Handheld laser welding gun, control method and storage medium

Abstract

The invention relates to the technical field of handheld laser welding, and provides a handheld laser welding gun, a control method and a storage medium. The handheld laser welding gun comprises a shell, a beam combiner and a focusing lens. A mounting cavity is formed in the shell; the beam combiner is mounted in the mounting cavity and used for combining the blue laser and the infrared laser and transmitting the combined laser to the focusing lens; and the focusing lens is used for collimating and focusing the combined laser and outputting the combined laser out of the mounting cavity.

Description

Technical Field

[0001] This application relates to the field of handheld laser welding technology, and more particularly to handheld laser welding guns, control methods, and storage media. Background Technology

[0002] Most laser welding guns in related technologies use a single wavelength near-infrared laser for welding. However, when welding highly reflective metals (such as copper and aluminum) with infrared lasers, the low absorption rate of infrared lasers on these metals can easily lead to uneven heat distribution during the welding process, thus affecting the welding quality. Summary of the Invention

[0003] This application aims to address at least one of the technical problems existing in the related art. To this end, this application proposes a handheld laser welding gun, a control method, and a storage medium.

[0004] A handheld laser welding gun according to a first aspect of this application includes a housing, a beam combiner, and a focusing lens;

[0005] The housing has a mounting cavity;

[0006] The beam combiner is installed in the mounting cavity. The beam combiner is used to combine blue laser light and infrared laser light, and to transmit the combined laser light to the focusing lens.

[0007] The focusing lens is used to collimate and focus the combined laser beam, so that the combined laser beam is output outside the mounting cavity.

[0008] According to one embodiment of this application, the beam combiner includes a first surface and a second surface disposed opposite to each other. The middle position of the first surface is located on the transmission path of the blue laser, and the middle position of the second surface is located on the transmission path of the infrared laser. The beam combiner is capable of reflecting the blue laser and transmitting the red laser.

[0009] According to one embodiment of this application, the handheld laser welding gun further includes a galvanometer located on the transmission path of the combined laser beam, and the galvanometer is used to reflect the combined laser beam to the focusing lens.

[0010] According to one embodiment of this application, the housing includes a handheld part, a main body part, and a welding part connected in sequence. The laser beam, after being combined, passes through the welding part and exits the mounting cavity to weld the workpiece.

[0011] The handheld laser welding gun also includes a shielding mask, a camera, and a display screen. The shielding mask is connected to the outer wall of the housing. The handheld part and the welding part are located on opposite sides of the shielding mask. The camera is located on the side of the shielding mask facing the welding part, and the display screen is located on the side of the shielding mask facing the handheld part. The camera is used to acquire images of the welded area of ​​the workpiece and display them on the display screen.

[0012] According to a second aspect of this application, a control method, applied to the handheld laser welding gun as described above, includes:

[0013] S100. Based on the real-time shooting data of the camera, determine the real-time image of the welded part of the workpiece;

[0014] S200: Transmit the real-time image to the display screen for real-time display.

[0015] According to one embodiment of this application, before transmitting the real-time image to the display screen for real-time display, the following steps are further included:

[0016] S210. Based on the real-time shooting data of the camera, acquire at least two frames of welding area images under different exposure times to obtain a multi-exposure image group;

[0017] S220. Based on the brightness distribution characteristics of each image in the multi-exposure image group, determine the position information of the highlight area and the dark area to obtain the dynamic range mapping basis;

[0018] S230. Based on the dynamic range mapping criteria, pixel-level fusion processing is performed on the multi-exposure image group to obtain a high dynamic range image that suppresses solder joint overexposure and retains dark details.

[0019] S240. The high dynamic range image is used to replace the real-time image and is input to the display screen as the image to be displayed in real time.

[0020] According to one embodiment of this application, before transmitting the real-time image to the display screen for real-time display, the following steps are further included:

[0021] S310. Based on the camera being configured with a polarizing filter or a near-infrared photosensitive module, the original image of the welding area containing polarized light components or near-infrared band information is obtained, and an auxiliary image of smoke penetration is obtained.

[0022] S320. Based on the differences in the scattering of polarized light or the transmission characteristics of near-infrared light by smoke particles in the smoke penetration auxiliary image, identify the spatial distribution of the smoke-covered area and obtain the smoke interference mask.

[0023] S330. Based on the smoke interference mask, perform local contrast enhancement and background transmission restoration on the smoke penetration auxiliary image to obtain a smoke penetration enhanced image.

[0024] S340. The real-time image is replaced by the smoke and dust penetration enhancement image, and the image to be displayed is input to the display screen for real-time display.

[0025] According to one embodiment of this application, before transmitting the real-time image to the display screen for real-time display, the following steps are further included:

[0026] S410. Calculate the gradient amplitude based on the real-time image to obtain the edge intensity distribution map of the welding area;

[0027] S420. Based on the connection relationship of continuous high gradient pixels in the edge intensity distribution map, extract the candidate contour of the molten pool boundary to obtain the preliminary molten pool contour.

[0028] S430. Based on the geometric closure and symmetry characteristics of the preliminary molten pool profile, non-molten pool interference edges are eliminated to obtain the accurate molten pool profile.

[0029] S440. Based on the precise molten pool contour, a highlighted marker line is superimposed on the real-time image to obtain an enhanced image with contour markers;

[0030] S450. The enhanced image with outline markings replaces the real-time image and is input to the display screen as the image to be displayed in real time.

[0031] According to one embodiment of this application, ...

[0032] According to one embodiment of this application, ...

[0033] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0034] To more clearly illustrate the technical solutions in the embodiments or related technologies of this application, the accompanying drawings used in the description of the embodiments or related technologies will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a schematic diagram of the handheld laser welding gun provided by the present invention;

[0036] Figure 2This is a flowchart of the handheld laser welding gun control method provided by the present invention. Detailed Implementation

[0037] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0038] The following is combined Figures 1 to 2 This application describes the handheld laser welding gun, control method, and storage medium.

[0039] According to the embodiments of the first aspect of this application, such as Figure 1 The handheld laser welding gun includes a housing 1, a beam combiner 2, and a focusing lens 3;

[0040] The housing 1 has a mounting cavity 11;

[0041] The beam combiner 2 is installed in the mounting cavity 11. The beam combiner 2 is used to combine blue laser and infrared laser, and to transmit the combined laser to the focusing lens 3.

[0042] The focusing lens 3 is used to collimate and focus the combined laser beam, and to output the combined laser beam outside the mounting cavity 11.

[0043] According to an embodiment of this application, the handheld laser welding gun has a housing 1 as its main structure, with a closed mounting cavity 11 inside to accommodate optical components and provide an external contour for the operator to hold. A beam combiner 2 is fixedly mounted within the mounting cavity 11, and its surface is coated with a specific wavelength-selective film layer, enabling it to simultaneously receive incident beams from both a blue laser source and an infrared laser source, and combine them into a coaxial or near-coaxial composite laser beam. The wavelength range of blue lasers is typically 400nm to 500nm, and the wavelength range of infrared lasers is typically 800nm ​​to 1100nm. Since highly reflective metals such as copper and aluminum have significantly higher absorption rates for blue light than for infrared light (for example, pure copper has an absorption rate of over 60% for 450nm blue light, but less than 5% for 1070nm infrared light), introducing blue laser light can effectively improve the initial energy coupling efficiency.

[0044] The combined laser beam propagates along the optical path to the focusing lens 3. The focusing lens 3, mounted at the front end of the mounting cavity 11, collimates and focuses the combined laser beam, forming a high-energy-density spot, which is then output to the outside of the mounting cavity 11 and applied to the surface of the workpiece to be welded. During the welding process, the blue laser is first efficiently absorbed by the highly reflective metal, rapidly increasing the material's surface temperature and reducing its reflectivity to the infrared laser. Subsequently, the infrared laser is more effectively absorbed by the heated area, achieving deep melting and stable heat conduction. The synergistic effect of the two laser wavelengths ensures uniform heat distribution on the material's surface and within, avoiding problems such as difficulty initiating welding, increased spatter, or discontinuous welds caused by the low absorption rate of a single infrared laser.

[0045] Therefore, by setting a beam combiner 2 and a focusing lens 3 inside the handheld laser welding gun, and utilizing the combined output of blue laser and infrared laser, the welding adaptability and process stability of highly reflective metals such as copper and aluminum are significantly improved. High-quality, low-defect laser welding effect is achieved without significantly increasing the complexity of the system.

[0046] In some embodiments, such as Figure 1 The beam combiner 2 includes a first surface 21 and a second surface 22 arranged opposite to each other. The middle position of the first surface 21 is located on the transmission path of the blue laser, and the middle position of the second surface 22 is located on the transmission path of the infrared laser. The beam combiner 2 can reflect the blue laser and transmit the red laser.

[0047] Understandably, the blue laser emitted from the blue laser source propagates along the first optical path and is incident on the middle position of the first surface 21 of the beam combiner 2; the infrared laser emitted from the infrared laser source propagates along the second optical path and is incident on the middle position of the second surface 22 of the beam combiner 2. The beam combiner 2 is coated with a wavelength-selective dielectric film inside or on its surface. This film is designed to have high reflectivity for blue laser light and high transmittance for infrared laser light.

[0048] When the blue laser light irradiates the center of the first surface 21, it is reflected by the beam combiner 2, changing its propagation direction. Simultaneously, the infrared laser light enters from the center of the second surface 22, passing through the beam combiner 2 with minimal attenuation. By appropriately setting the incident angles of the blue and infrared lasers and the installation orientation of the beam combiner 2, the reflected blue laser light and the transmitted infrared laser light form a collinear or nearly collinear composite beam on the exit side of the beam combiner 2. This composite beam then enters the focusing lens 3, is collimated and focused, and is output outside the mounting cavity 11, acting on the workpiece surface.

[0049] Because the blue laser is precisely reflected and the infrared laser is efficiently transmitted, the two wavelengths of laser light achieve precise spatial overlap, ensuring uniform energy distribution within the welding spot. Simultaneously, arranging the blue laser path on the first surface 21 and the infrared laser path on the second surface 22 avoids mutual obstruction or interference between the two laser beams during the incident phase, improving the compactness of the optical path layout and assembly tolerance. Furthermore, the dual-sided independent incident design of the beam combiner 2 facilitates separate adjustment of the incident angles of the blue and infrared lasers, optimizing the beam combining quality.

[0050] Thus, by configuring the beam combiner 2 with a structure having a first surface 21 and a second surface 22, and by having the blue laser and the infrared laser act on the middle position of the two surfaces respectively, while utilizing its optical properties of reflecting blue light and transmitting infrared light, the handheld laser welding gun achieves efficient and stable beam combining of the two lasers, providing a reliable beam foundation for high-quality welding of highly reflective metals.

[0051] In some examples, the beam combiner 2 is a fused silica substrate with a 450nm high reflectivity / 1070nm high transmittance dichroic film coated on its surface; a blue laser is incident on the first surface 21 at a 45° angle and propagates along the optical axis after reflection; an infrared laser is incident perpendicularly on the second surface 22 along the optical axis and is directly transmitted; the beam combiner 2 has a thickness of 2mm to 5mm and its edges are treated with an extinction effect to suppress stray light; the two laser sources are coupled into the housing 1 through optical fiber, and the optical path is guided to the beam combiner 2 through a reflector or collimator.

[0052] In some embodiments, such as Figure 1 The handheld laser welding gun also includes a galvanometer 4, which is located on the transmission path of the combined laser beam. The galvanometer 4 is used to reflect the combined laser beam to the focusing lens 3.

[0053] Understandably, by setting a galvanometer 4 on the laser transmission path after beam combining and using its reflection function to guide the laser to the focusing lens 3, the handheld laser welding gun achieves dynamic beam control of the composite laser, which expands the welding mode and application scenarios while ensuring the synergistic effect of the two wavelengths.

[0054] In some embodiments, such as Figure 1 The housing 1 includes a handheld part 12, a main body part 13 and a welding part 14 connected in sequence. The laser beam after beaming passes through the welding part 14 and exits the mounting cavity 11 to weld the workpiece.

[0055] The handheld laser welding gun also includes a shielding mask 5, a camera, and a display screen 6. The shielding mask 5 is connected to the outer wall of the housing 1. The handheld part 12 and the welding part 14 are located on both sides of the shielding mask 5. The camera is located on the side of the shielding mask 5 facing the welding part 14. The display screen 6 is located on the side of the shielding mask 5 facing the handheld part 12. The camera is used to acquire images of the welded area of ​​the workpiece and display them on the display screen 6.

[0056] Understandably, the handheld part 12 is for the operator to hold, and the welding part 14 is located at the front end of the housing 1. It is equipped with a focusing lens 3. The laser beam, after being combined, passes through the welding part 14 and exits the mounting cavity 11, acting on the surface of the workpiece to complete the welding operation. The shielding mask 5 is fixedly connected to the outer wall of the housing 1, and its position is between the handheld part 12 and the welding part 14, physically separating the area where the operator is located from the welding operation area.

[0057] The shielding mask 5 is made of optical protective materials with high absorption or high reflection characteristics for blue and infrared lasers (such as dark glass doped with metal oxides or special laser-protective plastics). It can effectively block high-energy lasers reflected or scattered from the workpiece surface during welding, preventing them from directly irradiating the operator's eyes or skin, and significantly improving operational safety. The shielding mask 5 is particularly important when welding highly reflective metals such as copper and aluminum, where the reflected laser energy is strong.

[0058] A camera is mounted on the side of the mask 5 facing the welding section 14, directly opposite the welding area, to acquire real-time image information of the workpiece welding point. A display screen 6 is mounted on the side of the mask 5 facing the handheld part 12, facing the operator, and is electrically connected to the camera to display the real-time images acquired by the camera. When the operator holds the handheld part 12 for welding, they do not need to look beyond the mask 5 to observe the weld point; they can clearly see the molten pool shape, weld trajectory, and surrounding environment through the display screen 6, achieving precise positioning and process monitoring. This visualization design not only improves welding accuracy but also avoids operational errors caused by obstructed vision.

[0059] Therefore, by placing the shielding mask 5 between the handheld part 12 and the welding part 14, and integrating the camera and display screen 6, the handheld laser welding gun effectively blocks dangerous laser reflections while providing operators with safe and intuitive real-time visual feedback, thus balancing personal protection and ease of operation.

[0060] In some examples, the shielding mask 5 is an arc-shaped or planar structure with a light transmittance of less than 0.1% at 450nm and 1070nm; the camera is an industrial-grade CMOS sensor with automatic exposure and strong light resistance; the display screen 6 is a 3.5-inch to 7-inch LCD or OLED screen that supports brightness adjustment; the camera and display screen 6 are connected via a flexible circuit board or wireless module with a signal delay of less than 50ms; the shielding mask 5 is detachable or angle-adjustable for easy maintenance and adaptation to different welding postures.

[0061] According to an embodiment of the second aspect of this application, such as Figure 2 The control method is applied to the aforementioned handheld laser welding gun. The control method includes:

[0062] S100: Based on real-time camera data, determine the real-time image of the weld joint of the workpiece;

[0063] The built-in control module of the handheld laser welding gun activates the camera mounted on the side of the shielding mask 5 facing the welding section 14. This camera is an industrial-grade CMOS image sensor with a resolution of 1280×720 pixels, a frame rate of 30 frames per second, and features automatic exposure and wide dynamic range. When the operator aligns the welding section 14 with the workpiece, the camera continuously captures image data of the workpiece surface area, including the weld, molten pool, and surrounding heat-affected zone. The control module extracts the current frame from the video stream as a real-time image of the weld joint on the workpiece. Here, the "weld joint" refers to the metal joining area where the laser beam is about to act or is currently acting, and its image must clearly show the geometric contours and surface condition.

[0064] S200: Transmit the real-time image to the display screen 6 for real-time display;

[0065] The control module transmits the real-time images acquired by the S100 to the display screen 6 mounted on the side of the face shield 5 facing the handheld part 12 via an internal high-speed data bus (such as a MIPI CSI-2 interface). This display screen 6 is a 3.5-inch TFT LCD screen with a brightness of 500 nits and supports anti-glare processing. After color correction and contrast enhancement, the image is displayed in full-screen mode in real time on the display screen 6 with a latency of no more than 80 milliseconds. Operators can clearly see the details of the welding area by observing the display screen 6 without needing to bring their eyes close to areas of high temperature, strong light, or splashes.

[0066] Understandably, this application determines the real-time image of the welding point of the workpiece based on the real-time shooting data of the camera in step S100, and transmits the real-time image to the display screen 6 for real-time display in step S200. Through the coordinated execution of the above two steps, the handheld laser welding gun realizes a visual remote observation mechanism for the welding process. When the operator holds the handheld part 12 to perform the operation, they only need to look at the display screen 6 inside the shielding mask 5 to obtain a clear, real-time image of the welding area, completely avoiding the risks of strong light stimulation, ultraviolet radiation, and high-temperature spatter damage caused by directly looking at the high-energy laser point. At the same time, since the camera and display screen 6 are integrated into the same device and directly connected through local lines, image transmission does not need to rely on external networks or additional display terminals, and the system has fast response and high reliability. Without changing the original laser beam combining and focusing optical path, this solution deeply integrates safety protection and operation assistance through integrated structural design, significantly improving the human-machine safety and operational accuracy of handheld laser welding operations.

[0067] In some embodiments, before transmitting the real-time image to the display screen 6 for real-time display, the following steps are also included:

[0068] S210. Based on the real-time shooting data of the camera, acquire at least two frames of welding area images under different exposure times to obtain a multi-exposure image group;

[0069] The built-in control module of the handheld laser welding gun controls the camera to continuously acquire at least two frames of images of the welding area within a single imaging cycle, with each frame using a different exposure time. Specifically, the first frame uses a short exposure time (set to 1 / 1000 of a second) to capture high-brightness areas (such as the laser pool and reflective metal surfaces) without overexposure; the second frame uses a long exposure time (set to 1 / 60 of a second) to preserve details in low-brightness areas (such as workpiece shadows and dark weld edges). The two frames have the same spatial alignment, together forming a multi-exposure image group. Here, a "multi-exposure image group" refers to a set of two or more images acquired simultaneously with different exposure parameters in the same scene, used to expand the overall dynamic range.

[0070] S220. Based on the brightness distribution characteristics of each image in the multi-exposure image group, determine the location information of the highlight area and the dark area to obtain the dynamic range mapping basis;

[0071] The control module performs brightness distribution analysis on the multi-exposure image group obtained by S210: First, each image is converted into a grayscale image, and then the brightness value of each pixel is calculated (ranging from 0 to 255). For short-exposure images, pixel areas with brightness values ​​greater than 220 are identified as highlight areas; for long-exposure images, pixel areas with brightness values ​​less than 30 are identified as shadow areas. Through coordinate mapping, the positional information of highlight and shadow areas in the image is recorded, forming a dynamic range mapping basis. Here, "dynamic range mapping basis" refers to the spatial position and brightness feature data used to guide the weight allocation of each region in the subsequent image fusion process.

[0072] S230. Based on the dynamic range mapping, pixel-level fusion processing is performed on the multi-exposure image group to obtain a high dynamic range image that suppresses overexposure of solder joints and retains details in dark areas.

[0073] Based on the dynamic range mapping obtained from S220, the control module performs pixel-level weighted fusion processing on the multi-exposure image group: in highlight areas, the short-exposure image is given a higher fusion weight (e.g., weight 0.9, weight 0.1 for the long-exposure image); in shadow areas, the long-exposure image is given a higher weight (e.g., weight 0.9, weight 0.1 for the short-exposure image); in intermediate brightness areas, a linear transition weight is used. The fusion formula is:

[0074] Output pixels ,

[0075] in These are the pixel values ​​of a short-exposure image. These are the pixel values ​​of the long-exposure image. and The normalized weights are dynamically calculated based on the position. This processing generates a high dynamic range image that suppresses overexposure of solder joints while preserving details in shadows. Here, a "high dynamic range image" refers to a composite image whose effective brightness range is significantly better than a single-frame image, while simultaneously presenting details in both bright and dark areas.

[0076] S240: A high dynamic range image is used to replace the real-time image and is input to the display screen 6 as the image to be displayed in real time.

[0077] The control module replaces the original single-frame real-time image with the high dynamic range image generated by S230, and transmits it to the display screen 6 via the internal data bus as the final image to be displayed. After receiving the image, the display screen 6 immediately refreshes the display, so that the operator sees a high-quality image with no overexposed molten pool, clearly visible weld outline and surrounding structure, which significantly improves the accuracy of visual judgment.

[0078] Understandably, in step S210, at least two frames of welding area images with different exposure times are acquired based on the real-time shooting data of the camera to obtain a multi-exposure image group. In step S220, the position information of the highlight area and the dark area is determined based on the brightness distribution characteristics of each image in the multi-exposure image group to obtain the dynamic range mapping basis. Based on the dynamic range mapping basis, the multi-exposure image group is subjected to pixel-level fusion processing in step S230 to obtain a high dynamic range image that suppresses overexposure of the welding point and retains the details of the dark area. Finally, in step S240, the high dynamic range image is input to the display screen 6 as the image to be displayed for real-time display. Through the coordinated execution of the above four sub-steps, the handheld laser welding gun overcomes the visual blind spot problem of "bright areas are all white and dark areas are all black" in the strong light welding scene of traditional single-exposure imaging. The system does not rely on external HDR displays or complex post-processing equipment. It can generate high-contrast, wide dynamic range welding views in real time using only a local camera and control module. This allows operators to clearly identify the molten pool shape, weld alignment, and details of the surrounding workpiece from a safe distance. This not only improves welding accuracy but also further enhances the safety of the human eye from direct exposure to strong light, achieving a dual optimization of safety and work quality.

[0079] In some embodiments, before transmitting the real-time image to the display screen 6 for real-time display, the following steps are also included:

[0080] S310: Based on a camera equipped with a polarizing filter or a near-infrared photosensitive module, it acquires the original image of the welding area containing polarized light components or near-infrared band information, and obtains an auxiliary image of smoke penetration.

[0081] The built-in control module of the handheld laser welding gun activates the camera, which integrates a switchable optical front end: one configuration uses a linear polarization filter (with the transmission axis fixed at 0 degrees), and the other uses a near-infrared photosensitive module (sensing wavelength from 800 to 1000 nanometers). During welding, the high-temperature fumes generated by metal evaporation strongly scatter visible light but have high transmittance for light with specific polarization states or near-infrared light. The control module selects one of these modes for image acquisition based on preset modes—for example, activating the polarization filter to acquire the original image of the welding area containing polarized light components; or activating the near-infrared photosensitive module to acquire the original image of the welding area containing near-infrared wavelength information. This image, due to its stronger fume penetration capability, is defined as a fume penetration-assisted image. Here, "fume penetration-assisted image" refers to the original image acquired using polarized light scattering suppression or near-infrared high transmittance characteristics, which is less affected by fume interference.

[0082] S320. Based on the differences in the scattering of polarized light by smoke particles or the transmission characteristics of near-infrared light in the smoke penetration auxiliary image, the spatial distribution of the smoke-covered area is identified to obtain the smoke interference mask.

[0083] The control module identifies smoke and dust regions in the smoke penetration auxiliary image obtained by S310: If polarization imaging is used, the polarization contrast of different regions in the image is analyzed—smoke particles scatter unpolarized light strongly, resulting in a decrease in polarization degree; therefore, regions with low polarization contrast are marked as smoke-occluded areas. If near-infrared imaging is used, the characteristics of smoke and dust—weak absorption and strong transmission of near-infrared light—are utilized, and occlusion regions are identified through brightness threshold segmentation (e.g., areas with pixel brightness below 80 are considered dense smoke and dust areas). The system generates a binary image as a smoke and dust interference mask, where white pixels represent smoke and dust-occluded areas, and black pixels represent clearly visible areas. Here, "smoke and dust interference mask" refers to a spatial distribution map used to identify areas affected by smoke and dust in the image.

[0084] S330. Based on the smoke and dust interference mask, local contrast enhancement and background transmission restoration are performed on the smoke and dust penetration auxiliary image to obtain the smoke and dust penetration enhanced image.

[0085] The control module performs local enhancement processing on the smoke and dust interference mask generated by the S320 to assist in smoke and dust penetration: In the smoke-covered areas (white parts of the mask), an adaptive histogram equalization algorithm is applied to improve local contrast, and the background transmittance is estimated by combining the dark channel prior model to restore the surface texture of the workpiece blurred by smoke and dust; in the uncovered areas (black parts of the mask), the original pixel values ​​are kept unchanged to avoid over-enhancement and the introduction of noise. After this processing, a smoke and dust penetration enhanced image is output, in which the weld contour, molten pool edge, and surrounding structure are still clearly distinguishable in the smoke and dust environment.

[0086] S340: The real-time image is replaced by a smoke and dust penetration enhanced image, which is then input to the display screen 6 as the image to be displayed in real time.

[0087] The control module replaces the original single-frame visible light real-time image with the smoke penetration enhancement image generated by S330, and transmits it to display screen 6 as the final image to be displayed via the internal high-speed data bus. Display screen 6 refreshes the image in real time, allowing operators to clearly observe the true state of the workpiece welding area even in a smoky welding environment, without waiting for the smoke to settle naturally or interrupting the operation.

[0088] Understandably, in step S310, the system acquires an original image of the welding area containing polarized light components or near-infrared band information using a camera equipped with a polarizing filter or a near-infrared photosensitive module to obtain a smoke penetration auxiliary image. In step S320, based on the differences in scattering of polarized light by smoke particles or their transmission characteristics to near-infrared light in the smoke penetration auxiliary image, the system identifies the spatial distribution of the smoke-obstructed area to obtain a smoke interference mask. In step S330, based on the smoke interference mask, the system performs local contrast enhancement and background transmission restoration on the smoke penetration auxiliary image to obtain a smoke penetration enhancement image. Finally, in step S340, the smoke penetration enhancement image is input to the display screen 6 as the image to be displayed in real time. Through the coordinated execution of these four sub-steps, the handheld laser welding gun effectively overcomes the inherent defects of traditional visible light imaging in welding smoke environments, such as blurred vision and loss of detail. The system utilizes the suppression effect of polarized light scattering or the high penetration capability of near-infrared light through smoke, combined with a local enhancement algorithm guided by an intelligent mask, to achieve visual assurance of "clear visibility" without adding external dust removal equipment. This solution significantly improves welding visibility and operational reliability under heavy dust conditions, while avoiding operators having to frequently approach the welding point due to poor visibility, further enhancing operational safety and achieving high-quality human-machine collaborative welding control in harsh environments.

[0089] In some embodiments, before transmitting the real-time image to the display screen 6 for real-time display, the following steps are also included:

[0090] S410. Calculate the gradient magnitude based on the real-time image to obtain the edge intensity distribution map of the welding area;

[0091] The built-in control module of the handheld laser welding gun calculates the gradient magnitude of the real-time image. Specifically, the Sobel operator is used to calculate the first-order partial derivatives of the image in the horizontal and vertical directions, respectively, to obtain the horizontal gradient component Gx and the vertical gradient component Gy; then, the gradient magnitude is calculated according to the formula. The gradient intensity of each pixel is calculated to generate a grayscale image, called the edge intensity distribution map of the welding area. In this image, the edge of the molten pool exhibits a high gradient value due to abrupt changes in brightness, while the background area has a lower gradient. Here, the "edge intensity distribution map" refers to a two-dimensional intensity map that reflects the degree of brightness change at various locations in the image.

[0092] S420. Based on the connection relationship of continuous high gradient pixels in the edge intensity distribution map, extract the candidate contour of the molten pool boundary to obtain the preliminary molten pool contour.

[0093] The control module performs threshold segmentation on the edge intensity distribution map obtained from S410, setting the gradient magnitude threshold to 80 (maximum value 255 for 8-bit images), and marking pixels above this threshold as edge candidate points. Subsequently, an eight-neighborhood connected component analysis algorithm is used to connect spatially continuous high-gradient pixels into closed or nearly closed curves, forming multiple edge loops. Based on the prior knowledge that the melt pool is usually elliptical and located in the central region of the image, the largest connected edge loop with an area between 500 and 3000 pixels and a center coordinate located in the 60% region of the image center is selected as the preliminary melt pool contour.

[0094] S430. Based on the geometric closure and symmetry characteristics of the preliminary molten pool profile, non-molten pool interference edges are eliminated to obtain the accurate molten pool profile.

[0095] The control module performs geometric feature verification on the preliminary molten pool profile obtained from S420: first, it determines whether it is a closed curve (the distance between the start and end points is less than 5 pixels); second, it calculates the ratio of its major axis to minor axis. If it is between 1.0 and 2.5, it is considered to have reasonable symmetry. If both closure and symmetry are satisfied, the profile is retained; otherwise, it is judged as spatter, weld slag, or reflective interference edges and is discarded. After this screening, a unique accurate molten pool profile is output. Here, "accurate molten pool profile" refers to a closed curve that represents the true boundary of the molten pool after geometric reasonableness verification.

[0096] S440. Based on the accurate molten pool contour, a highlight marker line is superimposed on the real-time image to obtain an enhanced image with contour markers.

[0097] The control module maps the precise molten pool contour obtained from S430 back to the original real-time image coordinate system and draws a 2-pixel-wide highlight line at the corresponding pixel position. This highlight line is cyan (RGB value 0, 255, 255) because it has high visual contrast against metallic backgrounds and in strong light. The superimposed image is called the enhanced image with contour markers, which retains the original scene information while highlighting the molten pool boundary.

[0098] S450: An enhanced image with outline markings is used to replace the real-time image and is input to the display screen 6 as the image to be displayed in real time;

[0099] The control module replaces the original real-time image with the enhanced image with outline markings generated by S440 as the final image to be displayed, and transmits it to the display screen 6 via the internal data bus. The operator can clearly see the real-time shape, size and position of the molten pool through the display screen 6, which facilitates precise control of the welding path and speed, and avoids defects such as incomplete fusion, undercut or burn-through.

[0100] Understandably, in step S410, the gradient magnitude is calculated based on the real-time image to obtain the edge intensity distribution map of the welding area. Based on the connection relationship of consecutive high-gradient pixels in the edge intensity distribution map, candidate contours of the molten pool boundary are extracted in step S420 to obtain a preliminary molten pool contour. Based on the geometric closure and symmetry features of the preliminary molten pool contour, non-molten pool interference edges are eliminated in step S430 to obtain a precise molten pool contour. Based on the precise molten pool contour, a highlighted marker line is superimposed on the real-time image in step S440 to obtain an enhanced image with contour markings. Finally, in step S450, the enhanced image with contour markings is input to the display screen 6 for real-time display. Through the coordinated execution of the above five sub-steps, the handheld laser welding gun achieves automatic identification and visualization enhancement of the molten pool boundary. This method does not rely on deep learning models or external calibration, completing real-time image processing locally, effectively solving the problem of difficulty in identifying the molten pool boundary under strong light, reflection, and spatter interference. Operators can intuitively grasp the dynamic shape of the molten pool through the display screen 6, significantly improving the controllability and consistency of the welding process, while avoiding approaching dangerous areas due to blurred vision, further enhancing operational safety and process quality.

[0101] In some embodiments, the control method further includes the following steps:

[0102] Step S1: Based on the workpiece surface image captured by the camera, obtain the position information of the welding standard line and the boundary identification information of the preset welding area;

[0103] Before welding begins, a camera captures high-resolution images of the workpiece surface. The central controller runs an image processing algorithm, first identifying the welding standard lines (e.g., a continuous thin white line) pre-printed or engraved on the workpiece surface using template matching or a deep learning model, and extracting their pixel coordinate sequence. Simultaneously, it identifies the closed boundary markers surrounding the area to be welded (e.g., area markers composed of QR codes or geometric contours), generating polygonal boundary data for that area. Thus, the positional information of the welding standard lines and the boundary marker information of the pre-defined welding area are obtained and stored in memory for subsequent steps.

[0104] Step S2: Based on the position information of the welding standard line and the center position of the indicator light projected onto the workpiece surface in real time, obtain the offset of the indicator light relative to the welding standard line;

[0105] The workpiece is then moved to the welding station, where the operator approaches it with a welding torch. At this point, an indicator light projects a circular spot onto the workpiece surface. A camera simultaneously captures a real-time image containing this spot. The central controller determines the position of the indicator light center in the image coordinate system through image segmentation and centroid calculation, and aligns it with the welding standard line position obtained in step S1 (by transforming it to the same coordinate system using a pre-calibrated camera-workpiece extrinsic parameter matrix). This allows the calculation of the vertical distance, or offset, from the indicator light center to the nearest standard line point.

[0106] Step S3: Based on the offset, determine whether the offset exceeds the preset first threshold. If it does, control the welding power supply to stop outputting welding energy.

[0107] The central controller compares the offset with a preset first threshold (e.g., ±1.0 mm). If the absolute value of the offset is greater than this threshold, it indicates that the welding torch has significantly deviated from the theoretical trajectory, and continuing welding will result in an incorrect weld position. At this point, the controller immediately sends a stop welding command to the welding power source, cutting off the arc or laser output and forcibly stopping the release of welding energy until the offset returns to the allowable range.

[0108] Step S4: Based on the boundary identification information and the current position of the welding torch, determine whether the current welding progress has covered the entire preset welding area.

[0109] During the welding process, the controller continuously records the actual trajectory of the welding torch's point of action (approximately by the sequence of the center position of the indicator light). A coverage analysis is performed on the area enclosed by this trajectory and the boundary markers obtained in step S1: if the trajectory covers more than 95% of the area (calculated using rasterization or integration methods), the current welding area is determined to be complete; otherwise, it is determined to be incomplete. This determination result is the "completed state".

[0110] Step S5: Based on the real-time acquisition of the reflection spectrum at the welding position of the workpiece by the spectral sensor, the current material type of the workpiece is obtained, and the corresponding welding parameter configuration table is retrieved based on the material type, thereby adjusting the output current and welding speed of the welding power supply.

[0111] At the moment of welding initiation and throughout the process, the spectral sensor on the support collects the spectral signal reflected from the workpiece welding position after excitation (e.g., using LIBS laser-induced breakdown spectroscopy, wavelength range 200–800 nm). The central controller matches the collected spectral curve with the built-in material database (containing characteristic spectral lines of typical materials such as stainless steel, carbon steel, and aluminum alloy) to identify the current workpiece material type. Subsequently, it retrieves the recommended parameter combination for the corresponding material from the pre-stored welding parameter configuration table (e.g., 180 A current and 0.8 m / min for stainless steel; 150 A current and 1.2 m / min for aluminum alloy), and adjusts the output current of the welding power supply and the wire feed / movement speed in real time to ensure the target penetration depth.

[0112] Step S6: Based on the completion status and the current penetration depth target signal, determine whether to allow entry into the next welding area; if the completion status is completed and the penetration depth target signal is valid, unlock the welding permission for the next welding area; otherwise, maintain the current area locked status.

[0113] The penetration depth compliance signal is generated by the penetration depth assessment module: This module predicts whether the current penetration depth meets the process requirements (e.g., ≥2.0 mm) based on the actual welding parameters adjusted in step S5, arc voltage fluctuation characteristics, and molten pool temperature distribution fed back by the infrared thermal imager, using a lightweight neural network model. If the penetration depth is predicted to meet the requirements, a valid signal is output. The central controller integrates the "completion status" of step S4 with the "penetration depth compliance signal" of this step: Only when both are affirmative (i.e., the area is completed and the penetration depth meets the requirements) will the "next area unlocked" prompt be displayed on the screen, allowing the operator to move to the next preset welding area for operation; otherwise, the system keeps the area locked, and even if the operator manually moves the welding torch to the next area, the welding power supply will not respond to the start command.

[0114] It is understandable that, based on the position information of the welding standard line and the boundary marker information of the preset welding area obtained in step S1, and based on the position information of the welding standard line and the real-time indicator light center position, the offset of the indicator light relative to the welding standard line is obtained in step S2. Based on this offset, dynamic start-stop control of the welding energy output is realized in step S3. Simultaneously, based on the boundary marker information and the welding torch action point position obtained in step S1, the completion status of the current welding area is obtained in step S4. On the other hand, based on the reflection spectrum collected by the spectral sensor at the welding position, the current workpiece material type is obtained in step S5, and the welding parameters are adjusted accordingly to adapt to the penetration depth requirements. A reliable penetration depth target signal is generated. Finally, based on the completion status of step S4 and the penetration depth target signal generated in step S5, step S6 determines whether the area switching conditions are met and executes access control. Through the close collaboration of the above data flow and control logic, the system realizes active intervention on welding torch trajectory deviation, forced constraint on the integrity of the welding area, and material adaptive adjustment of welding parameters during handheld welding. This solves the problems of welding deviation, missed welding, and insufficient penetration depth caused by complete reliance on manual operation in the background technology. It achieves the technical effect of significantly improving welding position accuracy and penetration depth consistency and ensuring welding quality reliability while retaining the flexibility of handheld operation.

[0115] For example, the indicator light source is integrated into the welded part of the housing.

[0116] In some embodiments, step S2 includes the following sub-steps:

[0117] Step S21: Based on the real-time image of the indicator light and the workpiece surface captured by the camera, obtain the light spot area of ​​the indicator light in the real-time image;

[0118] The central controller controls the camera to acquire images of the workpiece surface currently at the welding station, obtaining a real-time image that includes the indicator light and the workpiece surface. Since the indicator light source is a high-brightness red laser, it forms a bright spot area clearly distinguishable from the background in this real-time image. The central controller performs binarization processing by setting a brightness threshold, grouping all consecutive pixels with grayscale values ​​higher than the threshold into the same region, thereby extracting the indicator light spot area in the real-time image. Here, the "spot area" refers to a connected region composed of multiple adjacent pixels, whose overall brightness is significantly higher than the surrounding workpiece surface.

[0119] Step S22: Based on the pixel distribution of the light spot area, obtain the position of the indicator light center;

[0120] The central controller calculates a coordinate point based on the arithmetic mean of the x and y coordinates of all pixels in the light spot area. This coordinate point indicates the location of the light center. Specifically, the average x coordinate is obtained by adding the x coordinates of all pixels in the light spot area and dividing by the total number of pixels; the average y coordinate is obtained by adding the y coordinates of all pixels and dividing by the total number of pixels; the two-dimensional coordinate formed by this average x and y coordinates indicates the location of the light center.

[0121] Step S23: Based on the position information of the welding standard line, obtain the set of continuous line segments of the welding standard line in the real-time image;

[0122] The central controller retrieves the position information of the welding standard line obtained in step S1 of claim 1, which is stored in the form of a series of ordered coordinate points. The central controller connects these coordinate points in pairs sequentially to form multiple straight line segments connected end to end, constituting a continuous set of line segments of the welding standard line in the real-time image. Each straight line segment is defined by two endpoints, and all line segments together approximate the geometry of the original welding standard line.

[0123] Step S24: Based on the position of the indicator light center and the set of continuous line segments, obtain the pair of line segment endpoints closest to the position of the indicator light center;

[0124] The central controller iterates through each straight line segment in the continuous set of line segments, calculates the Euclidean distance from the center of the indicator light to the two endpoints of the line segment, and takes the smaller of the two values ​​as the representative distance of the line segment. Then, the central controller compares the representative distances of all line segments, selects the line segment corresponding to the minimum value, and records the two endpoints of the line segment to obtain the pair of line segment endpoints closest to the center of the indicator light.

[0125] Step S25: Based on the nearest pair of line segment endpoints, obtain the equation of the line connecting the pair of line segment endpoints;

[0126] The central controller constructs a straight line equation based on the x-coordinates and y-coordinates of the nearest pair of line segment endpoints. Specifically, the method is as follows: Let the x-coordinate of the first endpoint be X1 and the y-coordinate be Y1, and the x-coordinate of the second endpoint be X2 and the y-coordinate be Y2; if X1 is not equal to X2, then the slope is calculated as (Y2 - Y1) divided by (X2 - X1), and then substituted into the point-slope form to obtain the straight line equation; if X1 equals X2, then the line is perpendicular, and its equation is that the x-coordinate equals X1. The final straight line equation is used to characterize the direction and position of the local welding standard line.

[0127] Step S26: Based on the position of the indicator light center and the equation of the straight line, obtain the perpendicular distance from the position of the indicator light center to the straight line;

[0128] The central controller calculates the vertical distance from the center of the indicator light to the straight line based on the x and y coordinates of the indicator light's center position and the equation of the straight line. The calculation method is as follows: substitute the x and y coordinates of the indicator light's center position into the general equation of the straight line, take the absolute value, and divide by the magnitude of the direction vector of the straight line. The result is the vertical distance. This vertical distance represents the shortest spatial distance between the center of the indicator light and the welding standard line.

[0129] Step S27: Based on the vertical distance and the preset direction discrimination rule, obtain the offset direction of the indicator light relative to the welding standard line;

[0130] The central controller determines the direction vector of the line segment based on the nearest pair of line segment endpoints. This direction vector is calculated by subtracting the x-coordinate of the first endpoint from the x-coordinate of the second endpoint, and subtracting the y-coordinate of the first endpoint from the y-coordinate of the second endpoint, as the y-coordinate. Subsequently, the central controller constructs a vector pointing from the first endpoint to the center of the indicator light and calculates the cross product of this vector and the direction vector. If the cross product is greater than zero, the indicator light is determined to be to the left of the welding standard line; if the cross product is less than zero, the indicator light is determined to be to the right of the welding standard line; if the cross product is equal to zero, the indicator light is determined to be on the welding standard line. This determination rule is the preset direction discrimination rule, thus obtaining the offset direction of the indicator light relative to the welding standard line.

[0131] Step S28: Based on the vertical distance and offset direction, obtain the offset of the indicator light relative to the welding standard line;

[0132] The central controller uses the vertical distance as the offset amplitude and, combined with the offset direction, generates a signed numerical value as the offset amount: if the offset direction is to the left, the offset amount is negative, and its absolute value equals the vertical distance; if the offset direction is to the right, the offset amount is positive, and its absolute value equals the vertical distance; if there is no offset, the offset amount is zero. This offset amount fully expresses the spatial deviation of the indicator light relative to the welding standard line, which is used in subsequent step S3 for threshold judgment and welding stop control.

[0133] It is understood that in this embodiment, the indicator light spot area in the real-time image is obtained based on sub-step S21, and the center position of the indicator light is obtained in sub-step S22 based on the spot area; based on the welding standard line position information obtained in step S1 of claim 1, the set of continuous line segments in the real-time image is obtained in sub-step S23; based on the center position of the indicator light and the set of continuous line segments, the nearest pair of line segment endpoints is obtained in sub-step S24; based on the pair of line segment endpoints, the local straight line equation is obtained in sub-step S25; based on the center position of the indicator light and the straight line equation, the vertical distance is obtained in sub-step S26; based on the vertical distance and the direction discrimination rule defined by the pair of line segment endpoints, the offset direction is obtained in sub-step S27; finally, based on the vertical distance and the offset direction, the offset amount with symbolic meaning is obtained in sub-step S28. Through the deterministic geometric operations and logical connections of the above eight sub-steps, the central controller can accurately and in real time extract the offset of the indicator light relative to the welding standard line from the original image without relying on models, learning or fuzzy judgment. This provides a reliable and executable perception basis for solving the welding deviation problem caused by manual operation in the background technology, thereby supporting the entire system to realize active intervention on welding trajectory deviation.

[0134] It is understood that this embodiment uses a central controller as the main body to perform precise calculations of the offset of the indicator light relative to the welding standard line. Before the welding operation begins, the system has completed workpiece positioning, image acquisition equipment calibration, and welding standard line identification. The relevant data has been provided by step S1 and stored in the memory of the central controller.

[0135] In some embodiments, step S3 includes the following sub-steps:

[0136] Step S31: Based on the welding stage information of the current welding operation, obtain the current welding stage type;

[0137] The central controller uses the initial moment the welding torch trigger is pressed as the welding start time and, in conjunction with the built-in process timing table, determines the duration of the welding start time. If the duration is less than or equal to 500 milliseconds, the current welding operation is determined to be in the arc initiation stage; if the duration is greater than 500 milliseconds and the remaining length from the preset welding path end point is greater than 10 millimeters, the current welding operation is determined to be in the stable welding stage; if the remaining length is less than or equal to 10 millimeters, the current welding operation is determined to be in the arc termination stage. Thus, the central controller obtains the current welding stage type. Here, "welding stage type" refers to one of the arc initiation stage, stable welding stage, or arc termination stage, used to characterize the process state of the welding process.

[0138] Step S32: Based on the offset and the current welding stage type, obtain the corresponding offset tolerance threshold;

[0139] The central controller retrieves a pre-stored stage-threshold mapping table. This table specifies that the offset tolerance threshold is 2.0 mm for the arc initiation stage, 0.8 mm for the stable welding stage, and 1.2 mm for the arc termination stage. Based on the current welding stage type obtained in sub-step S31, the central controller looks up the corresponding value in this mapping table to obtain the offset tolerance threshold. This offset tolerance threshold represents the maximum distance that the indicator light center is allowed to deviate from the welding standard line under the current welding stage.

[0140] Step S33: Based on the offset and the offset tolerance threshold, obtain the offset exceeding the limit judgment result;

[0141] The central controller compares the absolute value of the offset obtained in sub-step S28 with the offset tolerance threshold obtained in sub-step S32. If the absolute value of the offset is greater than the offset tolerance threshold, a "offset exceeded" judgment result is generated; otherwise, a "offset not exceeded" judgment result is generated. This offset exceeded judgment result is used to determine whether intervention in the welding process is required.

[0142] Step S34: Based on the offset over-limit judgment result and the current output status of the welding power source, obtain the welding intervention command type;

[0143] The central controller determines the type of welding intervention command based on the offset exceeding the limit judgment result and the current output status of the welding power supply. Specifically:

[0144] If the offset exceeds the limit judgment result is "offset not exceeded", then the welding intervention command type is "maintain current output";

[0145] If the offset exceeds the limit judgment result is "offset exceeds the limit" and the current output status of the welding power supply is "not output", then the welding intervention command type is "maintain current output";

[0146] If the offset exceeds the limit judgment result is "offset exceeds the limit" and the current output status of the welding power supply is "outputting", then further determine the current welding stage type:

[0147] If it is the arc initiation stage or the arc termination stage, the welding intervention command type is "gradually reduce output";

[0148] If it is a stable welding stage, the welding intervention command type is "stop output immediately".

[0149] As a result, the central controller receives the welding intervention command type.

[0150] Step S35: Based on the welding intervention command type, control the welding power supply to perform the corresponding output adjustment action;

[0151] Based on the welding intervention command type obtained in sub-step S34, the central controller sends a corresponding control signal to the welding power source:

[0152] If the welding intervention command type is "maintain current output", no adjustment signal will be sent, and the welding power supply will maintain the current output state.

[0153] If the welding intervention command type is "gradually reduce output", then a linear attenuation command is sent to control the welding power supply to uniformly reduce the output current from the current value to zero within 200 milliseconds;

[0154] If the welding intervention command type is "stop output immediately", then an emergency stop command is sent to control the welding power supply to cut off all welding energy output within 10 milliseconds.

[0155] After performing the above operations, the welding power source returns an execution completion signal to the central controller.

[0156] Step S36: Update the running status flag of the welding operation based on the completion signal of the output adjustment action;

[0157] After receiving the completion signal from the welding power source in sub-step S35, the central controller updates the operating status flag of the welding operation according to the welding intervention command type:

[0158] If the command is "Maintain current output", the running status flag will remain "Running".

[0159] If the execution is "gradually reduce output" or "immediately stop output", the running status flag will be updated to "paused".

[0160] This running status flag is used in subsequent steps (such as step S6) to determine whether entry into the next welding area is permitted, ensuring system status consistency.

[0161] It is understood that in this embodiment, the current welding stage type is obtained based on sub-step S31, the corresponding offset tolerance threshold is obtained in sub-step S32 based on the current welding stage type and the offset, the offset exceeding the limit judgment result is obtained in sub-step S33 based on the offset and the offset tolerance threshold, the welding intervention command type is obtained in sub-step S34 based on the offset exceeding the limit judgment result and the current output state of the welding power supply, the welding intervention command type is obtained in sub-step S35 based on the welding intervention command type, the welding power supply is controlled to perform the corresponding output adjustment action, and the operation status flag of the welding operation is updated in sub-step S36 based on the completion signal of the output adjustment action. Through the coordinated execution of the above six sub-steps, the central controller realizes stage-adaptive and state-aware hierarchical intervention for welding deviation behavior, which not only avoids accidental welding stoppage caused by slight jitter in sensitive stages such as arc initiation or arc termination, but also ensures strict control over trajectory deviation in critical stable welding stages. Thus, without sacrificing the smoothness of operation, it effectively solves the welding deviation problem caused by unstable manual operation in the background technology, and significantly improves the reliability and process consistency of the handheld welding process.

[0162] It is understood that this embodiment uses a central controller as the main body to perform graded welding intervention control based on offset. Before entering this embodiment, the system has obtained the offset through step S2 of claim 1, and obtained the current output status of the welding power supply (including two states: "not outputting" and "outputting") through the welding power supply status monitoring module. At the same time, the welding stage information of the current welding operation is determined through the welding torch trigger signal and the timer.

[0163] In some embodiments, step S4 includes the following sub-steps:

[0164] Step S41: Based on the boundary identification information, obtain the closed polygonal outline of the preset welding area;

[0165] The central controller retrieves the boundary identification information identified and stored in step S1 of claim 1. This information consists of multiple coordinate points arranged in a clockwise or counterclockwise order. The central controller connects these coordinate points sequentially, forming a closed polygonal outline. This closed polygonal outline completely surrounds the area to be welded, and its interior does not contain any non-welding areas. Here, "closed polygonal outline" refers to a closed polygonal shape composed of at least three vertices, used to precisely define the geometric range of the predetermined welding area.

[0166] Step S42: Based on the closed polygonal contour, divide the preset welding area into multiple non-overlapping sub-region units;

[0167] The central controller, based on a closed polygonal contour, employs a regular meshing method to divide the preset welding area into multiple non-overlapping sub-region units. Specifically, the system constructs a square unit mesh with a side length of 2 mm within the minimum bounding rectangle of the closed polygonal contour. Then, it removes all mesh units that do not intersect with the closed polygonal contour, retaining mesh units that are completely or partially located within the contour, and treating each retained mesh unit as a sub-region unit. The resulting multiple sub-region units do not overlap, and their union covers the entire preset welding area. Here, a "sub-region unit" refers to a square area with a side length of 2 mm, serving as the basic unit for determining weld coverage.

[0168] Step S43: Based on the real-time images continuously acquired by the camera during the welding process, obtain the time series of the position of the indicator light center;

[0169] During the welding process, the central controller extracts the position of the indicator light center from the real-time image captured by the camera every 50 milliseconds, and records the extracted horizontal and vertical coordinates in chronological order to form an ordered list. This ordered list is the time series of the indicator light center position. Here, the "time series of the indicator light center position" refers to multiple two-dimensional coordinate points arranged chronologically to reflect the spatial trajectory of the welding torch's point of action over time.

[0170] Step S44: Based on the time series of the center position of the indicator light, obtain the actual trajectory point set of the welding torch's point of action;

[0171] The central controller extracts all coordinate points from the time sequence of the indicator light center position obtained in sub-step S43, removes the time attribute, and retains only the spatial coordinates, forming an unordered set. This set is the actual trajectory point set of the welding torch's point of action. Here, the "actual trajectory point set" refers to the set of all spatial positions where the indicator light center has appeared during the welding process, representing the area actually acted upon by the welding torch.

[0172] Step S45: Based on the actual trajectory point set and multiple non-overlapping sub-region units, obtain the set of sub-region units covered by the trajectory points;

[0173] The central controller iterates through each trajectory point in the actual trajectory point set obtained in sub-step S44, and for each sub-region unit obtained in sub-step S42, determines whether the trajectory point is located inside or on the boundary of that sub-region unit. If a sub-region unit contains at least one trajectory point, then that sub-region unit is marked as "covered". Finally, the central controller collects all marked sub-region units to form a set of sub-region units covered by trajectory points.

[0174] Step S46: Based on the set of sub-region units covered by the trajectory points and multiple non-overlapping sub-region units, obtain the number of uncovered sub-region units;

[0175] The central controller counts the total number of sub-region units generated in sub-step S42, denoted as N; then it counts the number of units in the set of sub-region units covered by trajectory points obtained in sub-step S45, denoted as M; finally, it calculates the number of uncovered sub-region units, which is equal to N minus M. This value is the number of uncovered sub-region units.

[0176] Step S47: Based on the number of uncovered sub-region units and the preset coverage integrity threshold, obtain the region coverage determination result;

[0177] The central controller presets a coverage integrity threshold to be five percent of the total number of sub-region units, rounded down. For example, if the total number of sub-region units is 200, the coverage integrity threshold is 10. The central controller compares the number of uncovered sub-region units obtained in sub-step S46 with the coverage integrity threshold: if the number of uncovered sub-region units is less than or equal to the coverage integrity threshold, the area coverage determination result is "complete coverage"; otherwise, the area coverage determination result is "incomplete coverage".

[0178] Step S48: Based on the area coverage determination result, determine whether the current welding progress has completely covered the preset welding area;

[0179] Based on the area coverage determination result obtained in sub-step S47, the central controller generates a completion status indicating whether the current welding progress completely covers the preset welding area: if the area coverage determination result is "complete coverage", the completion status is "completed"; if the area coverage determination result is "incomplete coverage", the completion status is "not completed". This completion status will be used in step S6 of claim 1 as one of the criteria for determining whether to allow entry into the next welding area.

[0180] It is understood that in this embodiment, a closed polygonal outline of a preset welding area is obtained based on sub-step S41. Based on this closed polygonal outline, multiple non-overlapping sub-region units are obtained in sub-step S42. Based on real-time images continuously acquired by the camera during the welding process, a time series indicating the position of the light center is obtained in sub-step S43. Based on this time series, the actual trajectory point set of the welding torch's action point is obtained in sub-step S44. Based on this actual trajectory point set and multiple non-overlapping sub-region units, a set of sub-region units covered by the trajectory points is obtained in sub-step S45. Based on this set of sub-region units covered by the trajectory points and multiple non-overlapping sub-region units, in sub-step S46... S46 obtains the number of uncovered sub-region units. Based on the number of uncovered sub-region units and the preset coverage integrity threshold, the region coverage determination result is obtained in sub-step S47. Finally, based on the region coverage determination result, the completion status of whether the current welding progress fully covers the preset welding area is obtained in sub-step S48. Through the deterministic geometric division and set coverage judgment of the above eight sub-steps, the central controller realizes the objective, quantifiable, and interference-resistant automatic identification of the welding area completion degree, effectively solving the problem of missed welding caused by the inability to ensure complete welding of the area due to manual operation in the background technology, and providing a reliable status basis for forced sequential welding control.

[0181] It is understood that this embodiment uses a central controller as the main body to perform an objective determination of whether the preset welding area is completely covered. Before entering this embodiment, the system has obtained boundary identification information through step S1 and continuously acquired the position of the center of the indicator light through step S2.

[0182] In some embodiments, step S5 includes the following sub-steps:

[0183] Step S51: Based on the reflection spectrum signal collected by the spectral sensor at the welding position of the workpiece, obtain the spectral intensity values ​​of multiple preset characteristic bands;

[0184] The central controller controls the spectral sensor to acquire the spectrum of the workpiece surface area to be welded by the welding torch, obtaining a reflected spectral signal containing wavelengths and corresponding light intensities. This reflected spectral signal covers a wavelength range of 380 nm to 780 nm. The central controller extracts the spectral intensity values ​​of three preset characteristic bands from this signal: the first preset characteristic band is 390 nm to 395 nm, with an average light intensity of 420 units; the second preset characteristic band is 510 nm to 515 nm, with an average light intensity of 350 units; and the third preset characteristic band is 650 nm to 655 nm, with an average light intensity of 280 units. Here, "preset characteristic bands" refer to narrow wavelength ranges selected in advance based on the characteristic emission peaks of common metallic materials, used to characterize the material composition.

[0185] Step S52: Based on the spectral intensity values ​​of multiple preset characteristic bands, obtain the intensity ratio sequence of each characteristic band;

[0186] The central controller divides the spectral intensity values ​​of the three preset characteristic bands obtained in sub-step S51 pairwise in a fixed order to generate an intensity ratio sequence. Specifically, the calculation is as follows: the spectral intensity value of the first preset characteristic band is divided by the spectral intensity value of the second preset characteristic band to obtain the first ratio; the spectral intensity value of the second preset characteristic band is divided by the spectral intensity value of the third preset characteristic band to obtain the second ratio; the first and second ratios are then arranged in order to form the intensity ratio sequence. In this example, the first ratio is 420 divided by 350, which equals 1.2, and the second ratio is 350 divided by 280, which equals 1.25. Therefore, the intensity ratio sequence is [1.2, 1.25]. Here, the "intensity ratio sequence" refers to an ordered list of two ratios arranged in a fixed order, used to eliminate the influence of light source intensity fluctuations on the absolute value.

[0187] Step S53: Based on the strength ratio sequence and the pre-stored material-ratio rule library, obtain the material identifier of the successful match;

[0188] The central controller calls a pre-stored material-ratio rule base, which contains multiple deterministic matching rules, such as: "If the first ratio in the strength ratio sequence is greater than or equal to 1.1 and less than or equal to 1.3, and the second ratio is greater than or equal to 1.2 and less than or equal to 1.4, then the successfully matched material identifier is stainless steel"; "If the first ratio is greater than or equal to 0.8 and less than or equal to 1.0, and the second ratio is greater than or equal to 0.9 and less than or equal to 1.1, then the successfully matched material identifier is carbon steel." The central controller compares the strength ratio sequence [1.2, 1.25] obtained in sub-step S52 with each rule one by one, and finds that it meets the matching conditions for stainless steel, thus obtaining the successfully matched material identifier as stainless steel. Here, "successfully matched material identifier" refers to the material type name that uniquely matches the current strength ratio sequence.

[0189] Step S54: Based on the successfully matched material identifier, obtain the corresponding welding parameter configuration table;

[0190] Based on the successfully matched material identifier "stainless steel" obtained in sub-step S53, the central controller searches for the corresponding record in the welding parameter configuration database to obtain the corresponding welding parameter configuration table. This table contains combinations of process parameters suitable for stainless steel.

[0191] Step S55: Based on the welding parameter configuration table, obtain the target output current value and the target welding speed value;

[0192] The central controller reads the target output current value and target welding speed value from the welding parameter configuration table obtained in sub-step S54. In this example, the target output current value is 180 amperes, and the target welding speed value is 0.8 meters per minute. Here, "target output current value" refers to the expected output current value of the welding power supply set to ensure the penetration depth of stainless steel; "target welding speed value" refers to the expected speed of the welding torch or workpiece movement.

[0193] Step S56: Based on the actual output current value and target output current value of the current welding power source, obtain the current adjustment direction and adjustment step size;

[0194] The central controller reads the current actual output current value of the welding power supply, assuming it is 150 amps. It compares this with the target output current value of 180 amps obtained in sub-step S55: since 150 amps is less than 180 amps, the current adjustment direction is "increase"; the adjustment step size is set to a fixed value of 5 amps. Therefore, the central controller determines that the current adjustment direction is "increase" and the adjustment step size is 5 amps.

[0195] Step S57: Based on the actual moving speed value of the current welding actuator and the target welding speed value, obtain the speed adjustment direction and adjustment step size;

[0196] The central controller reads the current actual moving speed of the welding actuator, assuming it to be 1.0 m / min. It compares this speed with the target welding speed of 0.8 m / min obtained in sub-step S55. Since 1.0 m / min is greater than 0.8 m / min, the speed adjustment direction is "decreasing"; the adjustment step size is set to a fixed value of 0.05 m / min. Therefore, the central controller determines that the speed adjustment direction is "decreasing" and the adjustment step size is 0.05 m / min.

[0197] Step S58: Based on the current adjustment direction and adjustment step size and the speed adjustment direction and adjustment step size, send parameter adjustment commands to the welding power source and the welding actuator respectively;

[0198] Based on the current adjustment direction of "increase" and the adjustment step size of 5 amperes obtained in sub-step S56, the central controller sends a command to the welding power source, instructing it to increase the output current by 5 amperes in the next control cycle. Simultaneously, based on the speed adjustment direction of "decrease" and the adjustment step size of 0.05 meters per minute obtained in sub-step S57, it sends a command to the welding actuator, instructing it to decrease its movement speed by 0.05 meters per minute in the next control cycle. These parameter adjustment commands are sent to the corresponding actuators, achieving synchronous and gradual adjustment of current and speed.

[0199] It is understood that in this embodiment, based on sub-step S51, multiple preset characteristic band spectral intensity values ​​are obtained; based on these spectral intensity values, an intensity ratio sequence is obtained in sub-step S52; based on this intensity ratio sequence and a pre-stored material-ratio rule base, a successfully matched material identifier is obtained in sub-step S53; based on this successfully matched material identifier, a corresponding welding parameter configuration table is obtained in sub-step S54; based on this welding parameter configuration table, the target output current value and target welding speed value are obtained in sub-step S55; based on the actual output current value of the current welding power source and the target output current value, the current adjustment direction and adjustment step size are obtained in sub-step S56; and based on the actual movement of the current welding actuator... In sub-step S57, the dynamic speed value and the target welding speed value are used to obtain the speed adjustment direction and adjustment step size. Finally, in sub-step S58, parameter adjustment commands are sent to the welding power source and the welding actuator, respectively, based on the current adjustment direction and adjustment step size and the speed adjustment direction and adjustment step size. Through the deterministic spectral feature extraction, rule matching and progressive parameter adjustment of the above eight sub-steps, the central controller realizes the automatic identification of different workpiece materials and the adaptive adjustment of welding parameters. It effectively solves the problem of insufficient penetration or burn-through caused by the inability of manual dynamic material adaptation in the background technology. Without relying on complex models, it ensures the consistency of penetration and the reliability of the process in the handheld welding process.

[0200] It is understood that this embodiment uses a central controller as the main body to perform material identification and welding parameter adjustment based on spectral feature ratio matching. Before proceeding to this embodiment, the system has already completed the acquisition of the reflected spectral signal at the welding position of the workpiece through a spectral sensor.

[0201] In some embodiments, step S6 includes the following sub-steps:

[0202] Step S61: Based on the area identification information of the current welding area, obtain the area number of the current welding area;

[0203] The central controller matches the corresponding welding area based on the current workpiece position of the welding torch and reads the area identification information preset in the process file for that welding area. This area identification information contains an Arabic numeral number. For example, if the current welding area is identified as "Area 3," its area number is 3. Here, the "area number" refers to a unique Arabic numeral assigned to each welding area according to the sequence of the welding process, used to indicate the area's position in the overall process.

[0204] Step S62: Based on the region number, query the pre-stored region sequence dependency table to obtain the set of region numbers of the preceding region;

[0205] The central controller calls a pre-stored region sequence dependency table, which is stored in key-value pair format. For example, the preceding region of "Region 1" is an empty set, the preceding region of "Region 2" is {1}, and the preceding region of "Region 3" is {1, 2}. Based on the region number 3 obtained in sub-step S61, the central controller queries this table and finds the set of preceding region numbers as {1, 2}. Here, the "set of preceding region numbers" refers to the set of region numbers that must be welded before the current region begins.

[0206] Step S63: Based on the set of region numbers of the preceding regions, obtain the completion status flag of each preceding region;

[0207] The central controller traverses the set of region numbers {1, 2} obtained in sub-step S62, and sequentially reads the completion status flags corresponding to region 1 and region 2. Assume the completion status flag for region 1 is "completed" and the completion status flag for region 2 is "completed". Here, the "completion status flag" refers to a status flag maintained separately by the system for each welding region, with a value of "completed" or "not completed", written by the determination result of step S4 in claim 1.

[0208] Step S64: Based on the completion status flags of each preceding region, obtain the overall completion determination result of the preceding region;

[0209] The central controller checks the completion status flags of all preceding regions obtained in sub-step S63: if the completion status flag of each region in the set is "completed", then the overall completion determination result of the preceding regions is "yes"; otherwise, it is "no". In this example, both regions 1 and 2 have been completed, so the overall completion determination result of the preceding regions is "yes".

[0210] Step S65: Based on the completion status of the current welding area obtained in step S4 of claim 1 and the penetration depth compliance signal generated in step S5 of claim 1, obtain the comprehensive compliance judgment result of the current area;

[0211] The central controller calls the completion status of the current welding area output in step S4 of claim 1 (assumed to be "completed") and the penetration depth compliance signal generated in step S5 (assumed to be "valid"), and performs a logical AND judgment: the overall compliance judgment result of the current area is "yes" only if both are satisfied; otherwise, it is "no". In this example, both are satisfied, so the overall compliance judgment result of the current area is "yes".

[0212] Step S66: Based on the overall completion judgment result of the previous region and the comprehensive compliance judgment result of the current region, obtain the region unlocking condition satisfaction judgment result;

[0213] The central controller performs a logical AND operation on the overall completion judgment result "yes" of the preceding region obtained in sub-step S64 and the comprehensive compliance judgment result "yes" of the current region obtained in sub-step S65. If both are "yes", the region unlocking condition is satisfied and the judgment result is "satisfied"; otherwise, it is "not satisfied". In this example, the judgment result is "satisfied".

[0214] Step S67: Based on the determination result that the area unlocking conditions are met, generate the welding permission status of the next welding area;

[0215] The central controller generates the welding permission status of the next welding area based on the region unlocking condition satisfaction determination result obtained in sub-step S66: if the determination result is "satisfied", the welding permission status of the next welding area (i.e., area 4) is set to "unlocked"; otherwise, it is set to "locked". In this example, the welding permission status of area 4 is set to "unlocked".

[0216] Step S68: Based on the welding permission status, output area switching prompt information to the display screen and send an area access control signal to the welding power source;

[0217] Based on the welding permission status "unlocked" generated in sub-step S67, the central controller sends a command to the display screen, showing the message "Area 4 unlocked, welding can begin." Simultaneously, it sends an area access control signal to the welding power supply. This signal is a high-level enable signal, allowing the welding power supply to start output upon receiving a welding torch trigger command. If the permission status is "locked," the message is "Please complete the current area," and the area access control signal is a low-level disable signal; the welding power supply will ignore the welding torch trigger command.

[0218] It is understood that in this embodiment, the region number of the current welding area is obtained based on sub-step S61; based on this region number, the set of region numbers of the preceding regions is obtained in sub-step S62; based on this set of region numbers of the preceding regions, the completion status flag of each preceding region is obtained in sub-step S63; based on the completion status flag of each preceding region, the overall completion judgment result of the preceding regions is obtained in sub-step S64; based on the completion status of the current welding area obtained in step S4 of claim 1 and the penetration depth compliance signal generated in step S5 of claim 1, the comprehensive compliance judgment result of the current region is obtained in sub-step S65; and based on the overall completion judgment result of the preceding regions and the comprehensive compliance judgment result of the current region, the following steps are taken: Step S66 obtains the determination result that the area unlocking condition is met. Based on the determination result, in sub-step S67, the welding permission status of the next welding area is generated. Finally, based on the welding permission status, in sub-step S68, the area switching prompt information is output to the display screen and the area access control signal is sent to the welding power source. Through the coordinated execution of the above eight sub-steps, the central controller constructs a forced process control system based on the triple constraints of process sequence, area integrity and penetration depth quality. This effectively solves the problems of missed welding and incomplete welding caused by arbitrary skipping of welding or entering the next process before completing the area in the background technology. It realizes the orderly, standardized and quality closed-loop control of welding operations.

[0219] It is understood that this embodiment uses a central controller as the main body to perform welding permission control based on regional sequence dependency and multi-condition judgment. Before entering this embodiment, the system has obtained the completion status of the current welding area through step S4 of claim 1, and generated a penetration depth compliance signal through step S5; at the same time, the process sequence information of each welding area on the workpiece has been pre-entered into the system.

[0220] In some embodiments, the control method further includes the following steps:

[0221] Step S81: Obtain the voltage fluctuation amplitude based on the welding arc voltage signal acquired by the voltage sensor;

[0222] The central controller acquires voltage signals from both ends of the welding arc in real time using voltage sensors. It records the maximum and minimum voltage values ​​within a continuous 100-millisecond time window and calculates the difference between them to obtain the voltage fluctuation amplitude. For example, if the maximum value is 28 volts and the minimum value is 23 volts, the voltage fluctuation amplitude is 5 volts. Here, "voltage fluctuation amplitude" refers to the difference between the maximum and minimum arc voltage values ​​within a fixed time window, used to reflect arc stability.

[0223] Step S82: Based on the arc sound signal collected by the microphone, obtain the main peak frequency of the sound spectrum;

[0224] The central controller acquires the sound signal generated by the electric arc through a microphone and performs a Fast Fourier Transform (FFT) on the signal to obtain a frequency-energy spectrum. The central controller then locates the frequency with the highest energy within the range of 3000 Hz to 6000 Hz, designating it as the dominant frequency of the sound spectrum. In this example, this frequency is 4200 Hz. Here, the "dominant frequency of the sound spectrum" refers to the single frequency value with the highest energy in the electric arc sound spectrum, and its variation is related to the dynamic behavior of the molten pool.

[0225] Step S83: Based on continuous imaging of the molten pool area using an infrared thermal imager, obtain the highest temperature value of the molten pool;

[0226] The central controller directs the infrared thermal imager to continuously image the current molten pool area, with each frame containing the temperature values ​​of multiple pixels. The central controller extracts the temperature values ​​of all pixels within the molten pool area and identifies the maximum value as the highest temperature of the molten pool. In this example, this value is 1750 degrees Celsius. Here, the "highest temperature of the molten pool" refers to the highest instantaneous temperature measured on the surface area of ​​the molten pool, reflecting whether the heat input is sufficient.

[0227] Step S84: Based on the voltage fluctuation amplitude and the preset first melting depth correlation rule, obtain the first melting depth status flag;

[0228] The central controller calls the pre-stored first melting depth association rule, which stipulates: "If the voltage fluctuation amplitude is less than or equal to 6 volts, the first melting depth status flag is 'compliant'; otherwise, it is 'non-compliant'." The central controller compares the voltage fluctuation amplitude of 5 volts obtained in sub-step S81 with this rule. Since 5 volts is less than 6 volts, the first melting depth status flag is "compliant".

