CN111179233B - Self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts - Google Patents
Self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts Download PDFInfo
- Publication number
- CN111179233B CN111179233B CN201911330636.1A CN201911330636A CN111179233B CN 111179233 B CN111179233 B CN 111179233B CN 201911330636 A CN201911330636 A CN 201911330636A CN 111179233 B CN111179233 B CN 111179233B
- Authority
- CN
- China
- Prior art keywords
- image
- deviation
- self
- values
- adaptive
- 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.)
- Active
Links
Images
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/13—Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/181—Segmentation; Edge detection involving edge growing; involving edge linking
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Quality & Reliability (AREA)
- Image Processing (AREA)
- Laser Beam Processing (AREA)
Abstract
The invention provides a self-adaptive correction method based on laser cutting of two-dimensional parts, which comprises signal acquisition, signal processing, deviation extraction and self-adaptive correction, wherein in the cutting process, information is processed through an upper computer when a track and a characteristic curve deviate through two-dimensional characteristics acquired in real time, and a correction system is controlled to realize automatic correction. According to the invention, the self-adaptive real-time correction is realized through robot vision, the positioning precision of the fixture is not limited, a sensor is not needed for acquiring the characteristic information, the cutting error caused by the processing technology defect of the thin-wall plate is effectively solved, and the self-adaptive correction of the cutting path can be realized.
Description
Technical Field
The invention belongs to the fields of laser application technology and machine vision, and particularly relates to a self-adaptive deviation correcting method based on laser cutting of two-dimensional parts.
Background
In recent years, the industrial high-technology industry exceeds the traditional industry, the laser processing technology and equipment are used as modern high-technology processing means, more and more traditional industries rely on the laser processing technology to improve the processing quality of products or solve the difficult problem that the traditional processing method and process cannot solve, the high-technology and high-growth laser industry is rapidly developed, and the application and the technical transformation of the laser technology are successful for partial traditional industry transformation.
When the laser cutting machine is used for cutting two-dimensional parts in the automobile industry, the characteristics of the two-dimensional parts are complicated due to the complex characteristics of the parts and defects (such as rebound or wrinkling and other shape defects) caused by the stamping forming technology, so that the laser cutting cannot be normally applied to a large number of cutting actions. And when facing the two-dimensional part structures, not every two-dimensional part structure has a regular contour curve, laser cutting cannot be performed normally,
aiming at the problems, the invention provides a self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts. The visual detection technology is introduced into the robot self-adaptive correction system, the image processing and recognition technology is adopted to extract the characteristics of the workpiece, and the curve characteristics are optimized in a curve fitting mode, so that the outline of the workpiece is obtained. The self-adaptive deviation correcting system mainly comprises a driving unit, a stepping motor and a cross sliding frame. The control system obtains the deviation amount according to the image processing result, and obtains corresponding control signals through calculation and processing, and drives the stepping motor and the cross sliding frame to realize automatic correction. Thereby realizing automatic correction of laser cutting and improving the efficiency and accuracy of laser cutting.
Disclosure of Invention
Aiming at the technical problems, the invention provides a self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts, which comprises the following steps:
(1) And (3) signal acquisition: integrating a camera with a DSP-based processing subsystem to form a vision sensor in a robot vision system;
(2) And (3) signal processing: obtaining meaningful features from an image by adopting an improved edge detection algorithm, wherein the image is divided into T, F, I and respectively comprises an object, a background and a noise edge, and preprocessing of the image adopts dual-kernel Gaussian elimination and gray level processing; selecting three thresholds a, b and c for segmentation according to the histogram of the image, selecting the first and last pixel values larger than the average value as a, c values and b values as pixel average values according to the average value Hmaxmean of the local peak value of the histogram, and obtaining T, F and I through the following formulas:
F＝1-T
I(i,j)＝devia(i,j)/devia max×dis(i,j)/dis max
wherein img is the image pixel value, devia is the average difference value of eight connected regions, dis is the image discontinuity,
obtaining edge images according to the threshold opt, opf and opi, wherein the following formula is as follows:
if(F＜opf or T＜opt)and(I＜opi):
Edge＝0
else:
Edge＝1
wherein opi is set to 0.1, opf and opt are selected according to the maximum criterion, such as opt, and according to the T pixel histogram, the minimum pixel value k is selected as the zero-time threshold, and according to the maximum criterion, k values are sequentially increased to obtain opt, and the specific formula is as follows:
p i ＝n i /N
where ni is the number of pixel values, N is the image size, wk is the sum of probabilities of the previous k pixel values, η is the pixel expectation,
is the maximum criterion;
(3) Deviation extraction: acquiring an edge image in real time by utilizing the step (2), performing upper computer programming treatment, performing curve fitting, comparing with an ideal cutting curve, and further acquiring a track deviation signal (x, y deviation);
(4) Self-adaptive deviation correction: after the deviation signal is obtained in the step (3), the extracted deviation value is sent to a deviation correction control module, and the stepping motor is controlled by the upper computer to drive the cross sliding frame to complete deviation compensation, so that the deviation values in the x direction and the y direction are corrected.
Furthermore, in step (1), the vision sensor integrates a light source, a camera, an image processor and a standard control and communication interface, so that the vision sensor becomes an intelligent image acquisition and processing unit, an internal program memory can store image processing algorithms, a PC (personal computer) can be used, and various algorithms can be programmed by using special configuration software and downloaded into the program memory of the vision sensor.
The beneficial effects are that: the invention provides a self-adaptive correction method based on laser cutting of two-dimensional parts, which is used for cutting of the two-dimensional parts, realizes self-adaptive real-time correction through robot vision, is not limited to the positioning precision of a clamp, does not need a sensor for acquiring characteristic information, effectively solves the cutting error caused by the processing technology defect of a thin-wall plate, and can realize self-adaptive correction of a cutting path. Meanwhile, the device is convenient to install, convenient to position and quick in deviation correction speed, can adapt to the requirements of a production line, can ensure the processing quality in the production process, is superior to the parts processed by the traditional laser cutting, avoids the defect from being continued to the next process, and reduces the cost caused by the related defect.
Drawings
FIG. 1 is a diagram showing the construction of a laser cutting adaptive deviation correcting structure according to the present invention.
Detailed Description
The present invention will be described in further detail with reference to examples and drawings, but embodiments of the present invention are not limited thereto.
As shown in fig. 1, the main purpose of the present invention is to provide a self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts, which effectively solves the cutting errors caused by the defects of the processing technology of the thin-wall plate, and includes signal acquisition, signal processing, deviation extraction and self-adaptive deviation rectifying.
Signal acquisition
The source of information for the machine vision system is from an image, the acquisition of which provides the most primitive information data, and all information processing is based on the acquired image. The acquired image thus directly determines the final achievable result, a well-practical acquisition system being an important component of the machine vision system.
The camera and the DSP-based processing subsystem are integrated together to form a vision sensor in the robot vision system. The vision sensor integrates a light source, a camera, an image processor and a standard control and communication interface, and is an intelligent image acquisition and processing unit, an internal program memory can store image processing algorithms, a PC (personal computer) can be used for programming various algorithms by using special configuration software, and the algorithms are downloaded into the program memory of the vision sensor. The vision sensor combines the flexibility of the PC, the reliability of the PLC, and the distributed network technology. The visual sensor and the PLC can be used for more easily forming a robot visual system so as to realize image acquisition. In addition, in the machine vision application of laser cutting, further pretreatment, respectively gray scale treatment, filtering and drying, is needed, so that rapid and accurate characteristic information extraction is facilitated.
The system adopts a dual-core Sobel composite filter, eliminates clutter, retains effective signals, reduces the calculated amount of image processing and ensures the real-time performance of the system.
Signal processing
In many computer vision tasks, a central problem is deriving meaningful features from images, which are used to solve practical problems such as object recognition, classification, or object tracking. The key issue is of course what is significant in the image, and what form of description should be taken. The human visual system can easily and accurately solve the problems. Methods for information (image) processing are divided into two types, namely a traditional operator algorithm and a deep learning model.
The present invention employs a new and improved edge detection algorithm. The image is divided into T, F, I, containing objects, background, noise edges, respectively. The preprocessing of the image adopts dual-kernel Gaussian elimination and gray processing; selecting three thresholds a, b and c for segmentation according to the histogram of the image, selecting the first and last pixel values larger than the two pixels as a, c values and b values as pixel average values according to the average Hmaxmean of the local peak values of the histogram, and obtaining T, F and I through the following formulas:
F＝1-T
I(i,j)＝devia(i,j)/devia max×dis(i,j)/dis max
wherein img is an image pixel value, devia is an eight-connected-region average difference value, and dis is image discontinuity.
Obtaining edge images according to the threshold opt, opf and opi, wherein the following formula is as follows:
if(F＜opf or T＜opt)and(I＜opi):
Edge＝0
else:
Edge＝1
wherein opi is set to 0.1, opf and opt are selected according to the maximum criterion, such as opt, and according to the T pixel histogram, the minimum pixel value k is selected as the zero-time threshold, and according to the maximum criterion, k values are sequentially increased to obtain opt, and the specific formula is as follows:
p i ＝n i /N
where ni is the number of pixel values, N is the image size, wk is the sum of probabilities of the previous k pixel values, η is the pixel expectation,
as a rule of maximum discrimination (rule of maximum discrimination),
finally, in order to obtain better precision, a skeleton refining algorithm is adopted to obtain a final edge image.
Deviation extraction
And (3) acquiring image information in real time by utilizing the step (2), performing upper computer programming treatment, performing curve fitting, and comparing with an ideal cutting curve to acquire a track deviation signal (x, y deviation).
Self-adaptive deviation correction
The final goal is to have the machine take control of the laser spray head so the actuator is the last component in the machine vision system. Once the vision software has completed the image analysis, it is then necessary to communicate with an external unit to complete control of the production process. The correction function is realized. The self-adaptive deviation rectifying comprises a deviation rectifying control module which mainly comprises an upper computer, a stepping motor and a cross sliding frame.
After the deviation signal is obtained in step (3), the deviation value needs to be converted into a specific control signal. The method is characterized in that the extracted deviation amount is sent to a deviation correction control module, a stepping motor is controlled by an upper computer to drive a cross carriage to complete deviation compensation, deviation values in the x direction and the y direction are corrected, the cutting speed at the moment is reduced (the cutting speed is ensured to be higher than the minimum speed of the cutting quality), and then the self-adaptive deviation correction function is realized.
Claims (2)
1. The self-adaptive deviation rectifying method based on laser cutting of the two-dimensional parts comprises the following steps:
(1) And (3) signal acquisition: integrating a camera with a DSP-based processing subsystem to form a vision sensor in a robot vision system;
(2) And (3) signal processing: obtaining meaningful features from an image by adopting an improved edge detection algorithm, wherein the image is divided into T, F, I and respectively comprises an object, a background and a noise edge, and preprocessing of the image adopts dual-kernel Gaussian elimination and gray level processing; selecting three thresholds a, b and c for segmentation according to the histogram of the image, selecting the first and last pixel values larger than the two pixels as a, c values and b values as pixel average values according to the average Hmaxmean of the local peak values of the histogram, and obtaining T, F and I through the following formulas:
F＝1-T
I(i，j)＝devia(i,j)/devia max×dis(i，j)/dis max
wherein img is the image pixel value, devia is the average difference value of eight connected regions, dis is the image discontinuity,
obtaining edge images according to the threshold opt, opf and opi, wherein the following formula is as follows:
if(F＜opf or T＜opt)and(I＜opi):
Edge＝0
else:
Edge＝1
wherein opi is set to 0.1, opf and opt are selected according to the maximum criterion, such as opt, and according to the T pixel histogram, the minimum pixel value k is selected as the zero-time threshold, and according to the maximum criterion, k values are sequentially increased to obtain opt, and the specific formula is as follows:
p i ＝n i /N
where ni is the number of pixel values, N is the image size, wk is the sum of probabilities of the previous k pixel values, η is the pixel expectation,
is the maximum criterion;
(3) Deviation extraction: acquiring an edge image in real time by utilizing the step (2), performing upper computer programming treatment, performing curve fitting, comparing with an ideal cutting curve, and further acquiring a track deviation signal (x, y deviation);
(4) Self-adaptive deviation correction: after the deviation signal is obtained in the step (3), the extracted deviation value is sent to a deviation correction control module, and the stepping motor is controlled by the upper computer to drive the cross sliding frame to complete deviation compensation, so that the deviation values in the x direction and the y direction are corrected.
2. The adaptive correction method based on laser cutting two-dimensional parts according to claim 1, wherein in the step (1), the vision sensor integrates a light source, a camera, an image processor and a standard control and communication interface, and becomes an intelligent image acquisition and processing unit, an internal program memory can store image processing algorithms, and a PC can be used for programming various algorithms by using special configuration software to download the algorithms into the program memory of the vision sensor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911330636.1A CN111179233B (en) | 2019-12-20 | 2019-12-20 | Self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911330636.1A CN111179233B (en) | 2019-12-20 | 2019-12-20 | Self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111179233A CN111179233A (en) | 2020-05-19 |
| CN111179233B true CN111179233B (en) | 2023-05-05 |
Family
ID=70650256
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911330636.1A Active CN111179233B (en) | 2019-12-20 | 2019-12-20 | Self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111179233B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111736528A (en) * | 2020-07-07 | 2020-10-02 | 华中科技大学 | Laser cutting automatic programming system based on vision deviation rectification |
| CN113478552B (en) * | 2021-07-12 | 2023-08-01 | 深圳正实智能装备有限公司 | Method and device for early warning of PCB cutting path |
| CN116184929B (en) * | 2023-04-27 | 2023-07-25 | 湖南隆深氢能科技有限公司 | Intelligent control method and system applied to cutting equipment |
| CN116780971B (en) * | 2023-08-25 | 2023-10-31 | 无锡弘德科技有限公司 | Galvanometer motor drive control system of digital-analog hybrid architecture |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104259670A (en) * | 2014-07-07 | 2015-01-07 | 河海大学常州校区 | Turbine blade laser cutting system based on machine vision and industrial robot |
| WO2015118080A1 (en) * | 2014-02-07 | 2015-08-13 | Blackbird Robotersysteme Gmbh | Method and device for laser welding or cutting with a dynamically adaptable analysis region |
| EP3108311A2 (en) * | 2014-07-11 | 2016-12-28 | Trumpf Werkzeugmaschinen GmbH + Co. KG | Method, machining unit and computer program product for the image-based positioning of workpiece machining processes |
| CN106952281A (en) * | 2017-05-15 | 2017-07-14 | 上海交通大学 | A kind of method that weld profile feature recognition and its welding bead are planned in real time |
| KR20180034767A (en) * | 2016-09-27 | 2018-04-05 | 주식회사 제이스텍 | Method for manufacturing laser cutting for film surface glass |
| CN108846819A (en) * | 2018-07-04 | 2018-11-20 | 深圳市创客工场科技有限公司 | Parameters of laser cutting acquisition methods and device, electronic equipment, storage medium |
| CN109226967A (en) * | 2018-07-25 | 2019-01-18 | 同高先进制造科技（太仓）有限公司 | A kind of steady seam tracking system of active laser vision for Laser-Arc Hybrid Welding |
| CN110193670A (en) * | 2019-04-18 | 2019-09-03 | 苏州迈为科技股份有限公司 | A kind of OLED cutting equipment compensation system and method |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE10112445C2 (en) * | 2001-03-13 | 2003-03-20 | Fraunhofer Ges Forschung | Process for cutting components to be joined with laser radiation |
| DE102010020183B4 (en) * | 2010-05-11 | 2013-07-11 | Precitec Kg | Laser cutting head and method for cutting a workpiece by means of a laser cutting head |
| US10043283B2 (en) * | 2016-05-31 | 2018-08-07 | Servo-Robot, Inc. | Process tracking laser camera with non-eye-safe and eye-safe operating modes |
-
2019
- 2019-12-20 CN CN201911330636.1A patent/CN111179233B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015118080A1 (en) * | 2014-02-07 | 2015-08-13 | Blackbird Robotersysteme Gmbh | Method and device for laser welding or cutting with a dynamically adaptable analysis region |
| CN104259670A (en) * | 2014-07-07 | 2015-01-07 | 河海大学常州校区 | Turbine blade laser cutting system based on machine vision and industrial robot |
| EP3108311A2 (en) * | 2014-07-11 | 2016-12-28 | Trumpf Werkzeugmaschinen GmbH + Co. KG | Method, machining unit and computer program product for the image-based positioning of workpiece machining processes |
| KR20180034767A (en) * | 2016-09-27 | 2018-04-05 | 주식회사 제이스텍 | Method for manufacturing laser cutting for film surface glass |
| CN106952281A (en) * | 2017-05-15 | 2017-07-14 | 上海交通大学 | A kind of method that weld profile feature recognition and its welding bead are planned in real time |
| CN108846819A (en) * | 2018-07-04 | 2018-11-20 | 深圳市创客工场科技有限公司 | Parameters of laser cutting acquisition methods and device, electronic equipment, storage medium |
| CN109226967A (en) * | 2018-07-25 | 2019-01-18 | 同高先进制造科技（太仓）有限公司 | A kind of steady seam tracking system of active laser vision for Laser-Arc Hybrid Welding |
| CN110193670A (en) * | 2019-04-18 | 2019-09-03 | 苏州迈为科技股份有限公司 | A kind of OLED cutting equipment compensation system and method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111179233A (en) | 2020-05-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111179233B (en) | Self-adaptive deviation rectifying method based on laser cutting of two-dimensional parts | |
| CN111721259B (en) | Underwater robot recovery positioning method based on binocular vision | |
| CN112819772B (en) | High-precision rapid pattern detection and recognition method | |
| CN111191546A (en) | Intelligent product assembling method based on machine vision recognition | |
| CN111951238A (en) | Product defect detection method | |
| CN111784655B (en) | Underwater robot recycling and positioning method | |
| CN110298840B (en) | Yarn defect detection method based on image | |
| CN113570631A (en) | Intelligent pointer instrument identification method and equipment based on images | |
| CN113822810A (en) | Method for positioning workpiece in three-dimensional space based on machine vision | |
| CN112381783A (en) | Weld track extraction method based on red line laser | |
| CN111881748B (en) | Lane line visual identification method and system based on VBAI platform modeling | |
| CN117058182A (en) | Target detection frame optimization method for visual ranging | |
| CN111738264A (en) | Intelligent acquisition method for data of display panel of machine room equipment | |
| CN116433514A (en) | Vehicle door defect image correction method based on improved space transformation network STN | |
| CN116665097A (en) | Self-adaptive target tracking method combining context awareness | |
| CN116071360A (en) | Workpiece appearance defect detection method, electronic equipment and storage medium | |
| CN113592947B (en) | Method for realizing visual odometer by semi-direct method | |
| CN115018751A (en) | Crack detection method and system based on Bayesian density analysis | |
| Grigorescu et al. | Controlling Depth Estimation for Robust Robotic Perception | |
| CN112950787A (en) | Target object three-dimensional point cloud generation method based on image sequence | |
| CN110728281A (en) | License plate segmentation and recognition method | |
| CN110059540B (en) | Image processing method for recognizing and positioning loose core rivets | |
| CN112614182B (en) | Deep learning-based method for identifying machining position, storage device and mobile terminal | |
| CN111242159B (en) | Image recognition and robot automatic positioning method based on product features | |
| CN118052973A (en) | Image recognition method for optical fiber positioning |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |