CN114523236A - Intelligent automatic detection platform based on machine vision - Google Patents
Intelligent automatic detection platform based on machine vision Download PDFInfo
- Publication number
- CN114523236A CN114523236A CN202210186066.9A CN202210186066A CN114523236A CN 114523236 A CN114523236 A CN 114523236A CN 202210186066 A CN202210186066 A CN 202210186066A CN 114523236 A CN114523236 A CN 114523236A
- Authority
- CN
- China
- Prior art keywords
- image
- molten pool
- welding
- module
- picture
- 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
- B23K37/00—Auxiliary devices or processes, not specially adapted to a procedure covered by only one of the preceding main groups
Abstract
The invention relates to the technical field of welding detection, and particularly discloses an intelligent automatic detection platform based on machine vision and a detection method thereof.
Description
Technical Field
The application relates to the technical field of welding detection, and particularly discloses an intelligent automatic detection platform based on machine vision.
Background
The study of weld quality control is an important component of automation of the welding process. In recent years, due to the rapid development of computer vision technology, it has become an important research direction to directly observe and capture an image of a weld puddle by using a camera and obtain characteristic information of the weld puddle through image processing. Whether a clear molten pool image can be obtained is an important link for ensuring the welding quality. Because the welding pool is dynamic, strong arc light generated in the welding process, mirror reflection of the welding pool and the like, the real-time acquisition of the information of the welding pool is difficult. The quality of image acquisition plays a crucial role in the subsequent image processing problem, and the image processing directly affects the real-time performance of the whole tracking system.
Direct vision inspection systems can be classified into active and passive types, depending on whether the imaging light source in the vision inspection system is an auxiliary light source or a light source generated by the welding region itself. The active direct visual detection method adopts auxiliary light sources such as laser and the like to artificially illuminate a welding area. Because the laser has the characteristics of high brightness, high directivity, high monochromaticity, high coherence and the like, a clearer molten pool image can be obtained by the method, but the method adopts a pulse light source with high energy density and a camera with a special electronic shutter, and the equipment is extremely expensive, so that the popularization and the application of the method in the actual production are limited. The passive direct visual sensing is a visual sensing mode which utilizes radiation light of a molten pool or reflected light of the molten pool to an electric arc or both as a signal source of a receiving device. The method does not need an auxiliary light source, only uses a common CCD camera to directly obtain the front image of the molten pool area, has simple equipment and low cost, and is suitable for production and application. At present, whether a passive or active imaging light source is adopted, the visual-based weld pool detection and control research is mainly focused on the TIG welding process. The arc combustion is stable in the TIG welding process, and the molten drop transition process is avoided. However, not only strong arc interference, but also spatter and smoke are generated in the welding process, and arc flicker is generated in the frequent short circuit transition process, which brings great difficulty to the sensing of the weld pool image.
When a molten pool image is collected, because a welding process always has a plurality of interference factors such as sound, light, electricity, heat, magnetism, smoke dust impurities and the like, background noise of the molten pool image is very strong, the molten pool image is subjected to post-processing aiming at the characteristics of a preprocessed image, image filtering is generally used for removing noise, then gray-scale morphology is used for expanding and corroding the image so as to obtain a complete and clear molten pool image, the image is generally used for TIG welding with stable arc combustion, the defects of large calculation amount and poor real-time performance exist, and partial details of the molten pool are lost by fuzzy image and expansion and corrosion processing in filtering processing, so that the inventor provides an intelligent automatic detection platform based on machine vision so as to solve the problems.
Disclosure of Invention
The invention aims to solve the problems that when the traditional method is used for preprocessing a molten pool image, the image is blurred in the filtering process, and partial details of the molten pool are lost in the expansion and corrosion processes.
In order to achieve the above object, the basic scheme of the present invention provides an intelligent automatic detection platform based on machine vision, comprising a workbench for placing a workpiece, a robot body for welding the workpiece, a welding torch arranged on the robot body, a mounting fixture for fixing a CCD camera and the welding torch, a trigger, a CCD camera and an image acquisition card, wherein the trigger comprises a signal input module, a current sensor module, a filter module, a waveform shaping conversion module, a variable counter module and a signal output module,
further, the robot body is provided with an installation shaft for installing a welding torch, and the welding torch and the CCD camera can synchronously rotate along the axis of the installation shaft.
Furthermore, the variable counter module is connected with a display technology module through signals.
Further, the invention also provides an intelligent detection method based on machine vision, which comprises the following steps:
s001, carrying out AND operation on the collected molten pool image picture A at the previous moment and the collected molten pool image picture B at the next moment;
s002, the image C obtained by the and operation is subjected to threshold segmentation.
Further, in S001, photograph a is a coordinate space TAThe image obtained is the picture B which is the coordinate space TB＝TA+d TBAThe images obtained were taken with the two photographs separated by a time Δ t.
Further, in S002, the between-class variance g ═ ω1×ω2×(μ1-μ2)2；
In the formula, the ratio of the number of pixels in the picture A to the whole image is denoted as ω1Average gray level mu of1；
The proportion of the number of pixels of the picture B to the whole image is omega2Average gray of μ2。
The principle and the effect of the scheme are as follows:
1. the method and the device carry out AND operation on the collected molten pool image picture A at the previous moment and the collected molten pool image picture B at the next moment, and carry out threshold segmentation, thereby obtaining a clear and real molten pool image.
2. The invention adopts threshold segmentation, divides the pixel set by gray level, and each obtained subset forms a region corresponding to the real scenery, and each region has consistent property, and the adjacent region layout has the consistent property, thus the invention has simple realization, small calculation amount and stable performance.
3. The invention does not need to additionally increase an auxiliary light source, but utilizes the welding pool tracking system which takes the electric welding arc light as an imaging light source, and the CCD camera is controlled by the sensor to collect the pool image at the short-circuit stage of welding, thereby obtaining the pool image information with the minimum interference, and the simple and high-efficiency image processing method is adopted to remove the noise and the splash interference, thereby obtaining the clear and accurate pool image.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
FIG. 1 shows a graph of current voltage waveforms in short circuit transition welding;
FIG. 2 illustrates a voltage signal waveform diagram during short circuit transition welding;
FIG. 3 shows a waveform of an output voltage signal at the short transition detection circuit;
FIG. 4 shows a schematic diagram of the flip-flop control hardware of the present invention;
fig. 5 shows a schematic structural diagram of an intelligent automatic detection platform based on machine vision.
Detailed Description
To further clarify the technical measures and effects of the present invention adopted to achieve the intended purpose, the following detailed description is given of specific embodiments, structures, characteristics and effects according to the present invention with reference to the accompanying drawings and preferred embodiments.
Reference numerals in the drawings of the specification include: the robot comprises a workbench 1, a workpiece 2, a welding torch 3, a robot body 4, a CCD camera 5 and a filter 6.
An intelligent automatic detection platform based on machine vision for CO2Short circuit transition welding, includingImage acquisition and image processing.
In image acquisition, as shown in fig. 5, including workstation 1 that is used for work piece 2 to place and the robot body 4 that is used for welding work piece 2, be equipped with welding torch 3 on the robot body 4, still include the CCD camera and with the CCD camera with the installation anchor clamps of welding torch 3 fixed, be equipped with the installation axle that is used for installing welding torch 3 on the robot body 4, welding torch 3 and CCD camera can follow installation axle axis synchronous revolution, and the utensil is internal, installs the axle circumference pot head and is equipped with the installation lantern ring, the installation lantern ring and welding torch 3 rigid coupling, be equipped with drive installation lantern ring pivoted motor on the robot body 4, the motor passes through the rotation of gear pair drive installation lantern ring.
In the image processing, the machine vision system comprises a trigger, a CCD camera and an image acquisition card, wherein the trigger comprises a signal input module, a current sensor module, a filter module, a waveform shaping conversion module, a variable counter module and a signal output module, and the variable counter module is in signal connection with a display technology module.