[0229] Step S85: Based on the main peak frequency of the sound spectrum and the preset second melting depth correlation rule, obtain the second melting depth status flag;

[0230] The central controller invokes a pre-stored second melting depth association rule, which stipulates: "If the main peak frequency of the sound spectrum is greater than or equal to 4000 Hz and less than or equal to 4500 Hz, the second melting depth status flag is 'compliant'; otherwise, it is 'non-compliant'." The central controller substitutes the main peak frequency of the sound spectrum, 4200 Hz, obtained in sub-step S82 into this rule. Since 4200 Hz is within the range of 4000 to 4500 Hz, the second melting depth status flag is "compliant".

[0231] Step S86: Based on the highest temperature value of the molten pool and the preset third melt depth association rule, obtain the third melt depth status flag;

[0232] The central controller calls the pre-stored third melt depth association rule, which stipulates: "If the highest temperature of the molten pool is greater than or equal to 1700 degrees Celsius and less than or equal to 1800 degrees Celsius, the third melt depth status flag is 'meets the standard'; otherwise, it is 'not met'." The central controller substitutes the highest molten pool temperature value of 1750 degrees Celsius obtained in sub-step S83 into this rule. Since 1750 degrees Celsius is within the range of 1700 to 1800 degrees Celsius, the third melt depth status flag is "meets the standard".

[0233] Step S87: Based on the first melt depth status flag, the second melt depth status flag, and the third melt depth status flag, when all three are "compliant", generate a melt depth compliance signal;

[0234] The central controller checks the first melt depth status flag, the second melt depth status flag, and the third melt depth status flag obtained in sub-steps S84, S85, and S86, respectively. A melt depth compliance signal is generated only if all three are "compliant," and its value is "valid." If any one is "non-compliant," the melt depth compliance signal is "invalid." In this example, all three are "compliant," therefore the melt depth compliance signal is "valid."

[0235] Step S88: When the penetration depth compliance signal is invalid, the area unlocking operation in step S6 of claim 1 is prohibited, and the current welding area is maintained in a locked state;

[0236] The central controller determines whether the penetration depth compliance signal generated in sub-step S87 is "invalid". If it is "invalid", the area unlocking operation in step S6 of claim 1 is prohibited, that is, the current welding area is kept locked, and even if the completion status is "completed", entry into the next area is not allowed; at the same time, a prompt is sent to the display screen: "Penetration depth not met, please re-weld the current area". In this example, because the signal is "valid", this prohibition operation is not triggered.

[0237] It is understood that in this embodiment, the voltage fluctuation amplitude is obtained in sub-step S81 based on the welding arc voltage signal collected by the voltage sensor; the main peak frequency of the sound spectrum is obtained in sub-step S82 based on the arc sound signal collected by the microphone; the highest temperature value of the molten pool is obtained in sub-step S83 based on the continuous imaging of the molten pool area by the infrared thermal imager; the first molten pool state flag is obtained in sub-step S84 based on the voltage fluctuation amplitude and the preset first molten pool depth association rule; the second molten pool state flag is obtained in sub-step S85 based on the main peak frequency of the sound spectrum and the preset second molten pool depth association rule; the third molten pool state flag is obtained in sub-step S86 based on the highest molten pool temperature value and the preset third molten pool depth association rule; and the first molten pool state flag is obtained in sub-step S85 based on the first molten pool depth association rule. In sub-step S87, the first, second, and third penetration depth status flags generate a penetration depth compliance signal. Finally, in sub-step S88, if the penetration depth compliance signal is invalid, the region unlocking operation is prohibited and the current welding region is maintained in a locked state. Through the fusion and logical AND judgment of multi-physics field signals in the above eight sub-steps, the central controller realizes non-intrusive, real-time, and highly reliable closed-loop verification of penetration depth quality. This effectively solves the problem in the background technology that relying solely on a single current parameter cannot truly reflect the penetration state. It ensures that penetration depth compliance is only recognized when the three conditions of arc stability, normal acoustic characteristics, and sufficient heat input are met simultaneously, thereby significantly improving the internal quality consistency and process reliability of handheld welding.

[0238] It is understood that this embodiment uses a central controller as the main body to perform a joint determination of the weld depth state based on three signals: arc voltage, arc sound, and molten pool temperature. Before entering this embodiment, the system has already deployed voltage sensors, microphones, and infrared thermal imagers, and is in the process of welding operations.

[0239] Finally, it should be noted that the above embodiments are only used to illustrate this application and are not intended to limit this application. Although this application has been described in detail with reference to the embodiments, those skilled in the art should understand that various combinations, modifications, or equivalent substitutions of the technical solutions of this application do not depart from the spirit and scope of the technical solutions of this application and should be covered within the scope of the claims of this application.

Claims

1. A handheld laser welding gun, characterized in that, Includes housing, beam combiner, and focusing lens; The housing has a mounting cavity; The beam combiner is installed in the mounting cavity. The beam combiner is used to combine blue laser light and infrared laser light, and to transmit the combined laser light to the focusing lens. The focusing lens is used to collimate and focus the combined laser beam, so that the combined laser beam is output outside the mounting cavity.

2. The handheld laser welding gun according to claim 1, characterized in that, The beam combiner includes a first surface and a second surface arranged opposite to each other. The middle position of the first surface is located on the transmission path of the blue laser, and the middle position of the second surface is located on the transmission path of the infrared laser. The beam combiner can reflect the blue laser and transmit the red laser.

3. The handheld laser welding gun according to claim 1 or 2, characterized in that, The handheld laser welding gun also includes a galvanometer, which is located on the transmission path of the combined laser beam and is used to reflect the combined laser beam to the focusing lens.

4. The handheld laser welding gun according to claim 1 or 2, characterized in that, The housing includes a handheld part, a main body part, and a welding part connected in sequence. The laser beam, after being combined, passes through the welding part and exits the mounting cavity to weld the workpiece. The handheld laser welding gun also includes a shielding mask, a camera, and a display screen. The shielding mask is connected to the outer wall of the housing. The handheld part and the welding part are located on opposite sides of the shielding mask. The camera is located on the side of the shielding mask facing the welding part, and the display screen is located on the side of the shielding mask facing the handheld part. The camera is used to acquire images of the welded area of ​​the workpiece and display them on the display screen.

5. A control method applied to the handheld laser welding gun as described in claim 4, characterized in that, include: S100. Based on the real-time shooting data of the camera, determine the real-time image of the welded part of the workpiece; S200: Transmit the real-time image to the display screen for real-time display.

6. The handheld laser welding gun control method according to claim 5, characterized in that, Before transmitting the real-time image to the display screen for real-time display, the following steps are also included: S210. Based on the real-time shooting data of the camera, acquire at least two frames of welding area images under different exposure times to obtain a multi-exposure image group; S220. Based on the brightness distribution characteristics of each image in the multi-exposure image group, determine the position information of the highlight area and the dark area to obtain the dynamic range mapping basis; S230. Based on the dynamic range mapping criteria, pixel-level fusion processing is performed on the multi-exposure image group to obtain a high dynamic range image that suppresses solder joint overexposure and retains dark details. S240. The high dynamic range image is used to replace the real-time image and is input to the display screen as the image to be displayed in real time.

7. The handheld laser welding gun control method according to claim 5, characterized in that, Before transmitting the real-time image to the display screen for real-time display, the following steps are also included: S310. Based on the camera being configured with a polarizing filter or a near-infrared photosensitive module, the original image of the welding area containing polarized light components or near-infrared band information is obtained, and an auxiliary image of smoke penetration is obtained. S320. Based on the differences in the scattering of polarized light or the transmission characteristics of near-infrared light by smoke particles in the smoke penetration auxiliary image, identify the spatial distribution of the smoke-covered area and obtain the smoke interference mask. S330. Based on the smoke interference mask, perform local contrast enhancement and background transmission restoration on the smoke penetration auxiliary image to obtain a smoke penetration enhanced image. S340. The real-time image is replaced by the smoke and dust penetration enhancement image, and the image to be displayed is input to the display screen for real-time display.

8. The handheld laser welding gun control method according to claim 5, characterized in that, Before transmitting the real-time image to the display screen for real-time display, the following steps are also included: S410. Calculate the gradient amplitude based on the real-time image to obtain the edge intensity distribution map of the welding area; S420. Based on the connection relationship of continuous high gradient pixels in the edge intensity distribution map, extract the candidate contour of the molten pool boundary to obtain the preliminary molten pool contour. S430. Based on the geometric closure and symmetry characteristics of the preliminary molten pool profile, non-molten pool interference edges are eliminated to obtain the accurate molten pool profile. S440. Based on the precise molten pool contour, a highlighted marker line is superimposed on the real-time image to obtain an enhanced image with contour markers; S450. The enhanced image with outline markings replaces the real-time image and is input to the display screen as the image to be displayed in real time.

9. The handheld laser welding gun control method according to claim 1 or 2, characterized in that, It also includes the following steps: Step S1: Based on the workpiece surface image captured by the camera, obtain the position information of the welding standard line and the boundary identification information of the preset welding area; Step S2: Based on the position information of the welding standard line and the center position of the indicator light projected onto the workpiece surface in real time, obtain the offset of the indicator light relative to the welding standard line; Step S3: Based on the offset, determine whether the offset exceeds the preset first threshold. If it does, control the welding power supply to stop outputting welding energy. Step S4: Based on the boundary identification information and the current position of the welding torch, determine whether the current welding progress has covered the entire preset welding area. Step S5: Based on the real-time acquisition of the reflection spectrum at the welding position of the workpiece by the spectral sensor, the current material type of the workpiece is obtained, and the corresponding welding parameter configuration table is retrieved based on the material type, thereby adjusting the output current and welding speed of the welding power supply. Step S6: Based on the completion status and the current penetration depth compliance signal, determine whether it is allowed to enter the next welding area; If the completion status is "completed" and the penetration depth meets the standard signal is valid, then the welding permission for the next welding area is unlocked; otherwise, the current area remains locked.

Images