CN115194323A - Positioning welding method of laser welding machine - Google Patents
Positioning welding method of laser welding machine Download PDFInfo
- Publication number
- CN115194323A CN115194323A CN202211026155.3A CN202211026155A CN115194323A CN 115194323 A CN115194323 A CN 115194323A CN 202211026155 A CN202211026155 A CN 202211026155A CN 115194323 A CN115194323 A CN 115194323A
- Authority
- CN
- China
- Prior art keywords
- welding
- image
- welded
- seams
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K31/00—Processes relevant to this subclass, specially adapted for particular articles or purposes, but not covered by only one of the preceding main groups
- B23K31/02—Processes relevant to this subclass, specially adapted for particular articles or purposes, but not covered by only one of the preceding main groups relating to soldering or welding
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K26/00—Working by laser beam, e.g. welding, cutting or boring
- B23K26/08—Devices involving relative movement between laser beam and workpiece
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B23—MACHINE TOOLS; METAL-WORKING NOT OTHERWISE PROVIDED FOR
- B23K—SOLDERING OR UNSOLDERING; WELDING; CLADDING OR PLATING BY SOLDERING OR WELDING; CUTTING BY APPLYING HEAT LOCALLY, e.g. FLAME CUTTING; WORKING BY LASER BEAM
- B23K26/00—Working by laser beam, e.g. welding, cutting or boring
- B23K26/70—Auxiliary operations or equipment
- B23K26/702—Auxiliary equipment
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
- G06T7/001—Industrial image inspection using an image reference approach
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30152—Solder
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Optics & Photonics (AREA)
- Mechanical Engineering (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Plasma & Fusion (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Quality & Reliability (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Laser Beam Processing (AREA)
Abstract
The invention provides a positioning welding method of a laser welding machine, which relates to the technical field of laser welding and comprises the following steps: the method comprises the following steps: collecting data of various welding spots and welding seams before welding, constructing a pre-resource package, collecting data of various welding spots and welding seams after welding, and constructing a post-resource package; step two: shooting an image of a position to be welded by using a CCD camera array, and performing noise reduction and resolution adjustment on the image; the method collects data of various welding spots and welding seams before welding, constructs a front resource package, shoots images of the positions to be welded, compares the data by extracting the welding spots and the welding seams in the front resource package and pixel characteristics, texture characteristics and shape characteristics of the images of the positions to be welded, determines the welding spots and the welding seams in the images of the positions to be welded, can position laser welding, records the data of the welding spots and the welding seams of various workpieces through the front resource package, is convenient to position various welding spots and welding seams, does not need manual assistance, automatically performs, and improves positioning efficiency.
Description
Technical Field
The invention relates to the technical field of laser welding, in particular to a positioning welding method of a laser welding machine.
Background
In the current laser welding market, a laser welding robot or an automatic laser welding machine replaces manual batch welding operation to reach 90% of the range, and the welding robot or the automatic welding machine is favored by more and more small and medium-sized enterprises due to the characteristics of higher efficiency, high quality, easiness in management and the like, but for some products requiring high precision, the problem cannot be solved by only depending on the welding robot and the like, and manual assistance is needed;
the existing laser welding machine and welding technology mainly have the following defects: the welding workpieces are various in variety and small in batch, and the requirement on welding technicians is high during welding positioning; the laser welding robot is often subjected to various technical conditions such as workpiece positioning deviation and the like in the welding production process; the laser welding robot has welding errors sometimes, and repair welding needs to be carried out after welding; therefore, the invention provides a positioning welding method of a laser welding machine to solve the problems in the prior art.
Disclosure of Invention
Aiming at the problems, the invention provides the positioning welding method of the laser welding machine, which is convenient for positioning various welding spots and welding seams, does not need manual assistance, is carried out automatically, and improves the positioning efficiency.
