric arc additive manufacturing according to claim 1, characterized in that, The smoothed noise-suppressed image is subjected to edge detection and pixel binarization processing to obtain an edge information image, specifically including: Gradient calculation is performed on the smoothed noise-suppressed image to obtain the gradient calculation result; The gradient calculation results are subjected to non-maximum suppression processing, and the data after non-maximum suppression processing is subjected to double threshold edge selection to generate edge selection values; Based on the set high and low thresholds, determine whether the edge selection value is a boundary value; If the edge selection value is less than the low threshold, then the edge selection value is not a boundary value; if the edge selection value is greater than the high threshold, then the edge selection value is a strong boundary value. Based on the determination result of the boundary value, an edge information image is obtained.

4. A visual inspection system for arc length in electric arc additive manufacturing, characterized in that, include: An image acquisition module is used to acquire raw electric arc images; the raw electric arc images are acquired by an industrial CCD camera located on a three-axis CNC machine tool support. The image preprocessing module is used to preprocess the original electric arc image to generate an edge information image. Specifically, it includes: using the color conversion method in the third-party computer vision library OpenCV to perform color conversion processing on the original electric arc image to generate a grayscale image; performing smoothing and noise suppression processing on the grayscale image to generate a smoothed and noise-suppressed image; and performing edge detection and pixel binarization processing on the smoothed and noise-suppressed image to obtain the edge information image. The arc peak width determination module is used to determine the width value of the arc peak within the first dashed box in the edge information image; the first dashed box is used to detect the arc width region of the edge information image. An arc peak stage determination module is used to determine that the edge information image is in the arc peak stage when the width value at the arc peak is greater than or equal to a set threshold, and to delete the edge information image data in the arc peak stage. The arc baseline stage determination module is used to determine that the edge information image is in the arc baseline stage when the width value at the arc peak is less than a set threshold, and to generate an arc length variation curve based on the edge information image in the arc baseline stage. Specifically, this includes: detecting the target image using a second dashed box to determine the ordinate value of the target point; the second dashed box is used to detect the arc length region of the target image; the target image is the edge information image in the arc baseline stage; and the ordinate value of the target point is the value retrieved within the second dashed box. The ordinate values ​​of non-zero pixel values ​​are used as follows: The ordinate values ​​of the target point and the arc calibration value are combined to form an array; the array includes two elements; each element includes the ordinate value of the target point and the arc calibration value; the arc calibration value is the ordinate value of the starting point of the arc within the second dashed box; the data in the element are subtracted to obtain a difference value, and the difference value is multiplied by a pixel conversion coefficient to obtain the arc length value; the arc length value is subjected to abrupt data suppression processing to generate an arc length variation curve; the abrupt data suppression method is amplitude limiting.

5. A visual inspection device for arc length in electric arc additive manufacturing, characterized in that, include: Industrial CCD cameras, three-axis CNC machine tools, and control computers; The industrial CCD camera is located on the support of the three-axis CNC machine tool and is used to acquire raw electric arc images; The three-axis CNC machine tool is used for three-way arc additive manufacturing. The control computer is connected to the industrial CCD camera and is used to execute the arc length visual detection method according to any one of claims 1-3.

Images