In CO2For short circuit transition welding, one short circuit period generally passes through four stages of arcing, arc gap short circuit, liquid bridge necking and breaking, arc reignition and the like; for CO2For short circuit transition welding, the arc voltage and the welding current contain rich information about the arc burning process and the droplet transition process, and there is an association relationship between the short circuit period of the welding and the current voltage, as shown in fig. 1:
the arc gap short-circuit phase helps to obtain a clear weld pool image, which is mainly represented by:
(1) in the short-circuit stage, the electric arc is extinguished, and the flickering phenomenon and the smoke interference of the electric arc do not exist;
(2) the surface temperature of the molten pool has been reduced to a relatively low degree, so there is almost no strong contrast between the welding arc and the molten pool during the arc stage, and the image gray scale distribution of the molten pool is relatively stable;
(3) splashing is rarely generated in the middle of short circuit, and noise interference generated by splashing in a molten pool image can be avoided to a great extent;
(4) when no re-arcing occurs, the electric arc impacts the molten pool, and the surface of the molten pool is relatively stable.
In the short-circuit transition, the different timings of the droplet transition can be associated with the change of the arc voltage, and as shown in fig. 2, the short-circuit transition voltage change curve is such that the arc voltage sharply drops when the droplet is short-circuited and the arc and splash disturbances are small in the short-circuit stage, so that when the exposure timing of the CCD camera is controlled to a certain timing at the arc short-circuit start stage by the external trigger, the arc intensity approaches and the disturbance is small.
Meanwhile, the exposure time of the CCD camera is set to be shorter, so that the exposure acquisition process of the image is completed in a short circuit stage, and the image with the brightness close to that and less interference can be acquired, thereby overcoming the problem of more interference such as image flickering and splashing caused by different acquisition brightness at different moments.
When the welder normally works, the working current is generally about 0-200A. Collecting voltage and current signals of the welding machine, and outputting alternating current 0-20 mA or voltage 0-5V with the same change rule as working current through conversion of a matched current sensor. Filtering is performed through a 300Hz low-pass filter, then a waveform shaping converter outputs a rectangular pulse waveform with the same frequency change as the working current of the welding machine, waveform characteristics are analyzed, short circuit transition is detected, TTL low level triggering a CCD camera is generated around the short circuit time of a molten drop, and therefore the shooting action of the camera is controlled, and the method is shown in figure 3. The flip-flop control hardware principle is shown in fig. 4.
An intelligent detection method based on machine vision comprises the following steps:
and S001, carrying out AND operation on the collected melt pool image picture A at the previous moment and the melt pool image picture B at the next moment.
In the motion process of the manipulator of the robot, the coordinate system { T } of the workpiece is known, and after delta T time, a new coordinate system represents T + dTComprises the following steps: t + d T ═ Trans (d)x,dy,dz)Rot(f,dθ)T
In the formula: trans (dx, dy, dz) represents the transformation of differential translations dx, dy and dz in the base system; rot (f, d θ) represents the transformation of differential rotation d θ around vector f in the basis system.
With Δ ═ Trans (dx, dy, dz) Rot (f, d θ) -I, the above formula can be simplified to:
in the formula: δ x, δ y and δ z are differential rotations around three axes x, y and z of the original coordinate system respectively; dx, dy, and dz are the differential translation amounts, respectively.
It is necessary to convert both photos a and B to the same space. In the process of plane welding, the movement direction of the top point of the welding torch is taken as symmetry, and the picture A is taken as a coordinate space TAThe image obtained is the coordinate space T of the photo BB＝TA+d TBAThe images obtained were taken with the two photographs separated by a time Δ t.
In the same space, the splashing arc light is dynamic quantity, the molten pool is relatively stable quantity, and the arc splashing arc light position is different due to different time. Dz is 0 due to the planar welding; from the view of welding process parameters, the walking angle and the working angle of the robot in the plane welding process are unchanged, so that the rotation around the y axis and the rotation around the x axis are zero, namely the corresponding position pixel is unchanged, the deformation of the image pixel around the x axis and the y axis is a quadratic differential, and the differential operator is
Since the micromotion of the second photograph is relative to the first, i.e.
Therefore, it is not only easy to use
In the second webIn space of (2), pixel position gB(i, j, s,1), wherein i, j, s is an integer. In the space of the first web are represented as
Therefore, it is not only easy to use
Since the coordinates of the pixels are integers, δzAnd dx,dyIt is a non-integer quantity, and therefore, must be integer-ized before the two pictures are 'and' operated. I.e. the result of the common area and after the integer:
according to the differential motion theory of robotics, the robot moves very slowly in units of milliseconds. The gray value of the image collected by the high-speed camera is changed by about 1 pixel in millisecond unit, and compared with the metal spattering movement speed generated by welding, the movement speed is very high, so that the image is similar to a static image after AND operation, and the blurring effect of a dynamic image can be ignored.
S002, the image C obtained by the and operation is subjected to threshold segmentation. And (5) carrying out AND operation on the photos A and B at the two moments before and after, and obtaining a common area of the photos A and B. The image at this time includes complete molten pool image information, no isolated noise particles exist around the molten pool, the gray level difference between the molten pool and the workpiece region is obvious, the gray level at the boundary jumps, the boundary of most of the molten pool is clearer, in order to distinguish the molten pool image from the background, the threshold segmentation needs to be carried out on the image C obtained after the AND operation, and the image is divided into a background and an object, namely the molten pool of our research object, and the other is the background.
The larger the inter-class variance between the background and the object, the larger the difference between the two parts constituting the image, and the smaller the difference between the two parts when part of the object is mistaken for the background or part of the background is mistaken for the object. Thus, a segmentation that maximizes the inter-class variance means that the probability of false positives is minimized. For image I (x, y), the segmentation threshold of foreground (target) and background is denoted as T, and the proportion of the number of pixels belonging to foreground in the whole image is denoted as omega1Average gray level mu of1(ii) a The proportion of the number of background pixels to the whole image is omega2Average gray of μ2. The total mean gray level of the image is denoted as μ and the inter-class variance is denoted as g.
Assuming that the background of the image is dark and the size of the image is M × N, the number of pixels in the image whose gray level value is less than the threshold value is denoted as N1The number of pixels having a pixel gray level greater than the threshold T is denoted by N2Then, there are:
N1+N2＝M×N
ω1+ω2＝1
μ＝μ1×ω1+μ2×ω2
g＝ω1×(μ-μ1)2+ω2×(μ-μ2)2
the above formula is arranged to obtain an equivalent formula:
g＝ω1×ω2×(μ1-μ2)2
and obtaining the threshold T which enables the inter-class variance to be maximum by adopting a traversal method, namely obtaining the threshold T. In the molten pool tracking system, the image acquired by the CCD camera is a 256-color gray scale image, and therefore the value range of T is [0, 255 ]. Therefore, when the threshold is obtained by the traversal method, 256 operations are needed, and the complexity is high. After the gray histogram of the molten pool image is researched, the gray values of pixels are mainly distributed in the range of 60-240, the image contrast is low, and the edge of the molten pool is clear. Therefore, when the threshold is obtained by using the Otsu method, the traversal interval is set to be between [30 and 80], so that the operation speed is increased. At present, when a molten pool image is processed, the molten pool image is generally directly binarized according to a threshold value, a target is set to be 0 (white), and a background is set to be 1 (black). However, the binarized image is only black and white, and the specific shape of the molten pool is difficult to observe and judge. Therefore, in the system, when the threshold value is segmented, the molten pool image is not directly binarized according to the threshold value, but pixel points with the gray value larger than the threshold value (the pixel points represent the molten pool part) are reserved, and the pixel points with the gray value smaller than the threshold value are re-assigned to be 0 (the pixel points are the background part and directly display black), so that the molten pool part is separated from the background, and the molten pool part is clearly displayed, and the observation and the judgment of the molten pool state are facilitated. When displaying the molten pool image, due to the threshold segmentation, jaggies may occur at the edge of the molten pool, and therefore, the entire image needs to be subjected to grayscale stretching to reduce the image contrast and alleviate the jaggies.
Based on the image processing method and the characteristics of dynamic change of a molten pool, large splashing and high contrast.
The invention adopts a passive molten pool vision sensing system, does not need an auxiliary light source, and utilizes self-radiation light of a molten pool and reflected arc light as a light source to acquire a molten pool image. As shown in FIG. 5, the system consists of a welding power supply, an arc welding manipulator, a high-speed CCD camera 5, a trigger, an image acquisition card and an industrial personal computer.
The optical signal collected by the CCD camera 5 consists of two parts: firstly, the electric arc and the self thermal radiation of the molten pool emit light which is directly received by the camera; secondly, the arc light irradiates the molten pool, and the molten pool reflects the arc light into the camera. To obtain high-definition images of the weld pool, the electric arc is optically suppressedThe molten pool light is received by the camera as completely as possible, and the essence of the molten pool light is to increase the signal-to-noise ratio of the molten pool image. The image signal-to-noise ratio is given by the following equation:
wherein Σ J is the sum of the metal characteristic spectrums in the image capture window;
Σ F is the total reflected light of the molten pool to the arc in the image capture window;
Σ Z is the sum of the radiation light of the molten pool in the image capture window;
Σ Q is the sum of the feature spectra of the shielding gas in the arc atmosphere within the image capture window;
Σ L is the sum of the continuous spectrum of the arc within the imaging window.
From the visible wave band to the near infrared light band, the continuous spectrum intensity of the electric arc is gradually reduced along with the increase of the wavelength, the self heat radiation of the molten pool is gradually enhanced, namely sigma Z is increased, and sigma Q is reduced, so that the high signal-to-noise ratio is favorably obtained. Although the central area of the electric arc can not be completely filtered and the head of the molten pool is partially covered by the electric arc in the near infrared wave band (such as 980nm and 1064nm) compared with the pulse stroboscopic imaging and the middle and far infrared imaging, the image definition is still higher and the characteristics of the molten pool are clear and identifiable. In the visible light band, such as the narrow-band filtering of 611nm, the radiation of the molten pool itself in this region is weak, i.e. Σ Z is small, and a high signal-to-noise ratio is to be obtained, only increasing Σ F and decreasing Σ Q, while when the neutral filter is used for the subtraction, Σ F and Σ Q are simultaneously reduced in equal proportion, so the signal-to-noise ratio is not as high as the image obtained in the near infrared region.
According to the Plunk blackbody radiation law and the Wein displacement law, the spectral radiation emittance of a molten pool increases along with the increase of the wavelength and the light intensity of an arc weakens along with the increase of the wavelength, namely the signal-to-noise ratio gradually increases. The reflected light of the arc is used as a signal by simply using the molten pool, so that a clear image is difficult to acquire. Meanwhile, the wavelength range of the sensitivity of a common CCD silicon wafer to light waves is from 400nm to 1060nm, and the sensitivity of the CCD is gradually reduced along with the increase of the wavelength in an infrared region, so that the quality of an image is influenced.
The main function of the filter 6 is to attenuate the intensity of the light radiation of the arc to a large extent, while not or less affecting the radiation intensity of the bath radiation and of the emitted light. In the test, the situation that the arc light intensity is weak when the molten pool image is collected in the short circuit stage of welding is considered, so that the narrow-band filter 6 with the near-infrared wavelength is selected, the bandwidth is 20nm, and the anti-splashing protection plate is arranged. The method is economical, easy to realize and very suitable for scientific research and actual production requirements.
The image collecting card is used to send the image video signal of the camera to the memory of the computer and VGA frame memory in frame or field unit for processing, storing, displaying, transmitting and other uses.
Although the present invention has been described with reference to the preferred embodiments, it is to be understood that the invention is not limited to the disclosed embodiments, but is intended to cover various modifications, equivalents and alternatives falling within the spirit and scope of the invention as defined by the appended claims
Although the present invention has been described with reference to the preferred embodiments, it should be understood that various changes, substitutions and alterations can be made herein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (6)
1. The utility model provides an intelligent automatic checkout platform based on machine vision which characterized in that: the welding robot comprises a workbench for placing workpieces, a robot body for welding the workpieces, a welding torch arranged on the robot body, a mounting fixture for fixing a CCD camera and the welding torch, a trigger, the CCD camera and an image acquisition card, wherein the trigger comprises a signal input module, a current sensor module, a filter module, a waveform shaping conversion module, a variable counter module and a signal output module.
2. The machine vision-based intelligent detection method according to claim 1, characterized by comprising the following steps:
s001, carrying out AND operation on the collected molten pool image picture A at the previous moment and the collected molten pool image picture B at the next moment;
s002, the image C obtained by the and operation is subjected to threshold segmentation.
3. The intelligent detection method based on machine vision according to claim 2, wherein in S001, the picture a is a coordinate space TAThe image obtained is the picture B which is the coordinate space TB＝TA+d TBAThe images obtained were taken with the two photographs separated by a time Δ t.
4. The machine-vision-based intelligent detection method according to claim 2 or 3, wherein in S002, the between-class variance g ═ ω1×ω2×(μ1-μ2)2；
In the formula, the ratio of the number of pixels in the picture A to the whole image is denoted as ω1Average gray level mu of1；
The proportion of the number of pixels of the picture B to the whole image is omega2Average gray of μ2。
5. The intelligent automatic detection platform based on machine vision according to claim 1, wherein the robot body is provided with an installation shaft for installing a welding torch, and the welding torch and the CCD camera can rotate synchronously along the axis of the installation shaft.
6. The machine-vision-based intelligent automatic detection platform is characterized in that the variable counter module is in signal connection with a display technology module.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210186066.9A CN114523236A (en) | 2022-02-28 | 2022-02-28 | Intelligent automatic detection platform based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210186066.9A CN114523236A (en) | 2022-02-28 | 2022-02-28 | Intelligent automatic detection platform based on machine vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114523236A true CN114523236A (en) | 2022-05-24 |
Family
ID=81624219
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210186066.9A Pending CN114523236A (en) | 2022-02-28 | 2022-02-28 | Intelligent automatic detection platform based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114523236A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114841999A (en) * | 2022-07-01 | 2022-08-02 | 湖南科天健光电技术有限公司 | Method and system for adjusting monitoring image of welding area |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102642068A (en) * | 2012-03-14 | 2012-08-22 | 桂林航天工业高等专科学校 | Current trigger control circuit of high-speed camera for electrical arc welding monitoring system |
| CN104002020A (en) * | 2014-05-04 | 2014-08-27 | 柳州铁道职业技术学院 | Digital camera automatic focusing system and method based on weld pool |
| CN104966099A (en) * | 2015-06-15 | 2015-10-07 | 北京航空航天大学 | Foot type classification method based on human foot images |
| CN106599790A (en) * | 2016-11-22 | 2017-04-26 | 桂林航天工业学院 | Current trigger camera module and welding pool image tracking system |
| CN108320280A (en) * | 2018-01-16 | 2018-07-24 | 南京理工大学 | The crater image method for detecting abnormality of view-based access control model clarity and contours extract |
| CN109444134A (en) * | 2018-07-10 | 2019-03-08 | 南京理工大学 | Molten bath vision optimal imaging band selection method based on spectrum analysis |
| CN110111350A (en) * | 2019-04-24 | 2019-08-09 | 桂林航天工业学院 | A kind of welding pool edge detection method, device and storage medium |
| CN111354009A (en) * | 2020-02-27 | 2020-06-30 | 西安交通大学 | Method for extracting shape of laser additive manufacturing molten pool |
| CN111696107A (en) * | 2020-08-05 | 2020-09-22 | 南京知谱光电科技有限公司 | Molten pool contour image extraction method for realizing closed connected domain |
| CN112183665A (en) * | 2020-10-28 | 2021-01-05 | 广西科技大学 | CO based on convolutional neural network2Welding pool image identification method |
-
2022
- 2022-02-28 CN CN202210186066.9A patent/CN114523236A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102642068A (en) * | 2012-03-14 | 2012-08-22 | 桂林航天工业高等专科学校 | Current trigger control circuit of high-speed camera for electrical arc welding monitoring system |
| CN104002020A (en) * | 2014-05-04 | 2014-08-27 | 柳州铁道职业技术学院 | Digital camera automatic focusing system and method based on weld pool |
| CN104966099A (en) * | 2015-06-15 | 2015-10-07 | 北京航空航天大学 | Foot type classification method based on human foot images |
| CN106599790A (en) * | 2016-11-22 | 2017-04-26 | 桂林航天工业学院 | Current trigger camera module and welding pool image tracking system |
| CN108320280A (en) * | 2018-01-16 | 2018-07-24 | 南京理工大学 | The crater image method for detecting abnormality of view-based access control model clarity and contours extract |
| CN109444134A (en) * | 2018-07-10 | 2019-03-08 | 南京理工大学 | Molten bath vision optimal imaging band selection method based on spectrum analysis |
| CN110111350A (en) * | 2019-04-24 | 2019-08-09 | 桂林航天工业学院 | A kind of welding pool edge detection method, device and storage medium |
| CN111354009A (en) * | 2020-02-27 | 2020-06-30 | 西安交通大学 | Method for extracting shape of laser additive manufacturing molten pool |
| CN111696107A (en) * | 2020-08-05 | 2020-09-22 | 南京知谱光电科技有限公司 | Molten pool contour image extraction method for realizing closed connected domain |
| CN112183665A (en) * | 2020-10-28 | 2021-01-05 | 广西科技大学 | CO based on convolutional neural network2Welding pool image identification method |
Non-Patent Citations (2)
| Title |
|---|
| 刘晓刚;谢存禧;张昌年;徐家园;: "基于弧光反射的焊缝图像获取及图像处理", 焊接学报, no. 04, pages 73 - 76 * |
| 王克鸿,汤新臣,刘永,徐越兰,虞剑: "富氩气保焊熔池信息视觉检测方法试验研究", 机械工程学报, no. 06, pages 161 - 164 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114841999A (en) * | 2022-07-01 | 2022-08-02 | 湖南科天健光电技术有限公司 | Method and system for adjusting monitoring image of welding area |
| CN114841999B (en) * | 2022-07-01 | 2022-10-11 | 湖南科天健光电技术有限公司 | Method and system for adjusting monitoring image of welding area |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN210046133U (en) | Welding seam visual tracking system based on laser structured light | |
| WO2022179186A1 (en) | Stroboscopic stepped illumination defect detection system | |
| US20200269340A1 (en) | Active Laser Vision Robust Weld Tracking System and Weld Position Detection Method | |
| CN103998177B (en) | The welding position checkout gear of laser weld and welding position detection method | |
| CN102590218B (en) | Device and method for detecting micro defects on bright and clean surface of metal part based on machine vision | |
| CN107824940A (en) | Welding seam traking system and method based on laser structure light | |
| CN108765416A (en) | PCB surface defect inspection method and device based on fast geometric alignment | |
| US5509597A (en) | Apparatus and method for automatic monitoring and control of a soldering process | |
| CN108247179B (en) | I shape groove CO based on crater image2Welding deviation detection method and device | |
| Ma et al. | Binocular vision system for both weld pool and root gap in robot welding process | |
| CN101718912B (en) | Digitalized detail visualizer of industrial X-ray negative with variable zooming ratio | |
| JP2000002518A (en) | Three dimensional input device | |
| CN110480127A (en) | A kind of seam tracking system and method based on structured light visual sensing | |
| CN109668904A (en) | A kind of optical element flaw inspection device and method | |
| CN110355465B (en) | High dynamic vision system based on industrial camera and synthesis method | |
| CN108032011A (en) | Initial point guiding device and method are stitched based on laser structure flush weld | |
| CN108067714A (en) | A kind of thin-walled circumferential weld termination quality on-line monitoring and defect positioning system and method | |
| CN114523236A (en) | Intelligent automatic detection platform based on machine vision | |
| CN104668739A (en) | Consumable electrode gas shielded-welding welding line identification visual sensor | |
| CN102519387B (en) | Visual inspection method of electron beam welding pool shape parameter | |
| JP2001509627A (en) | Method for operating a device for detecting the position of at least one characteristic region of a person and generating a control signal for a device dependent on this position | |
| CN106599790A (en) | Current trigger camera module and welding pool image tracking system | |
| CN111203655B (en) | Three-dimensional space spattering recognition device in laser welding process | |
| CN113751920A (en) | Embedded device and method for detecting welding quality of lockhole TIG welding in real time | |
| CN114932292B (en) | Narrow-gap passive vision weld joint tracking method and system |
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 |