In order to realize the purpose of the invention, the invention is realized by the following technical scheme: a positioning welding method of a laser welding machine comprises the following steps:
the method comprises the following steps: collecting data of various welding spots and welding seams before welding, constructing a pre-resource package, collecting data of various welding spots and welding seams after welding, and constructing a post-resource package;
step two: shooting an image of a position to be welded by using a CCD camera array, and performing noise reduction and resolution adjustment on the image;
step three: extracting pixel characteristics, textural characteristics and shape characteristics of welding spots and welding seams in the resource package before welding, and collecting the pixel characteristics, the textural characteristics and the shape characteristics of an image to be welded;
step four: comparing the three types of characteristics of the image to be welded with the three types of characteristics in the front resource packet, and determining a welding point and a welding line in the image to be welded;
step five: vectorizing the Path and Row numerical values corresponding to the image to form a point, line and plane model, obtaining the Path and Row numerical values of the welding points and the welding lines, and restoring the Path and Row numerical values into real coordinates according to a proportion;
step six: adjusting the motion track of the adaptive arm, driving a laser welding robot to weld a welding spot and a welding seam according to the coordinates, and shooting a welding image in real time by a CCD camera array during welding;
step seven: selecting welding spot and welding line data welded by a corresponding process in a rear resource packet according to the welding process of the laser welding robot;
step eight: and extracting the size data of the welding spot welding seam welded by the corresponding process in the resource packet, extracting the size data of the welding image shot in real time, comparing, and finely adjusting the motion track of the adaptive arm.
The further improvement lies in that: in the first step, the internet is accessed, various data of welding spots and welding seams before welding are collected, the data of the welding spots and the welding seams of the workpiece matched with various laser welding devices are collected, the data of the welding spots and the welding seams are included, and the data of the welding spots and the welding seams after welding are collected in the first step, and the data of the welding spots and the welding seams after the laser welding devices adopt various welding processes are included.
The further improvement lies in that: in the second step, the noise reduction of the image is specifically as follows: by adopting a bilateral filtering method, taking a 3 × 3 neighborhood around a target pixel as an example, taking the average value of the gray values of nine pixels in the 3 × 3 neighborhood as the gray value of the target pixel, then changing the average value into a Gaussian weighted average value, generating a Gaussian template when the gray value of the pixel farther away from the target pixel in the neighborhood has a smaller weight, and then further subtracting the influence of the gray gradient of the pixel in weight design to obtain the image after noise reduction.
The further improvement lies in that: in the second step, the resolution adjustment specifically comprises: firstly, enhancing an image by using a Laplace operator, disintegrating the image by using secondary differential of the image, namely, improving the contrast by using neighborhood pixels; and then, by a hyper-resolution algorithm SRCNN, based on a convolutional neural network, taking an original video image as input, increasing the resolution to a set value by using an up-sampling algorithm, and enhancing the resolution of the image through convolution operation.
The further improvement is that: in the third step, color pixel characteristics of the welding points and the welding seams in the front resource package and the images to be welded are obtained by using an HOG characteristic extraction algorithm, texture characteristics of the welding points and the welding seams in the front resource package and the images to be welded are extracted by using a filtering and density analysis process, and shape characteristics of the welding points and the welding seams in the front resource package and the images to be welded are determined by using stretching vectorization.
The further improvement lies in that: and in the fourth step, the YOYO neural network is utilized to take the three types of features in the front resource packet as a training set, and the training set is compared with the three types of features of the image to be welded to determine the welding points and the welding seams in the image to be welded.
The further improvement lies in that: and fifthly, vectorizing the Path and Row numerical values corresponding to the image by using the SVG to form a point, line and plane model, scaling the coordinates in the model and the actual coordinates according to a set proportion, and embedding the Path and Row numerical values into the positions of the welding points and the welding seams determined in the model.
The further improvement lies in that: in the sixth step, during welding, the CCD camera array shoots welding images in real time, the synchronous master control system records the specific welding process of the laser welding robot, and timestamps the welding images shot in real time.
The further improvement lies in that: and seventhly, recording the specific welding process of the laser welding robot according to the master control system, selecting welding spot and welding line data after welding of the corresponding process in the rear resource packet, wherein the data comprises all image data from the beginning to the end of welding of the whole process flow, and marking time nodes according to process time.
The further improvement is that: in the step eight, the size data of the welding spot welding seam welded by the corresponding process in the resource package after being extracted by stretching vectorization is zoomed according to a set proportion, the size data of the welding image shot in real time is synchronously measured, synchronous comparison is carried out according to the time node after the welding by the corresponding process in the resource package after being shot in real time, whether the welding is normal or not is determined, and the motion track of the adaptive arm is finely adjusted when the welding is abnormal.
