Water-guided laser light-water coupling automatic calibration method

Abstract

The invention provides a water-guided laser light-water coupling automatic calibration method which is applied to a water-guided laser light-water coupling automatic calibration system, and the calibration method comprises the steps: obtaining a relative position image of a light spot generated by a laser beam and a nozzle when a laser collimation head receives the laser beam, performing edge detection on a temporary image obtained through conversion according to the relative position image to extract edge features in the temporary image, and performing circle detection on the light spot edge features to obtain a light spot detection result; detecting according to the nozzle edge features to obtain a nozzle detection result; and based on a light spot detection result and a nozzle detection result, changing a light path of the laser beam through an electric light path calibration module to adjust the position of the light spot so as to realize automatic calibration of light-water coupling. By automatically adjusting the position of the light spot, errors possibly caused by manual adjustment are eliminated, the calibration efficiency and stability are improved, and meanwhile consumption of human resources is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of water-guided laser coupling, and more specifically, to a method for automatically calibrating the light-water coupling of water-guided lasers. Background Art

[0002] In the water-guided laser processing technology, the effective coupling of the laser beam and the water jet is a key link to achieve high-quality processing. Among them, the laser beam needs to undergo precise optical path calibration to ensure that the laser spot is perfectly docked with the sapphire nozzle in the water-guided laser system. However, due to the heating effect of the laser beam on the optical components (such as the sapphire nozzle) in the optical path during the processing, the position of the spot will drift, further increasing the frequent need for optical path calibration.

[0003] Currently, the optical path calibration mainly relies on manual adjustment. The operator manually adjusts through visual observation and experience. This method has many deficiencies in practical applications. First, the calibration accuracy highly depends on the operator's experience and technical level, with a large degree of subjectivity. Second, errors are prone to occur during the manual calibration process, resulting in unstable calibration results, which in turn affect the processing accuracy. Finally, the manual calibration efficiency is low. Especially in the case of multiple calibrations, it is easy to cause a decline in production efficiency. Summary of the Invention

[0004] The problem solved by the present invention is one or more of the above-related technical problems.

[0005] To solve the above problems, the present invention provides a method for automatically calibrating the light-water coupling of water-guided lasers, which is applied to an automatic calibration system for the light-water coupling of water-guided lasers. The automatic calibration system for the light-water coupling of water-guided lasers includes a laser collimator head, a coaxial monocular vision detection module, an electric optical path calibration module, and a coupling cavity provided with a nozzle. The method for automatically calibrating the light-water coupling of water-guided lasers includes:

[0006] When the laser collimator head receives a laser beam,

[0007] Obtain an image of the relative position between the spot generated by the laser beam and the nozzle through the coaxial monocular vision detection module, and convert the relative position image to obtain a temporary image;

[0008] Perform edge detection on the temporary image to extract the edge features therein. The edge features include spot edge features and nozzle edge features;

[0009] Perform circle detection on the spot edge features to obtain a spot detection result;

[0010] Based on the improved Hough gradient circle transform, perform detection according to the nozzle edge features to obtain a nozzle detection result;

[0011] Based on the spot detection result and the nozzle detection result, the optical path of the laser beam is changed by the electric optical path calibration module to adjust the position of the spot, so as to realize the automatic calibration of the light-water coupling.

[0012] Optionally, the detection based on the improved Hough gradient circle transformation is performed according to the nozzle edge feature to obtain the nozzle detection result, including:

[0013] Determine the gradient information of each edge point in the nozzle edge feature, and construct a parameter space based on the gradient information; the gradient information includes the gradient amplitude and the gradient direction;

[0014] Determine a plurality of candidate center points according to the gradient amplitude and the gradient direction of each edge point, and vote for each candidate center point in the parameter space to generate a plurality of temporary candidate center points;

[0015] Determine the corresponding candidate circles according to each temporary candidate center point and the preset radius parameter;

[0016] Perform a coincidence degree detection on each candidate circle based on the nozzle edge feature, and use the candidate circle with the highest coincidence degree as the nozzle detection result.

[0017] Optionally, the electric optical path calibration module includes an electric deflection mirror. Based on the spot detection result and the nozzle detection result, changing the optical path of the laser by the electric optical path calibration module to adjust the position of the spot to realize the automatic calibration of the light-water coupling includes:

[0018] Adjust the electric deflection mirror to the initial position;

[0019] Control the electric deflection mirror to continuously move in a preset step length, and collect the image change data of the nozzle;

[0020] Determine the mapping relationship between the image change data and the preset step length based on a preset algorithm;

[0021] Adjust the electric deflection mirror according to the mapping relationship, the spot detection result and the nozzle detection result to change the optical path of the laser so that the spot is aligned with the image of the nozzle.

[0022] Optionally, the coaxial monocular vision detection module includes an auxiliary light source, an adjustable lens barrel and a monocular CCD camera. The monocular CCD camera is connected to the adjustable lens barrel, and the auxiliary light source is on the same optical axis as the laser beam and is used to illuminate the spot; obtaining the relative position image of the spot and the nozzle through the coaxial monocular vision detection module includes:

[0023] By adjusting the focal length of the adjustable lens barrel, the imaging plane of the monocular CCD camera is made to coincide with the plane where the nozzle is located;

[0024] Illuminate using the auxiliary light source, and collect the relative position image of the light spot and the nozzle through the monocular CCD camera.

[0025] Optionally, the converting the relative position image to obtain a temporary image includes:

[0026] Determine the segmentation threshold of the relative position image through an image processing algorithm;

[0027] Perform region segmentation on the relative position image according to the segmentation threshold to obtain the temporary image.

[0028] Optionally, the performing circle detection on the light spot edge feature to obtain a light spot detection result includes:

[0029] Count all the pixel points in the light spot edge feature, and select a preset number of the pixel points for random sample consensus detection to obtain a set of edge pixel points;

[0030] Fit the set of edge pixel points to obtain a target fitting circle, and use the target fitting circle as the light spot detection result.

[0031] Optionally, the fitting the set of edge pixel points to obtain a target fitting circle includes:

[0032] Select a preset number of the pixel points in the set of edge pixel points for fitting to obtain a current fitting circle, and use the number of edge points included in the current fitting circle as the inlier number;

[0033] Repeat the above steps. When the inlier number of the current fitting circle is higher than the inlier number of the fitting circle in the previous iteration, and the radius of the current fitting circle satisfies the preset radius range, use the current fitting circle as a temporary fitting circle, or until the preset number of iterations is reached, use the current fitting circle as the target fitting circle.

[0034] Optionally, the electric optical path calibration module further includes a motion controller, and the motion controller is connected to the electric deflection mirror; the adjusting the electric deflection mirror according to the mapping relationship, the light spot detection result and the nozzle detection result to align the image of the light spot and the nozzle includes:

[0035] Based on the mapping relationship, the light spot detection result and the nozzle detection result, determine an execution instruction through the motion controller;

[0036] According to the execution instruction, deflect by adjusting the electric deflection mirror so that the light spot is aligned with the image of the nozzle.

[0037] Optionally, the segmentation threshold includes a first segmentation threshold and a second segmentation threshold; the region segmentation of the relative position image according to the segmentation threshold to obtain the temporary image includes:

[0038] Processing the relative position image based on the first segmentation threshold and the second segmentation threshold to obtain the temporary image;

[0039] Wherein, the first segmentation threshold is used to segment the high-brightness gray area and the medium-brightness gray area in the relative position image; the second segmentation threshold is used to segment the medium-brightness gray area and the low-brightness gray area in the relative position image.

[0040] Optionally, the temporary image includes a light spot image; after converting the relative position image to obtain the temporary image, it further includes:

[0041] Performing median filtering on the light spot image, and performing opening operation on the filtered image according to a first preset circular convolution kernel to obtain a temporary light spot image;

[0042] Performing closing operation on the temporary light spot image according to a second preset circular convolution kernel to obtain a processed light spot image.

[0043] The beneficial effects of the water-guided laser light-water coupling automatic calibration method of the present invention are:

[0044] When the laser collimation head receives the laser beam, first, the coaxial monocular vision detection module is used to obtain the relative position image between the light spot generated by the laser beam and the nozzle. This process captures images in real time through the camera and generates initial data, laying a foundation for subsequent processing. The obtained relative position image is then converted to obtain a temporary image. This process aims to improve the accuracy of subsequent steps (such as edge detection), reduce the influence of background noise, and thus more effectively extract key features.

[0045] Edge detection is performed on the temporary image to extract the edge features of the light spot and the nozzle. This process uses an edge detection algorithm (such as the Canny operator) to identify the obvious boundaries in the image, especially the contours of the light spot and the nozzle. By accurately extracting the edge features, the reliability of subsequent feature recognition and circle detection is ensured, laying a foundation for the precise positioning of the light spot.

[0046] Subsequently, circle detection is performed on the extracted spot edge features to obtain the detection result of the spot. This is usually achieved through the parametric equation of a circle and corresponding algorithms (such as the Hough transform). The spot detection result of this process provides the accurate spot position for the calibration process, ensuring that the laser beam can be effectively coupled with the water jet.

[0047] Next, the improved Hough gradient circle transform is used to detect the edge features of the nozzle to obtain the detection result of the nozzle. This step ensures that the position and shape of the nozzle are accurately identified and located. Therefore, the accuracy of nozzle detection is enhanced, especially in complex backgrounds, guaranteeing the correct positioning of the nozzle and helping to balance the relationship between the laser spot and the nozzle.

[0048] Finally, according to the spot detection result and the nozzle detection result, the electric optical path calibration module automatically adjusts the spot to achieve automatic calibration of the light-water coupling. By automatically adjusting the spot position, the errors that may be brought by manual adjustment are eliminated, the calibration efficiency and stability are improved, and at the same time, the consumption of human resources is reduced.

[0049] Therefore, the automatic light-water coupling calibration method of the water-guided laser in the present invention can achieve efficient and accurate automatic calibration in a complex environment, improving the stability and processing accuracy of the light-water coupling system. The use of automation technology not only improves the processing efficiency, but also reduces the dependence on the operator's technical level, providing higher consistency and repeatability, thus making the water-guided laser processing process more reliable and efficient. Description of the Drawings

[0050] Figure 1 One of the flow schematic diagrams of an automatic light-water coupling calibration method of a water-guided laser according to an embodiment of the present invention;

[0051] Figure 2 The water-guided laser light-water coupling automatic calibration system according to an embodiment of the present invention;

[0052] Figure 3 A schematic diagram of the relative position image according to an embodiment of the present invention;

[0053] Figure 4 A schematic diagram of the temporary image according to an embodiment of the present invention;

[0054] Figure 5 A schematic diagram of the automatic coupling of the water-guided laser spot into the water jet according to an embodiment of the present invention;

[0055] Figure 6 An optical path schematic diagram of the water-guided laser light-water coupling automatic calibration system according to an embodiment of the present invention;

[0056] Figure 7 A schematic diagram of the processed spot image according to an embodiment of the present invention;

[0057] Figure 8 This is the second flowchart of the water-guided laser light-water coupling automatic calibration method according to the embodiment of the present invention.

[0058] Explanation of reference numerals:

[0059] 1 - Laser collimation head, 101 - Laser input QBH interface, 2 - Variable aperture, 3 - Electric optical path calibration module, 301 - Electric deflection mirror, 302 - Reflecting mirror, 303 - Dichroic mirror, 4 - Coaxial monocular vision detection module, 401 - Auxiliary light source, 402 - Adjustable lens barrel, 403 - Monocular CCD camera, 5 - Focusing lens, 6 - Coupling cavity. Specific embodiments

[0060] To make the above objects, features and advantages of the present invention more obvious and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Although some embodiments of the present invention are shown in the drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments described herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not used to limit the protection scope of the present invention.

[0061] The term "including" and its variants used herein are open-ended, that is, "including but not limited to"; the term "based on" is "at least partially based on"; the term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "optional embodiments". The relevant definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc. mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0062] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0063] As Figure 1 shown, a water-guided laser light-water coupling automatic calibration method provided by an embodiment of the present invention is applied to a water-guided laser light-water coupling automatic calibration system. The water-guided laser light-water coupling automatic calibration system includes a laser collimation head 1, a coaxial monocular vision detection module 4, an electric optical path calibration module 3, and a coupling cavity 6 provided with a nozzle. The water-guided laser light-water coupling automatic calibration method includes:

[0064] Step S100, when the laser collimator head 1 receives a laser beam, the coaxial monocular vision detection module 4 acquires an image of the relative position between the light spot generated by the laser beam and the nozzle, and converts the relative position image to obtain a temporary image.

[0065] Specifically, as Figure 2 shown in the structural schematic diagram of the water-guided laser light-water coupling automatic calibration system, which includes a laser collimator head 1: used to receive a laser beam and receive laser input through the laser input QBH interface 101. The variable aperture 2 is immediately behind the laser collimator head to adjust the size of the laser light spot entering the optical path. The electric optical path calibration module 3 includes: an electric deflection mirror 301: used to adjust the direction of the laser to achieve precise calibration of the optical path. A reflecting mirror 302: used to reflect the laser beam so that it propagates in the correct direction. A dichroic mirror 303 is used to separate lasers of different wavelengths to improve the flexibility and functionality of the system. The coaxial monocular vision detection module 4 includes: an auxiliary light source 401: providing additional illumination to enhance the visual detection effect. An adjustable lens barrel 402: By changing its distance from the CCD camera 403, the focal length can be adjusted, so as to maintain the clarity of the image at different working distances, that is, it allows adjustment at different viewing distances to obtain a clear image. A monocular CCD camera 403: used to capture an image of the laser light spot to achieve real-time monitoring. A focusing mirror 5 is used to focus the laser beam to ensure that the laser beam has an accurate focus within the target area. The coupling cavity 6 finally combines the laser and the imaging system to form a complete optical path system.

[0066] After the laser collimator head 1 receives the laser beam through the laser input QBH interface 101, it sends the laser beam to the variable aperture (2, and after adjustment, it is transmitted to the electric optical path calibration module 3. The electric deflection mirror 301, the reflecting mirror 302, and the dichroic mirror 303 in the electric optical path calibration module 3 jointly adjust and calibrate the optical path of the laser beam. The calibrated laser beam continues to be transmitted to the coaxial monocular vision detection module 4, where illumination is provided by the auxiliary light source 401 and focused onto the monocular CCD camera 403 via the adjustable lens barrel 402. Finally, the focusing mirror 5 and the coupling cavity 6 ensure that the laser beam is correctly focused and coupled.

[0067] And the coaxial setting in the water-guided laser light-water coupling automatic calibration system includes:

[0068] The laser collimation head receives the laser beam. After passing through the variable aperture, the laser beam continues to propagate along the same optical axis. Then, it passes through the electric optical path calibration module 3, which also maintains the same optical axis when adjusting the beam. Auxiliary light source 401: Usually designed on the path of the laser beam to illuminate the laser spot. When this light source is on the same axis as the laser beam, it can provide a clear view. Adjustable barrel 402: Through the design of the barrel, light can directly pass from the auxiliary light source to the industrial monocular CCD camera 403, ensuring distortion-free imaging. Monocular CCD camera 403: The optical path of the CCD camera is arranged coaxially with other components to ensure that the captured image directly originates from the same laser beam.

[0069] When the laser collimation head 1 receives the laser beam, the coaxial monocular vision detection module 4, equipped with a camera and an appropriate light source, is used to capture the relative position image between the spot and the nozzle in real time. This capture process ensures the acquisition of the latest image data in the laser processing environment. The image mode is usually a grayscale image, for example Figure 3 As shown, a schematic diagram of the relative position image. Its image acquisition resolution is 2448×2048, and the acquisition frequency is 15fps. This ensures high resolution and real-time acquisition of the spot and nozzle images, achieving micron-level positioning of the self-alignment system.

[0070] If there is relative motion during the acquisition process (e.g., slight vibrations of the machine tool), image stabilization technology can be used to correct the image. This technology can reduce the blur caused by motion and improve the clarity of the image.

[0071] The obtained relative position image is transformed, which usually includes, such as image enhancement: by adjusting the contrast and brightness, improving the visibility of the image and ensuring more obvious edge features. Binarization processing for edge detection and feature extraction.

[0072] After the above processing, a temporary image is generated, as Figure 4 As shown, a schematic diagram of the temporary image (the result of binarization segmentation), which is used for subsequent edge detection and feature extraction steps. This temporary image will be used as the input for subsequent algorithm analysis to ensure the accuracy of subsequent processing.

[0073] By accurately obtaining the relative position image, the entire water-guided laser processing system can be more stable during automatic calibration. Optimized initial data input will reduce the calibration error caused by data differences and improve the overall stability of the system. Moreover, through automated image acquisition and processing steps, the time and labor intensity required for manual calibration are significantly reduced, improving the automation level and consistency of processing.

[0074] Step S200: Perform edge detection on the temporary image to extract the edge features therein, where the edge features include the light spot edge features and the nozzle edge features.

[0075] Specifically, first, use the temporary image generated in step S100 as the input data. This image has been preprocessed, with high clarity and low noise level, providing a reliable basis for edge detection.

[0076] Select a suitable edge detection algorithm, such as the Canny operator, Sobel operator, or Laplacian operator, etc. For example, the Canny operator can be selected. This algorithm has strong noise suppression and edge detection capabilities. After edge detection, the edge features of the light spot and the nozzle are extracted and output in the form of a binary image for subsequent feature recognition and position localization.

[0077] By selecting an efficient edge detection algorithm, the edges of the light spot and the nozzle can be automatically recognized in a relatively complex background, ensuring that the extracted features are accurate. And multiple steps in the edge detection process (such as threshold processing and non-maximum suppression) effectively reduce false alarms and missed detections, making the extraction of the light spot and nozzle edges more precise, which is crucial for subsequent calibration. And accurate edge feature extraction provides reliable data for subsequent feature recognition and circle detection, ensuring that the light spot can be effectively docked with the nozzle, thereby improving the overall effect of water-guided laser coupling.

[0078] The above use of an automated edge detection algorithm saves the time of manual inspection, improves the efficiency of the entire calibration process, and enables the system to quickly adapt to changes in the production environment. At the same time, the automated edge detection technology reduces the dependence on the operator's technical level, ensuring that even non-professionals can effectively complete the calibration, improving the consistency and operability of the system.

[0079] Step S300: Perform circle detection on the light spot edge features to obtain the light spot detection result.

[0080] Specifically, adopt a suitable circle detection algorithm, such as RCD (Randomized Circle Detection) to detect the light spot position. This algorithm is widely used in circular detection processing in images due to its high efficiency and adaptability.

[0081] The automated circle detection process reduces human intervention and adjustment time, making the calibration process of light-water coupling faster and more efficient. This improves the operation efficiency of the entire production line.

[0082] Step S400: Based on the improved Hough gradient circle transform, detect according to the nozzle edge features to obtain the nozzle detection result.

[0083] Specifically, the improved Hough gradient circle transform is an improved method for the traditional Hough transform, aiming to improve the efficiency and accuracy of circle detection. This method combines the advantages of edge detection and Hough transform and can perform better under high noise or complex backgrounds.

[0084] The process of applying the improved Hough gradient circle transform usually includes calculating the gradient: calculating the gradient of the edge intensity map to obtain the direction and intensity of the edge information. Circle accumulation: Establish an accumulator to record possible combinations of center coordinates and radii of circles simultaneously. By combining the edge gradient information, optimize the update process of the accumulator to improve the detection accuracy rate. Finally, output the detected nozzle results, including the center coordinates and radius information of the nozzle. These parameters can be used for subsequent control and adjustment.

[0085] Based on the combination of edge features and gradient information, the improved Hough gradient circle transform can more accurately identify the circular features of the nozzle, thereby improving the detection accuracy. This process has strong adaptability to noise and background interference and can still maintain good detection performance in a relatively complex environment, thus enhancing the reliability of the system.

[0086] Moreover, by combining the advantages of edge detection and Hough transform, unnecessary calculations are reduced, and the overall detection speed is increased, providing support for real-time or online detection. The improved algorithm can maintain high efficiency and adaptability in the detection of nozzles of different sizes and shapes, and has good effects on the rapid detection of new nozzles and adaptation to new environments.

[0087] Step S500, based on the spot detection result and the nozzle detection result, change the optical path of the laser beam through the electric optical path calibration module 3 to adjust the position of the spot, so as to realize the automatic calibration of the light-water coupling.

[0088] Specifically, after obtaining the positions of the spot and the nozzle, calculate the relative position deviation between the center of the spot and the center of the nozzle. This can usually be achieved through simple geometric calculations (such as Euclidean distance measurement). The purpose of this step is to determine the specific direction and distance by which the light beam needs to be offset to ensure the best coupling effect between the two.

[0089] Input the calculated deviation value into the electric optical path calibration module 3. This module may include a stepper motor, a servo motor, or other driving devices, which are responsible for adjusting the angles and positions of the optical path system (such as mirrors or lenses). According to the deviation value, control the electric calibration module to make precise adjustments and correct the position in real time according to the feedback information to ensure fast and accurate calibration.

[0090] After the adjustment is completed, the light-water coupling state can be verified, which can be achieved by real-time monitoring of parameters such as coupling efficiency.

[0091] Based on the accurate detection results of the light spot and the nozzle, highly accurate light-water coupling can be achieved, significantly improving the efficiency and quality of laser processing. This process reduces manual intervention through automatic calibration, improves the automation level of operation, reduces errors caused by human operation, and thus improves the overall stability of the system. At the same time, the deviation is calculated in real time and adjusted immediately, enabling the system to quickly adapt to the dynamically changing working environment and enhancing the overall production efficiency.

[0092] Through the automatic adjustment function of the electric optical path calibration module 3, the operation process is simplified, enabling even unprofessionally trained operators to easily manage and adjust the system, thus reducing the requirements for operating skills.

[0093] In this embodiment, when the laser collimator head receives the laser beam, first, the coaxial monocular vision detection module 4 is used to obtain the relative position image between the light spot generated by the laser beam and the nozzle. This process captures images in real time through the camera and generates initial data, laying the foundation for subsequent processing. The obtained relative position image is then converted to obtain a temporary image. This process aims to improve the accuracy of subsequent steps (such as edge detection), reduce the influence of background noise, and thus more effectively extract key features.

[0094] Edge detection is performed on the temporary image to extract the edge features of the light spot and the nozzle. This process uses an edge detection algorithm (such as the Canny operator) to identify the obvious boundaries in the image, especially the contours of the light spot and the nozzle. By accurately extracting the edge features, the reliability of subsequent feature recognition and circle detection is ensured, laying the foundation for the precise positioning of the light spot.

[0095] Subsequently, circle detection is performed on the extracted edge features of the light spot to obtain the detection result of the light spot. This is usually achieved through the parametric equation of the circle and the corresponding algorithm (such as the Hough transform). The detection result of the light spot in this process provides the accurate position of the light spot for the calibration process, ensuring that the laser beam can effectively couple with the water jet.

[0096] Next, the improved Hough gradient circle transform is used to detect the edge features of the nozzle to obtain the detection result of the nozzle. This step ensures that the position and shape of the nozzle are accurately identified and located. Therefore, the accuracy of nozzle detection is enhanced, especially in a complex background, ensuring the correct positioning of the nozzle, which helps to balance the relationship between the laser light spot and the nozzle.

[0097] Finally, according to the spot detection result and the nozzle detection result, the electric optical path calibration module 3 automatically adjusts the spot to achieve automatic calibration of the light-water coupling. By automatically adjusting the spot position, the errors that may be brought by manual adjustment are eliminated, the calibration efficiency and stability are improved, and at the same time, the consumption of human resources is reduced.

[0098] Therefore, the automatic light-water coupling calibration method of the present invention can achieve efficient and accurate automatic calibration in a complex environment, improving the stability and processing accuracy of the light-water coupling system. The use of automation technology not only improves the processing efficiency, but also reduces the dependence on the operator's technical level, providing higher consistency and repeatability, thus making the water-guided laser processing process more reliable and efficient.

[0099] Optionally, the detection based on the improved Hough gradient circle transform according to the nozzle edge feature to obtain the nozzle detection result includes:

[0100] Determine the gradient information of each edge point in the nozzle edge feature, and construct a parameter space based on this gradient information; the gradient information includes the gradient amplitude and the gradient direction;

[0101] Determine a plurality of candidate center points according to the gradient amplitude and the gradient direction of each edge point, and vote for each candidate center point in the parameter space to generate a plurality of temporary candidate center points;

[0102] Determine the corresponding candidate circles according to each temporary candidate center point and the preset radius parameter;

[0103] Perform a coincidence degree detection on each candidate circle based on the nozzle edge feature, and use the candidate circle with the highest coincidence degree as the nozzle detection result.

[0104] Specifically, the detection of the nozzle is performed using the improved Hough gradient circle transform. The traditional Hough gradient circle transform traverses all edges, calculates the pixel gradient direction of each edge point, and forms intersection points in the space, and these intersection points are regarded as candidate center points. At the same time, the gradient directions of the intersection points are accumulated as the voting quantity, and the more the voting quantity, the greater the possibility that the intersection point is considered to be the best center point; when all edge points are processed, there will be some points in the Hough space with a high voting quantity. These points with a high voting quantity are regarded as potential center points. By performing a local maximum detection on these voting points, the best center point position and the corresponding radius can be found.

[0105] However, due to the existence of some background noise interference in the nozzle edge features (nozzle edge detection results) and the inability to completely eliminate it, the background noise edges also participate in the Hough gradient circle transformation, interfering with the voting results and causing deviations in the voting results of the candidate circles. To solve this problem, based on the candidate circle results (candidate circles) obtained from the traditional Hough gradient circle transformation, it is necessary to calculate the coincidence degree between the edges of the candidate circles and the nozzle edge detection results (nozzle edge features). The candidate circle with the highest coincidence degree will be identified as the best circle and fed back as the final detection result. As Figure 5 shown, the schematic diagram of the automatic coupling of the water-guided laser spot into the water jet Figure 5 in which the image before the automatic calibration of the driving electric deflection mirror contains the nozzle detection results.

[0106] Optionally, the electric optical path calibration module 3 includes an electric deflection mirror 301. Based on the spot detection result and the nozzle detection result, the electric optical path calibration module 3 changes the optical path of the laser to adjust the position of the spot to achieve automatic calibration of the light-water coupling, including:

[0107] Adjust the electric deflection mirror 301 to the initial position, control the electric deflection mirror 301 to move continuously at a preset step size, and collect the image change data of the nozzle;

[0108] Determine the mapping relationship between the image change data and the preset step size based on a preset algorithm;

[0109] According to the mapping relationship, the spot detection result, and the nozzle detection result, adjust the electric deflection mirror 301 to change the optical path of the laser to align the spot with the image of the nozzle.

[0110] Optionally, the electric optical path calibration module 3 further includes a motion controller, which is connected to the electric deflection mirror. The adjustment of the electric deflection mirror 301 according to the mapping relationship, the spot detection result, and the nozzle detection result to align the spot with the image of the nozzle includes:

[0111] Based on the mapping relationship, the spot detection result, and the nozzle detection result, determine an execution instruction through the motion controller;

[0112] According to the execution instruction, deflect the electric deflection mirror 301 to align the spot with the image of the nozzle.

[0113] Specifically, as Figure 2As shown in the figure, the laser collimation head includes an overall beam coupling device, a self-collimation mirror group, and a laser input QBH interface 101 that can be directly connected to an optical fiber. The overall beam coupling device is mainly responsible for effectively coupling the laser beam from the laser source into the subsequent optical system to ensure the light transmission efficiency. The output end of the overall beam coupling device is connected to the self-collimation mirror group. In the design, the overall coupling device ensures that the beam is transmitted to the subsequent optical elements at an appropriate angle and size. The self-collimation mirror group is used to correct and adjust the path of the laser beam to ensure the straight-line transmission of the beam in space and maintain the best focusing and divergence characteristics. The input end of the self-collimation mirror group receives the beam from the overall beam coupling device and further collimates it. The output end of the self-collimation mirror group should be connected to the laser input QBH interface 101 that can be directly connected to an optical fiber to ensure the efficient transmission of the beam to the optical fiber. The laser input QBH interface 101 is a standardized interface that allows the laser to be effectively transmitted through the optical fiber and forms a connection with other optical or laser systems. The structural design of the QBH interface allows it to directly receive the beam from the self-collimation mirror group to ensure seamless docking. In many systems, the QBH interface may also include some mechanical and optical alignment mechanisms to reduce the optical loss during the connection process and ensure the efficient transmission of the laser beam. As Figure 6 shown in the schematic diagram of the optical path of the water-guided laser light-water coupling automatic calibration system.

[0114] As Figure 2 shown in the figure, the electric optical path calibration module 3 further includes a mirror 302, a dichroic mirror 303, and a motion controller. The electric deflection mirror 301 is used to adjust the direction of the beam, accurately direct the beam to the nozzle, and realize the relative position adjustment between the light spot and the nozzle. The electric deflection mirror 301 is directly connected to the motion controller to receive control commands. When the motion controller issues a command, the electric deflection mirror deflects at a set angle to change the propagation direction of the beam. That is, the electric deflection mirror 301 can deflect according to the motion controller command to adjust the relative position between the light spot and the nozzle and complete the self-collimation of the optical path. The mirror 302 is used to make the first 90° turn of the beam and introduce the beam into the dichroic mirror. The mirror 302 is connected to the output end of the electric deflection mirror 301. After passing through the electric deflection mirror, the beam first reaches the mirror and is reflected to the dichroic mirror 303. The dichroic mirror 303 is used to perform specific optical processing on the beam (such as phase modulation, dispersion, etc.) and realize the second 90° turn. The dichroic mirror 303 is placed opposite to the mirror 302 to receive the beam from the mirror. The design of the mirror ensures that the beam turns smoothly after passing through the dichroic mirror 303 and maintains the quality and consistency of the beam. The motion controller is responsible for controlling the movement of the electric deflection mirror to ensure the precise alignment between the beam and the nozzle. The motion controller is connected to the electric deflection mirror 301 to send control signals for deflection adjustment.

[0115] Specifically, heat is generated during the laser processing, which can cause thermal deformation of the optical components. In addition, when replacing the components of the water-jet guiding processing system, it will cause a position change in the Z-axis direction. Therefore, it is necessary to adjust the adjustable lens barrel 402 to ensure that the focal plane of the camera (monocular CCD camera 403) is always located in the plane where the nozzle and the light spot are located, so as to achieve clear imaging. This results in a change in the mapping relationship between the image pixels collected by the camera and the step size of the electric deflection mirror 301 during each processing. Therefore, it is necessary to initialize the electric deflection mirror 301 or its corresponding motor. Then, control the electric deflection mirror 301 or the motor to continuously move in a preset step size, and collect the image change data of the nozzle; determine the mapping relationship between the image change data and the preset step size based on a preset algorithm; and adjust the electric deflection mirror 301 in combination with the light spot detection result and the nozzle detection result to align the light spot with the image of the nozzle.

[0116] In some embodiments, the motor is controlled to continuously move 8 times, with each movement being 50 steps, to detect the pixel changes (image change data) of the nozzle in the X and Y directions. Using a preset algorithm, such as the least squares method, to fit the mapping relationship between the motor step size and the pixels. Finally, as Figure 5 shown, by adjusting the electric deflection mirror 301, the light spot is moved and aligned with the nozzle, thereby realizing the self-alignment function of the water-jet guiding laser processing optical path.

[0117] Optionally, as Figure 2 shown, the coaxial monocular vision detection module 4 includes an auxiliary light source 401, an adjustable lens barrel 402, and a monocular CCD camera 403. The monocular CCD camera 403 is connected to the adjustable lens barrel 402. The auxiliary light source 401 is on the same optical axis as the laser beam and is used to illuminate the light spot; obtaining the relative position image of the light spot and the nozzle through the coaxial monocular vision detection module 4 includes:

[0118] Adjusting the focal length of the adjustable lens barrel 402 so that the imaging plane of the monocular CCD camera 403 coincides with the plane where the nozzle is located;

[0119] Using the auxiliary light source 401 for illumination, and collecting the relative position image of the light spot and the nozzle through the monocular CCD camera 403.

[0120] Specifically, by adjusting the focal length of the adjustable lens barrel 402, it is ensured that the imaging plane of the monocular CCD camera 403 perfectly coincides with the plane where the nozzle is located. This precise alignment can provide high-quality images, enabling the system to accurately capture the relative position of the light spot and the nozzle, thereby improving the accuracy and reliability of the measurement.

[0121] The auxiliary light source 401 provides sufficient illumination for the field of view, enabling the monocular CCD camera 403 to capture images with sufficient brightness and contrast. By using the auxiliary light source 401 for illumination instead of a traditional ring light source, shadow and reflection problems caused by space limitations or improper light source layout can be effectively reduced. This flexible light source selection ensures uniform lighting conditions, significantly enhancing the contrast and clarity of the images, making the spot and nozzle features more distinct. The use of the monocular CCD camera 403 makes the image acquisition process efficient and real-time, capable of quickly obtaining the relative position image of the spot and the nozzle. Such real-time feedback is crucial for dynamic adjustment and control systems, enabling faster response.

[0122] Selecting the coaxial monocular vision detection module 4 simplifies the system design, especially in space-constrained environments. The integration of the entire module not only saves space but also reduces interference between components, enhancing the stability and usability of the system.

[0123] After automatically processing the relative position image of the spot and the nozzle, it can be used for subsequent control and calibration tasks, improving the automation level of the entire laser processing system. Such an advanced detection system can reduce human intervention and improve work efficiency.

[0124] Optionally, the converting the relative position image to obtain a temporary image includes:

[0125] Determining the segmentation threshold of the relative position image through an image processing algorithm;

[0126] Performing region segmentation on the relative position image according to the segmentation threshold to obtain the temporary image.

[0127] Optionally, the segmentation threshold includes a first segmentation threshold and a second segmentation threshold; the performing region segmentation on the relative position image according to the segmentation threshold to obtain the temporary image includes:

[0128] Processing the relative position image based on the first segmentation threshold and the second segmentation threshold to obtain the temporary image;

[0129] Wherein, the first segmentation threshold is used to segment the high-brightness gray area and the medium-brightness gray area in the relative position image; the second segmentation threshold is used to segment the medium-brightness gray area and the low-brightness gray area in the relative position image.

[0130] Specifically, due to system space limitations, the camera illumination does not use a ring light source in traditional vision, but instead selects an offset auxiliary light source 401 for illumination. This results in changes in the gray-scale distribution in the image as the relative position relationship between the light spot and the nozzle varies. Therefore, it is necessary to use image processing algorithms such as Otsu's method to calculate the segmentation threshold in the current image. By traversing the gray-scale values from 0 to 255, the gray-scale value that maximizes the between-class variance is taken as the optimal threshold, thereby achieving adaptive image segmentation.

[0131] The calculation formula for the first segmentation threshold t is as follows:

[0132]

[0133] In the formula: q1(t) is the within-class mean of pixels with gray-scale values less than the threshold t; q2(t) is the within-class mean of pixels with gray-scale values greater than t; is the within-class variance of pixels with gray-scale values less than the threshold t; is the within-class variance of pixels with gray-scale values greater than the threshold t.

[0134] Since there is sputtering of the light spot in the high-pressure water chamber and there is noise in the image due to impurities such as dust near the nozzle, this will affect the detection accuracy of the light spot and the nozzle. To enhance the image, weaken the background, and reduce the influence of noise, the relative position image is binarized according to the threshold obtained by Otsu's method (image processing algorithm) to segment the images of the light spot and the nozzle (as Figure 4 shown).

[0135]

[0136] Among them, g(x, y) is the binarized image; ti is the pixel gray-scale value of the relative position image; the second segmentation threshold is: 255 - t; when segmenting the light spot, T = 255 - t, and when segmenting the nozzle, it is segmented based on the first segmentation threshold t.

[0137] In some embodiments, Figure 4 is obtained through binarization conversion according to Figure 3 and Figure 3 is the original gray-scale image (relative position image). The original gray-scale image can generally be divided into a high-brightness gray-scale area, a medium-brightness gray-scale area, and a low-brightness gray-scale area according to the gray-scale values. In the gray-scale image, the light spot corresponds to the high-brightness gray-scale area, the nozzle is located in the low-brightness gray-scale area, and the background noise belongs to the medium-brightness gray-scale area. Among them, high-brightness gray-scale area (light spot): In the gray-scale image, the gray-scale value of the light spot area is relatively high, usually close to 255. Medium-brightness gray-scale area (background noise): The gray-scale value of the noise area is in the middle, between the light spot and the nozzle. Low-brightness gray-scale area (nozzle): The gray-scale value of the nozzle area is relatively low, close to 0.

[0138] To distinguish these three grayscale regions, it is necessary to perform a three-segment division on the grayscale values of the entire image, that is, to determine two division thresholds. Using Otsu's method to calculate the grayscale of the original grayscale image, a t value is obtained, as shown in the above formula (1). This t value is used as the first segmentation threshold, and 255 - t is used as the second segmentation threshold, and the threshold is applied for segmentation: High-brightness grayscale region (light spot): Pixels with grayscale values greater than t are marked as the "high-brightness grayscale region", that is, the light spot region. Medium-brightness grayscale region (background noise): Pixels with grayscale values between t and 255 - t are marked as the "medium-brightness grayscale region", that is, the background noise region. Low-brightness grayscale region (nozzle): Pixels with grayscale values less than 255 - t are marked as the "low-brightness grayscale region", that is, the nozzle region.

[0139] By applying these segmentation thresholds, the high-brightness grayscale region (light spot), medium-brightness grayscale region (background noise), and low-brightness grayscale region (nozzle) can be effectively distinguished, thereby achieving region segmentation. Finally, the pixel values of the high-brightness grayscale region (light spot) and low-brightness grayscale region (nozzle) will be set to 0, which appears as black on the image, while the pixel values of the medium-brightness grayscale region (background noise) will be set to 255. As Figure 4 shown in the schematic diagram of the temporary image.

[0140] Optionally, performing circle detection on the edge features of the light spot to obtain a light spot detection result includes:

[0141] Counting all pixel points in the edge features of the light spot, and selecting a preset number of the pixel points for random sample consensus detection to obtain a set of edge pixel points;

[0142] Fitting the set of edge pixel points to obtain a target fitting circle, and taking the target fitting circle as the light spot detection result.

[0143] Optionally, fitting the set of edge pixel points to obtain a target fitting circle includes:

[0144] Selecting a preset number of the pixel points in the set of edge pixel points for fitting to obtain a current fitting circle, and taking the number of edge points included in the current fitting circle as the inlier number;

[0145] Repeating the above steps, when the inlier number of the current fitting circle is higher than the inlier number of the fitting circle in the previous iteration, and the radius of the current fitting circle satisfies the preset radius range, taking the current fitting circle as a temporary fitting circle, or until the preset number of iterations is reached, taking the current fitting circle as the target fitting circle.

[0146] In some embodiments, after image preprocessing, there are still slight unevenness problems in the spot edge features that cannot be eliminated. Therefore, the RCD (Random Circle Detection method) is used as the main means for spot circle detection. During the detection process, a preset number of pixel points (such as 90% of the points) in the spot edge are selected to participate in the detection, and the number of iterations is set to 700 times (the preset number of iterations), and the reasonable circle radius range is from 95 pixels to 500 pixels (the preset radius range). In each iteration, a preset number, usually 3 edge points, are randomly selected for circle fitting, and the number of edge points contained within the fitted circle, that is, the number of inliers, is calculated. If the number of inliers in the current iteration is greater than the number of inliers of the fitted circle in the previous iteration, and the radius is within the preset radius range, then the fitted circle of the current iteration is stored as a temporary fitted circle. The above process is repeated until the set number of iterations is reached, and the finally stored temporary fitted circle will be used as the target fitted circle, that is, the detection result is feedback. The spot detection result, such as Figure 5 The image before driving the electric deflection mirror to automatically calibrate in Figure 5 contains the spot detection result.

[0147] Optionally, the temporary image includes a spot image; after converting the relative position image to obtain the temporary image, the following further includes:

[0148] Performing median filtering on the spot image, and performing opening operation on the filtered image according to the first preset circular convolution kernel to obtain a temporary spot image;

[0149] Performing closing operation on the temporary spot image according to the second preset circular convolution kernel to obtain a processed spot image.

[0150] Specifically, for the spot image after binary segmentation (the spot image in the temporary image), there are still problems such as independent sputtering light points and uneven edges. First, after performing median filtering on the spot image, median filtering replaces the value of each pixel with the median of the pixel values in its neighborhood, thereby suppressing larger noise values. This provides a clean base image for subsequent morphological operations.

[0151] Next, morphological operations are performed. An opening operation is performed on the spot image using a first preset circular convolution kernel with a diameter, such as a circular convolution kernel with a diameter of 25 pixels, to eliminate independent sputtering light points around the spot; then a closing operation is performed on the spot image according to the second preset circular convolution kernel, such as a circular convolution kernel with a diameter of 21 pixels, to close the gap at the edge of the spot, eliminate unevenness and smooth the edge. The result of the morphological operation, such as Figure 7 shown, is a schematic diagram of the processed spot image.

[0152] Optionally, edge detection is performed on the temporary image to extract the edge features therein, and the edge features include spot edge features and nozzle edge features.

[0153] In some embodiments, the Canny operator is used to perform edge detection on the spot image and the nozzle image after filtering and morphological processing. The Canny operator includes a strong edge threshold, a weak edge threshold, and a Sobel operator. First, the Sobel operator is used to scan the image to calculate the gradient intensity of the image. When the gradient intensity is greater than the strong edge threshold, the region is considered a strong edge; when the gradient intensity is between the strong edge threshold and the weak edge threshold, the region is considered a weak edge. For example, in the extraction of the spot edge features, the strong edge threshold of the Canny operator is set to 47.5, the weak edge threshold is 20, and the size of the Sobel operator is 3 pixels. In the extraction of the nozzle edge features, the strong edge threshold of the Canny operator is 47.5, the weak edge threshold is 17.5, and the size of the Sobel operator is also 3 pixels.

[0154] Among them, appropriately increasing the strong edge threshold can improve the sensitivity of detecting clear edges, while ignoring the blurred edges caused by image noise, reducing false detections, and improving the resistance to noise. However, too high a strong edge threshold may lead to missed detections. Therefore, considering comprehensively, the strong edge thresholds of both the spot and the nozzle are set to 47.5.

[0155] Appropriately reducing the weak edge threshold can retain more edge details and make the detected edges more continuous and complete. However, if the weak edge threshold is set too small, it may lead to false detections. Since there are many tiny splashes at the edge of the spot, these splashes are equivalent to noise in the edge detection process. Therefore, the weak edge threshold of the spot is set to 20 to avoid false detection of the splash edges, thus affecting the accuracy of edge detection. The shape of the nozzle is regular and it occupies a large area in the figure. Appropriately reducing the weak edge threshold to 17.5 can obtain a more continuous and complete nozzle edge.

[0156] The Sobel operator is used for image smoothing. The larger its value, the less noise, but the edges will also be more blurred, and the size must be an odd number greater than 1. This is because the Sobel operator is equivalent to a square window, and the calculated result must be fed back to the central pixel. Only when the side length is an odd number of pixels will there be a central pixel in the window. After the above-mentioned region segmentation, most of the background noise in the image is filtered out. Therefore, only the smallest Sobel operator (3 pixels) needs to be used to ensure the clearest edges for subsequent detection.

[0157] In some specific embodiments, as Figure 8 shown, the water-guided laser light-water coupling automatic calibration method includes:

[0158] Collecting a grayscale image (relative position image) through a laser coaxial CCD camera (monocular CCD camera).

[0159] The Otsu method is used to binarize the acquired image, that is, to binarize the relative position image to obtain the binary threshold of the image, namely the first segmentation threshold and the second segmentation threshold.

[0160] The binarized segmentation of the light spot / nozzle image is to segment the relative position image by the first segmentation threshold and the second segmentation threshold, and segment the light spot image and the nozzle image according to the binarization result, that is, to obtain a temporary image, which includes the light spot image and the nozzle image.

[0161] Processing process of the light spot image: Median filtering to remove noise: Perform median filtering on the light spot image to remove the noise in the image. Morphological processing: Perform morphological operations on the light spot image to further improve the image quality. Apply the Canny operator for edge detection to extract the edge features of the light spot, and finally detect the light spot circle based on the random circle detection algorithm to obtain the light spot detection result;

[0162] Processing process of the nozzle image: Median filtering to remove noise: Perform median filtering on the nozzle image to remove the noise in the image. Canny edge detection: Apply the Canny operator for edge detection to extract the edge features of the nozzle. Finally, detect the nozzle circle based on the improved Hough gradient circle transformation to obtain the nozzle detection result.

[0163] Calculate the relative height of the light spot / nozzle: Calculate the distance between the height of the light spot and the nozzle based on the nozzle detection result and the light spot detection result.

[0164] Initialize by controlling the corresponding motor (the motor corresponding to the electric deflecting mirror). Then, control the motor to continuously move in a preset step length and collect the image change data of the nozzle; Determine the mapping relationship between the image change data and the preset step length based on the least squares method, that is, fit the mapping relationship between the step length of the motor and the pixel change.

[0165] Finally, adjust the electric deflecting mirror based on the mapping relationship and the distance between the height of the light spot and the nozzle to align the images of the light spot and the nozzle, realizing automatic calibration of the light-water coupling.

[0166] Although the present invention is disclosed as above, the protection scope of the present invention is not limited thereto. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will all fall within the protection scope of the present invention.

Claims

1. A method for automatic calibration of water-guided laser light-water coupling, characterized in that: The invention is applied to a water-guided laser light-water coupling automatic calibration system, the water-guided laser light-water coupling automatic calibration system comprising a laser collimation head (1), a coaxial monocular vision detection module (4), an electric optical path calibration module (3) and a coupling cavity (6) provided with a nozzle, and the water-guided laser light-water coupling automatic calibration method comprises: When the laser collimation head (1) receives the laser beam, The coaxial monocular vision detection module (4) is used to obtain a relative position image between the light spot generated by the laser beam and the nozzle, and the relative position image is converted to obtain a temporary image; Performing edge detection on the temporary image to extract edge features therein, wherein the edge features include light spot edge features and nozzle edge features; Performing circle detection on the edge features of the light spot to obtain a light spot detection result; Performing detection based on the nozzle edge feature based on improved Hough gradient circle transform to obtain a nozzle detection result; Based on the light spot detection result and the nozzle detection result, the light path of the laser beam is changed by the electric light path calibration module (3) to adjust the position of the light spot, so as to realize automatic calibration of light-water coupling.

2. The water-guided laser light-water coupling automatic calibration method according to claim 1, characterized in that: The detecting according to the nozzle edge feature based on the improved Hough gradient circle transform to obtain the nozzle detection result includes: Determine the gradient information of each edge point in the nozzle edge feature, and construct a parameter space based on the gradient information; the gradient information includes a gradient amplitude and a gradient direction; Determine a plurality of candidate circle centers according to the gradient magnitude and the gradient direction of each edge point, and vote on each of the candidate circle centers in the parameter space to generate a plurality of temporary candidate circle centers; Determine a corresponding candidate circle according to each of the temporary candidate circle centers and a preset radius parameter; Based on the nozzle edge features, the overlap degree of each candidate circle is detected, and the candidate circle with the highest overlap degree is used as the nozzle detection result.

3. The water-guided laser light-water coupling automatic calibration method according to claim 1, characterized in that: The electric optical path calibration module (3) comprises an electric deflection mirror (301), and based on the light spot detection result and the nozzle detection result, the electric optical path calibration module (3) changes the optical path of the laser to adjust the position of the light spot, so as to realize automatic calibration of light-water coupling, comprising: Adjusting the electric deflection mirror (301) to an initial position, controlling the electric deflection mirror (301) to move continuously at a preset step length, and collecting image change data of the nozzle; Determine a mapping relationship between the image change data and the preset step length based on a preset algorithm; The electric deflection mirror (301) is adjusted according to the mapping relationship, the light spot detection result and the nozzle detection result to change the optical path of the laser so that the light spot is aligned with the image of the nozzle.

4. The water-guided laser light-water coupling automatic calibration method according to claim 2, characterized in that: The coaxial monocular visual detection module (4) comprises an auxiliary light source (401), an adjustable lens barrel (402) and a monocular CCD camera (403), wherein the monocular CCD camera (403) is connected to the adjustable lens barrel (402), and the auxiliary light source (401) and the laser beam are located on the same optical axis and are used to illuminate the light spot; The method of obtaining the relative position image of the light spot and the nozzle through the coaxial monocular vision detection module (4) comprises: By adjusting the focal length of the adjustable lens barrel (402), the imaging plane of the monocular CCD camera (403) is made to coincide with the plane where the nozzle is located; The auxiliary light source (401) is used for illumination, and the monocular CCD camera (403) is used to collect an image of the relative position of the light spot and the nozzle.

5. The water-guided laser light-water coupling automatic calibration method according to claim 4, characterized in that: The converting the relative position image to obtain a temporary image includes: Determine the segmentation threshold of the relative position image by an image processing algorithm; The relative position image is segmented into regions according to the segmentation threshold to obtain the temporary image.

6. The water-guided laser light-water coupling automatic calibration method according to claim 1, characterized in that: The performing circle detection on the edge feature of the light spot to obtain a light spot detection result includes: Counting all the pixel points in the edge features of the light spot, and selecting a preset number of the pixel points for random sampling consistency detection to obtain an edge pixel point set; The edge pixel point set is fitted to obtain a target fitting circle, and the target fitting circle is used as the light spot detection result.

7. The water-guided laser light-water coupling automatic calibration method according to claim 6, characterized in that: The step of fitting the edge pixel point set to obtain a target fitting circle includes: Selecting a preset number of pixel points in the edge pixel point set for fitting to obtain a current fitting circle, and taking the number of edge points included in the current fitting circle as the number of inliers; Repeat the above steps. When the number of inner points of the current fitting circle is higher than the number of inner points of the fitting circle of the previous iteration and the radius of the current fitting circle meets the preset radius range, the current fitting circle is used as the target fitting circle; or until the preset number of iterations is reached, the current fitting circle is used as the target fitting circle.

8. The water-guided laser light-water coupling automatic calibration method according to claim 3, characterized in that: The electric optical path calibration module (3) further comprises a motion controller, the motion controller being connected to the electric deflection mirror, and the electric deflection mirror (301) being adjusted according to the mapping relationship, the light spot detection result and the nozzle detection result so as to align the light spot with the image of the nozzle, comprising: Based on the mapping relationship, the light spot detection result and the nozzle detection result, determining an execution instruction through a motion controller; According to the execution instruction, the electric deflection mirror (301) is adjusted to perform deflection so as to align the light spot with the image of the nozzle.

9. The water-guided laser light-water coupling automatic calibration method according to claim 5, characterized in that: The segmentation threshold includes a first segmentation threshold and a second segmentation threshold; and performing region segmentation on the relative position image according to the segmentation threshold to obtain the temporary image includes: Processing the relative position image based on the first segmentation threshold and the second segmentation threshold to obtain the temporary image; Among them, the first segmentation threshold is used to segment the high-brightness grayscale area and the medium-brightness grayscale area in the relative position image; the second segmentation threshold is used to segment the medium-brightness grayscale area and the low-brightness grayscale area in the relative position image.

10. The water-guided laser light-water coupling automatic calibration method according to claim 1, characterized in that: The temporary image includes a spot image; and after converting the relative position image to obtain the temporary image, the following further comprises: Performing median filtering on the light spot image, and performing an opening operation on the filtered image according to a first preset circular convolution kernel to obtain a temporary light spot image; A closing operation is performed on the temporary light spot image according to a second preset circular convolution kernel to obtain a processed light spot image.

Images