The invention has the beneficial effects that:
1. the method collects data of various welding spots and welding seams before welding, constructs a front resource package, shoots an image of a to-be-welded part, compares the data by extracting the welding spots and the welding seams in the front resource package and pixel characteristics, texture characteristics and shape characteristics of the image of the to-be-welded part, determines the welding spots and the welding seams in the image of the to-be-welded part, can position the laser welding, records the data of various welding spots and welding seams of workpieces through the front resource package, is convenient to position various welding spots and welding seams, does not need manual assistance, automatically performs the positioning, and improves the positioning efficiency.
2. The method comprises the steps of shooting an image of a position to be welded, reducing noise and adjusting resolution of the image to enable the image to be more accurate, vectorizing Path and Row numerical values corresponding to the image to form a point model, a line model and a plane model, scaling coordinates in the model and actual coordinates according to a set proportion, obtaining the determined welding point and welding line Path and Row numerical values, and restoring the welding point and welding line Path and Row numerical values into actual coordinates according to the proportion, so that positioning accuracy is improved, and errors are avoided.
3. The method comprises the steps of shooting images during welding in real time, stamping a timestamp, selecting welding spot and welding line data corresponding to a welding process in a rear resource package, marking a time node, extracting welding spot and welding line size data after welding in the rear resource package corresponding to the process, comparing the size data with the synchronously measured size data of the welding images shot in real time according to the time node and the timestamp, and determining whether the welding is normal or not, so that fine adjustment is realized, the welding quality is ensured, and subsequent repair welding is not needed.
Drawings
FIG. 1 is a flow chart of the present invention.
Detailed Description
For the purpose of enhancing understanding of the present invention, the present invention will be further described in detail with reference to the following examples, which are provided for illustration only and are not intended to limit the scope of the present invention.
Example one
According to the illustration in fig. 1, the embodiment provides a positioning welding method of a laser welding machine, which comprises the following steps:
the method comprises the following steps: collecting data of various welding spots and welding seams before welding, constructing a pre-resource package, collecting data of various welding spots and welding seams after welding, and constructing a post-resource package;
step two: shooting an image of a position to be welded by using a CCD camera array, and performing noise reduction and resolution adjustment on the image;
step three: extracting pixel characteristics, textural characteristics and shape characteristics of welding spots and welding seams in the resource package before welding, and collecting the pixel characteristics, the textural characteristics and the shape characteristics of an image to be welded;
step four: comparing the three types of characteristics of the image to be welded with the three types of characteristics in the front resource packet, and determining a welding point and a welding line in the image to be welded;
step five: vectorizing Path and Row numerical values corresponding to the image to form a point model, a line model and a surface model, obtaining the Path and Row numerical values of the welding spot and the welding line, and restoring the Path and Row numerical values into real coordinates according to a proportion;
step six: adjusting the motion track of the adaptive arm, driving a laser welding robot to weld a welding spot and a welding seam according to the coordinates, and shooting a welding image in real time by a CCD camera array during welding;
step seven: selecting welding spot and welding line data welded by a corresponding process in a rear resource packet according to the welding process of the laser welding robot;
step eight: and extracting the size data of the welding spot welding seam welded by the corresponding process in the resource packet, extracting the size data of the welding image shot in real time, comparing, and finely adjusting the motion track of the adaptive arm.
The method comprises the steps of collecting data of various welding spots and welding seams before welding, constructing a front resource package, shooting an image of a to-be-welded position by utilizing a CCD camera array, comparing the image by extracting the welding spots and the welding seams in the front resource package and pixel characteristics, texture characteristics and shape characteristics of the image of the to-be-welded position, determining the welding spots and the welding seams in the image of the to-be-welded position, restoring the image into real coordinates according to proportions according to Path and Row vectorization values corresponding to the image, positioning laser welding can be carried out on the real welding spots and the welding seams, recording the data of various welding spots and welding seams of a workpiece through the front resource package, facilitating the positioning of various welding spots and welding seams, avoiding manual assistance, automatically carrying out and improving the positioning efficiency.
Example two
According to fig. 1, the present embodiment provides a tack welding method for a laser welding machine, including the following steps:
the method comprises the steps of accessing the Internet, collecting data of various welding spots and welding seams before welding, collecting data of the welding spots and the welding seams of workpieces matched with various laser welding devices, collecting data of various welding spots and welding seams after welding, and collecting data of the welding spots and the welding seams after welding by the laser welding devices by adopting various welding processes.
Shooting an image of a position to be welded by using a CCD camera array, and performing noise reduction and resolution adjustment on the image; the noise reduction of the image is specifically as follows: adopting a bilateral filtering method, taking a 3 × 3 neighborhood around a target pixel as an example, taking the average value of the gray values of nine pixels in the 3 × 3 neighborhood as the gray value of the target pixel, then changing the average value into a Gaussian weighted average value, generating a Gaussian template when the gray value of the pixel farther away from the target pixel in the neighborhood has smaller weight, and then further subtracting the influence of the gray gradient of the pixel in weight design so as to obtain an image after noise reduction; the bilateral filtering is specifically: consider first the gaussian partial weight:
then consider the pixel gray gradient partial weight:
where (i, j) is the neighborhood pixel coordinate in the template, (k, l) is the template center pixel coordinate, f (i, j) is the pixel value at coordinate (i, j), σ d And σ r The two weights are respectively the variance of the two weights, and the two weights are multiplied to obtain the weight of the bilateral filtering:
the adjustment resolution specifically comprises: firstly, enhancing an image by using a Laplace operator, disintegrating the image by using secondary differential of the image, wherein the differential is sharpening and the integral is fuzzy in the image field, and disintegrating the image by using the secondary differential, namely, improving the contrast by using neighborhood pixels; then, by a hyper-resolution algorithm SRCNN, based on a convolution neural network, an original video image is used as input, the resolution is improved to a set value by an up-sampling algorithm, and the hyper-resolution output and the image resolution are enhanced by 3 layers of convolution operation of 9 multiplied by 128,3 multiplied by 3 multiplied by 64,5 multiplied by 5. The invention shoots the image at the position to be welded, and performs noise reduction and resolution adjustment on the image, so that the image is more accurate.
Extracting pixel characteristics, texture characteristics and shape characteristics of welding spots and welding seams in the resource packet before welding, and collecting the pixel characteristics, the texture characteristics and the shape characteristics of an image at the position to be welded; the method specifically comprises the following steps: color pixel characteristics of the welding spots and the welding seams in the front resource package and the images at the positions to be welded are obtained by using an HOG characteristic extraction algorithm, texture characteristics of the welding spots and the welding seams in the front resource package and the images at the positions to be welded are extracted by using a filtering and density analysis process, and shape characteristics of the welding spots and the welding seams in the front resource package and the images at the positions to be welded are determined by using stretching vectorization. Wherein, using the HOG feature extraction algorithm, an image (an object or a picture to be detected) is: graying (treating the image as a three-dimensional image in x, y, z (gray); carrying out color space standardization (normalization) on the input image by adopting a Gamma correction method; the purpose being to adjust the imageThe contrast reduces the influence caused by local shadow and illumination change of the image, and simultaneously can inhibit the interference of noise; calculating the gradient (including magnitude and direction) of each pixel of the image; the method mainly aims to capture contour information and further weakens the interference of illumination; dividing the image into small cells (e.g., 6*6 pixels/cell); counting the gradient histogram (the number of different gradients) of each cell to form a descriptor of each cell; forming each plurality of cells into a block (for example, 3*3 cells/block), and connecting the feature descriptors of all the cells in the block in series to obtain the HOG feature descriptor of the block; connecting HOG feature descriptors of all blocks in an image in series to obtain the HOG feature descriptors of the image (a target to be detected), wherein the HOG feature descriptors are final feature vectors which can be used for classification, so that color pixel features are obtained; then, filtering the image, placing the texture result obtained by filtering and extracting the image in an ARCGIS (autoregressive moving average) for density analysis, determining the texture roughness of the image, then solving a second derivative of the spectrum of the image, writing a second derivative operation algorithm in ENVIIDL (enhanced image feature value), wherein the second derivative algorithm is as follows: derm (flash [, S ]) and]) ) is used, S represents spectral, and the second derivative formula y "= d is input in ENVIIDL 2 y/dx 2 Programming to complete second derivative calculation of the image, converting the characteristic value of the image into raster data by using an overlay logic analysis method, and determining texture characteristics; the image is then subsequently subjected to electrical, line, plane stretching, and vectorization to determine shape frame features.
And comparing the three types of features in the front resource packet as a training set with the three types of features of the image to be welded by using the YOYO neural network, and determining the welding points and the welding seams in the image to be welded. The method comprises the steps of collecting data of various welding spots and welding seams before welding, constructing a front resource package, shooting an image of a to-be-welded position by utilizing a CCD camera array, comparing the image by extracting the welding spots and the welding seams in the front resource package and pixel characteristics, texture characteristics and shape characteristics of the image of the to-be-welded position, determining the welding spots and the welding seams in the image of the to-be-welded position, restoring the image into real coordinates according to proportions according to Path and Row vectorization values corresponding to the image, positioning laser welding can be carried out on the real welding spots and the welding seams, recording the data of various welding spots and welding seams of a workpiece through the front resource package, facilitating the positioning of various welding spots and welding seams, avoiding manual assistance, automatically carrying out and improving the positioning efficiency.
Vectorizing the Path and Row numerical values corresponding to the image by using the SVG to form a point, line and plane model, scaling the coordinates in the model and the actual coordinates according to a set proportion, embedding the Path and Row numerical values into the positions of the welding points and the welding seams determined in the model, obtaining the Path and Row numerical values of the welding points and the welding seams, and restoring the Path and Row numerical values into real coordinates according to the proportion; the method vectorizes the Path and Row values corresponding to the image to form a point, line and plane model, scales the coordinates in the model and the actual coordinates according to a set proportion, obtains the determined Path and Row values of the welding points and the welding seams, and can restore the parameters to the actual coordinates according to the proportion, thereby improving the positioning accuracy and avoiding errors.
The motion trail of the self-adaptive arm is adjusted, the laser welding robot is driven to weld welding spots and welding seams according to coordinates, during welding, the CCD camera array shoots welding images in real time, the synchronous master control system records the specific welding process of the laser welding robot, and timestamps are printed on the welding images shot in real time.
Recording the specific welding process of the laser welding robot according to a master control system, selecting welding spot welding line data after welding of the corresponding process in a rear resource packet, wherein the data comprises all image data from the beginning to the end of welding of the whole process flow, and marking time nodes according to process time.
And extracting size data of welding spots welded by corresponding processes in the resource package by utilizing stretching vectorization, zooming according to a set proportion, synchronously measuring size data of a welding image shot in real time, synchronously comparing according to time nodes welded by corresponding processes in the resource package after the timestamp shot in real time is corresponding, determining whether the welding is normal, and finely adjusting the motion track of the adaptive arm when the welding is abnormal. During welding, shooting a welding image in real time and stamping a timestamp, recording a welding process, selecting welding spot welding line data after welding of a corresponding process in a back resource package, including all image data from the beginning of welding to the end of welding of the whole process flow, and marking a time node, thereby extracting welding spot welding line size data after welding of the corresponding process in the back resource package, comparing the size data with the size data of the welding image shot in real time and synchronously measured according to the time node and the timestamp, determining whether welding is normal or not, and thus fine-tuning, ensuring the welding quality and avoiding subsequent repair welding.
The method comprises the steps of collecting data of various welding spots and welding seams before welding, constructing a front resource package, shooting an image of a to-be-welded position by utilizing a CCD camera array, comparing the image by extracting pixel characteristics, texture characteristics and shape characteristics of the welding spots and the welding seams in the front resource package and the image of the to-be-welded position, determining the welding spots and the welding seams in the image of the to-be-welded position, restoring the image into real coordinates according to a ratio according to Path and Row vectorization values corresponding to the image, performing positioning laser welding on the real welding spots and the welding seams, recording data of various workpiece welding spots and the welding seams through the front resource package, facilitating positioning of various welding spots and welding seams, automatically performing the positioning without manual assistance, improving the positioning efficiency, shooting the image of the to-be-welded position, performing noise reduction and resolution adjustment on the image, enabling the image to be more accurate, forming point, line and vectorization plane models according to Path and Row values corresponding to the image, obtaining the determined welding spots and welding seams and Row values according to the set ratio, and restoring the image into real coordinates, improving the positioning accuracy and avoiding errors. Meanwhile, when welding, the invention shoots the welding image in real time and stamps, records the welding process, selects the welding spot welding line data after the welding of the corresponding process in the back resource package, comprises all image data from the beginning to the end of the welding in the whole process flow, and marks time nodes, thereby extracting the welding spot welding line size data after the welding of the corresponding process in the back resource package, and comparing the size data with the synchronously measured real-time shot welding image size data according to the time nodes and the time stamps to determine whether the welding is normal, thereby fine-tuning, ensuring the welding quality and having no need of subsequent repair welding.
The foregoing illustrates and describes the principles, general features, and advantages of the present invention. It will be understood by those skilled in the art that the present invention is not limited to the embodiments described above, which are given by way of illustration of the principles of the present invention, but that various changes and modifications may be made without departing from the spirit and scope of the invention, and such changes and modifications are within the scope of the invention as claimed. The scope of the invention is defined by the appended claims and equivalents thereof.
Claims (10)
1. A positioning welding method of a laser welding machine is characterized by comprising the following steps:
the method comprises the following steps: collecting data of various welding spots and welding seams before welding, constructing a pre-resource package, collecting data of various welding spots and welding seams after welding, and constructing a post-resource package;
step two: shooting an image of a position to be welded by using a CCD camera array, and performing noise reduction and resolution adjustment on the image;
step three: extracting pixel characteristics, texture characteristics and shape characteristics of welding spots and welding seams in the resource packet before welding, and collecting the pixel characteristics, the texture characteristics and the shape characteristics of an image at the position to be welded;
step four: comparing the three types of characteristics of the image to be welded with the three types of characteristics in the front resource packet, and determining a welding point and a welding line in the image to be welded;
step five: vectorizing the Path and Row numerical values corresponding to the image to form a point, line and plane model, obtaining the Path and Row numerical values of the welding points and the welding lines, and restoring the Path and Row numerical values into real coordinates according to a proportion;
step six: adjusting the motion track of the adaptive arm, driving a laser welding robot to weld a welding spot and a welding seam according to the coordinates, and shooting a welding image in real time by a CCD camera array during welding;
step seven: selecting welding spot and welding line data welded by a corresponding process in a rear resource packet according to the welding process of the laser welding robot;
step eight: and extracting the size data of the welding spot welding seam welded by the corresponding process in the resource packet, extracting the size data of the welding image shot in real time, comparing, and finely adjusting the motion track of the adaptive arm.
2. A tack welding method of a laser welder according to claim 1, characterized in that: in the first step, the internet is accessed, various data of welding spots and welding seams before welding are collected, the data of the welding spots and the welding seams of the workpiece matched with various laser welding devices are collected, the data of the welding spots and the welding seams are included, and the data of the welding spots and the welding seams after welding are collected in the first step, and the data of the welding spots and the welding seams after the laser welding devices adopt various welding processes are included.
3. A tack welding method of a laser welder according to claim 2, characterized in that: in the second step, the noise reduction of the image is specifically as follows: by adopting a bilateral filtering method, taking a 3 × 3 neighborhood around a target pixel as an example, taking the average value of the gray values of nine pixels in the 3 × 3 neighborhood as the gray value of the target pixel, then changing the average value into a Gaussian weighted average value, generating a Gaussian template when the gray value of the pixel farther away from the target pixel in the neighborhood has a smaller weight, and then further subtracting the influence of the gray gradient of the pixel in weight design to obtain the image after noise reduction.
4. A tack welding method of a laser welder according to claim 3, characterized in that: in the second step, the resolution adjustment specifically comprises: firstly, enhancing an image by using a Laplace operator, disintegrating the image by using secondary differential of the image, namely, improving the contrast by using neighborhood pixels; and then, by a hyper-resolution algorithm SRCNN, based on a convolutional neural network, taking an original video image as input, increasing the resolution to a set value by using an up-sampling algorithm, and enhancing the resolution of the image through convolution operation.
5. A tack welding method for a laser welder, as claimed in claim 4, characterized in that: in the third step, color pixel characteristics of the welding points and the welding seams in the front resource package and the images to be welded are obtained by using an HOG characteristic extraction algorithm, texture characteristics of the welding points and the welding seams in the front resource package and the images to be welded are extracted by using a filtering and density analysis process, and shape characteristics of the welding points and the welding seams in the front resource package and the images to be welded are determined by using stretching vectorization.
6. A tack welding method for a laser welder, as claimed in claim 5, characterized in that: and in the fourth step, the YOYO neural network is utilized to take the three types of characteristics in the front resource packet as a training set, and the training set is compared with the three types of characteristics of the image to be welded to determine the welding point and the welding line in the image to be welded.
7. A tack welding method for a laser welder, as claimed in claim 6, characterized in that: and fifthly, vectorizing the Path and Row numerical values corresponding to the image by using the SVG to form a point, line and plane model, scaling the coordinates in the model and the actual coordinates according to a set proportion, and embedding the Path and Row numerical values into the positions of the welding points and the welding seams determined in the model.
8. A tack welding method for a laser welder, as claimed in claim 7, characterized in that: in the sixth step, during welding, the CCD camera array shoots welding images in real time, the synchronous master control system records the specific welding process of the laser welding robot, and timestamps the welding images shot in real time.
9. The tack welding method of a laser welding machine according to claim 8, characterized in that: and seventhly, recording the specific welding process of the laser welding robot according to the master control system, selecting welding spot and welding line data after welding of the corresponding process in the rear resource packet, wherein the data comprises all image data from the beginning to the end of welding of the whole process flow, and marking time nodes according to process time.
10. A tack welding method of a laser welder according to claim 9, characterized in that: in the step eight, the size data of the welding spot welding seam welded by the corresponding process in the resource package after being extracted by stretching vectorization is zoomed according to a set proportion, the size data of the welding image shot in real time is synchronously measured, synchronous comparison is carried out according to the time node after the welding by the corresponding process in the resource package after being shot in real time, whether the welding is normal or not is determined, and the motion track of the adaptive arm is finely adjusted when the welding is abnormal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211026155.3A CN115194323A (en) | 2022-08-25 | 2022-08-25 | Positioning welding method of laser welding machine |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211026155.3A CN115194323A (en) | 2022-08-25 | 2022-08-25 | Positioning welding method of laser welding machine |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115194323A true CN115194323A (en) | 2022-10-18 |
Family
ID=83573121
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211026155.3A Pending CN115194323A (en) | 2022-08-25 | 2022-08-25 | Positioning welding method of laser welding machine |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115194323A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116727840A (en) * | 2023-08-08 | 2023-09-12 | 深圳市青虹激光科技有限公司 | Laser drilling detection method and detection device |
-
2022
- 2022-08-25 CN CN202211026155.3A patent/CN115194323A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116727840A (en) * | 2023-08-08 | 2023-09-12 | 深圳市青虹激光科技有限公司 | Laser drilling detection method and detection device |
| CN116727840B (en) * | 2023-08-08 | 2023-11-14 | 深圳市青虹激光科技有限公司 | Laser drilling detection method and detection device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111223088B (en) | Casting surface defect identification method based on deep convolutional neural network | |
| CN109934826B (en) | Image feature segmentation method based on graph convolution network | |
| CN111307818A (en) | Visual online detection method for laser welding spot of lithium battery tab | |
| CN113077453A (en) | Circuit board component defect detection method based on deep learning | |
| CN114821114A (en) | Groove cutting robot image processing method based on visual system | |
| CN112561807B (en) | End-to-end radial distortion correction method based on convolutional neural network | |
| CN115131268A (en) | Automatic welding system based on image feature extraction and three-dimensional model matching | |
| CN112488207A (en) | High-similarity part visual identification method and system based on simulation image | |
| CN115194323A (en) | Positioning welding method of laser welding machine | |
| CN117392097A (en) | Additive manufacturing process defect detection method and system based on improved YOLOv8 algorithm | |
| CN115810133A (en) | Welding control method based on image processing and point cloud processing and related equipment | |
| CN114612412A (en) | Processing method of three-dimensional point cloud data, application of processing method, electronic device and storage medium | |
| CN113205511B (en) | Electronic component batch information detection method and system based on deep neural network | |
| CN113674218A (en) | Weld characteristic point extraction method and device, electronic equipment and storage medium | |
| CN117253062A (en) | Relay contact image characteristic quick matching method under any gesture | |
| CN111695560A (en) | Method for actively positioning and focusing crop diseases and insect pests based on convolutional neural network | |
| CN111627059A (en) | Method for positioning center point position of cotton blade | |
| CN111461230A (en) | Dental model classification and identification method based on dental stl model and two-dimensional image registration | |
| CN113670268A (en) | Binocular vision-based unmanned aerial vehicle and electric power tower distance measuring method | |
| CN117314906B (en) | Wafer grain detection method based on array distribution information guidance | |
| CN114851569B (en) | Intelligent control method and system for ion exchange membrane welding forming | |
| CN112364783B (en) | Part detection method and device and computer readable storage medium | |
| CN118060669B (en) | Automatic welding control method based on welding pool feature recognition | |
| CN113450331B (en) | Special-shaped component pin detection method | |
| CN117115482A (en) | Image local feature matching method based on multi-model fusion multi-thread reasoning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |