CN105728903B - A kind of obstacle recognition method of laser type complicated track weld seam - Google Patents
A kind of obstacle recognition method of laser type complicated track weld seam Download PDFInfo
- Publication number
- CN105728903B CN105728903B CN201610311753.3A CN201610311753A CN105728903B CN 105728903 B CN105728903 B CN 105728903B CN 201610311753 A CN201610311753 A CN 201610311753A CN 105728903 B CN105728903 B CN 105728903B
- Authority
- CN
- China
- Prior art keywords
- obstacle recognition
- laser
- wag
- wig
- amplitude
- 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
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
- B23K9/00—Arc welding or cutting
- B23K9/12—Automatic feeding or moving of electrodes or work for spot or seam welding or cutting
- B23K9/127—Means for tracking lines during arc welding or cutting
- B23K9/1272—Geometry oriented, e.g. beam optical trading
- B23K9/1274—Using non-contact, optical means, e.g. laser means
Abstract
The present invention relates to a kind of obstacle recognition method of laser type complicated track weld seam, it mainly solves the problems, such as the identification of barrier during complicated track workpiece Automation of Welding.Its drip irrigation device is：Position sampling is carried out using laser type obstacle recognition sensor, the sensor is made up of laser vision signal pickup assembly, adaptive strain amplitude wig-wag, obstacle recognition controller, is fixed on walking mechanism and preposition in welding torch；Obstacle recognition controller carries out data sampling and signal transacting to laser system output signal, extracts feature locations peak value, forms the dynamic matrix of certain sample size, and the barrier attribute of each sample matrix is analyzed using multivariate statistics distinguished number；Meanwhile, the multiple feedback signal of change analyzed according to the output characteristics and distinguished number of obstacle recognition sensor swings up and down amplitude, and real-time intelligent adjusts wig-wag to ensure adaptivity of the laser on complicated track workpiece, realizes the accurate identification of barrier.
Description
Technical field
The invention belongs to Automation of Welding control device technical field, and in particular to a kind of laser type complicated track weld seam
Obstacle recognition method.
Background technology
For the welding production line of complicated track workpiece, China is also in the semiautomatic welding stage.During actual welding,
Most of identification and control avoidance action manually that barrier is carried out by people's mesh, generally existing welding efficiency is low or because of human error
And manually caused by control the problem of obstacle recognition efficiency is low, accuracy of identification is not high.With the hair of computer vision technique
Exhibition, visual sensing technology has been also applied to assisted automated such as weld and HAZ and the obstacle recognition process of weld joint tracking and correlation
In.Current existing conventional visual Sensor Analog Relay System visual performance obtains the characteristic information of weld seam and all kinds of barriers, but should
Technology has certain limitation, is only capable of the hasty obstacle on detection of straight lines or simple circle arc curved welding seam track, once occur
Parametric curve track and barrier, conventional visual sensing differentiate real-time difficult to realize and high accuracy, so as to can not expire
Foot is full-automatic.Furthermore, it is applied to butt plates welding more laser vision sensor now and stitches tracking, workpiece three-dimensional appearance is surveyed
Amount system does not possess acquisition range regulation adaptivity, requires higher to the frock clamp of weld task, Universal flexible is not high, fits
Groove scope is not wide.For to sum up, vision sensor is applied to the obstacle on complicated track Automation of Welding production line
The research work of thing identification technology is not yet ripe, the invention be intended to break the barriers distinguished number preferably with additional adaptive strain
Width wig-wag realizes the accurate of complicated track weld seam obstacle recognition and stably, is obstacle recognition, the weld seam of welding production line
The function integrations such as tracking and full-automation lay the first stone.
The content of the invention
In view of the above-mentioned problems, present invention seek to address that during complicated track workpiece Automation of Welding the identification of barrier ask
Topic.For the hasty obstacle and its accuracy and stability being only capable of at present on detection of straight lines or simple circle arc curved welding seam track
Not high the problem of, on the basis of existing visual sensing technology, the distinguished number that breaks the barriers preferably with it is additional adaptive
A kind of combination of luffing wig-wag, it is proposed that obstacle recognition method of laser type complicated track weld seam.Its technical scheme
It is：Position sampling is carried out using laser type obstacle recognition sensor, the sensor is by laser vision signal pickup assembly, adaptive
Strain amplitude wig-wag, obstacle recognition controller composition, are fixed on walking mechanism and preposition in welding torch；Obstacle recognition controller
Data sampling and signal transacting are carried out to laser system output signal, feature locations peak value is extracted, certain sample size is formed
Dynamic matrix, the barrier attribute of each dynamic matrix is analyzed using multivariate statistics distinguished number；Meanwhile, passed according to obstacle recognition
The multiple feedback signal of change that the output characteristics and distinguished number of sensor are analyzed swings up and down amplitude, Intelligent adjustment wig-wag
Amplitude of fluctuation realizes the accurate identification of barrier to ensure adaptivity of the laser on complicated track workpiece.
Laser type obstacle recognition sensor of the present invention, it is characterized in that：Laser type obstacle recognition sensor is by swashing
Light visual signal harvester, adaptive strain amplitude wig-wag, obstacle recognition controller composition, are fixed on walking mechanism and preposition
In welding torch；Adaptive strain amplitude wig-wag of the present invention, the signal acquisition range for adjusting laser vision signal pickup assembly,
To ensure adaptivity of the laser on complicated track workpiece.It is characterized in that：The device is adjusted by swing performing module and the amplitude of oscillation
Control module is constituted, and during the seam track change of workpiece, is sentenced according to the output characteristics and multivariate statistics of obstacle recognition sensor
The multiple feedback signal of other Algorithm Analysis, amplitude of oscillation regulation control module calculate above and below amplitude of oscillation regulated quantity, and driven by swing luffing
The real-time intelligent adjustment of dynamic device module control executing agency.
The specific method of the obstacle recognition of laser type complicated track weld seam of the present invention：Laser vision signal acquisition is filled
Put and be driven in adaptive strain amplitude wig-wag, with certain frequency perpendicular to weld seam reciprocally swinging, be initially located at complicated track workpiece and rise
Side is welded, walking mechanism and laser vision system synchronization are started working, laser data acquisition module from workpiece according to reflecting
Characteristic value, extracts its position peak value, and preceding p peak volume is sequentially constituted a p dimensional feature value by obstacle recognition controller successively
Matrix, often increases p/n position peak volume newly and substitutes most preceding equivalent position peak value afterwards, forms a new p dimension position square
Battle array, by that analogy, forms polynary with dimension dynamic matrix.Two groups of sample matrix are gathered, respectively there are p dimension matrixes during barrier total
P dimension matrixes when body and clear are overall, and the overall mean vector of every group of matrix, association are calculated using multivariate statistics distinguished number
The overall desired value of poor battle array, sample matrix and dynamic matrix and there are barrier matrix totality and clear matrix overall
Mahalanobis distance and discriminant function, the overall barrier attribute of every group of matrix is judged by the positive negativity of discriminant function；Acquisition process
Multigroup mean vector and desired value, according to the output characteristics of obstacle recognition sensor, obtain experience sample multiple feedback signal
With the corresponding relation of corresponding wig-wag wobble amplitude, fit respective function relation, when laser and complicated track weld seam away from
During from changing, amplitude of oscillation regulating calculation module draws amplitude of oscillation regulated quantity, and is adjusted by swing module real-time intelligent, now laser
The position on complicated track workpiece can adaptively be adjusted.
The beneficial effects of the invention are as follows：Present invention seek to address that barrier during complicated track workpiece Automation of Welding
Identification problem, proposes a kind of obstacle recognition method of laser type complicated track weld seam.In existing laser-vision sensing technology
On the basis of, the laser type obstacle recognition sensor of invention selects multivariate statistics distinguished number, can be to the number that collects in real time
According to statistical classification is carried out, barrier attribute is accurately distinguished；Invention adaptive strain amplitude wig-wag overcome it is existing be only capable of identification
The technology limitation of single straight line or the curved line tracking of simple circle, realizes adaptivity of the laser on complicated track workpiece, is multiple
The function integrations such as obstacle recognition, the weld joint tracking of miscellaneous track workpiece welding production line and full-automation lay the first stone.
Brief description of the drawings
Fig. 1 is the theory structure block diagram of the obstacle recognition of laser type complicated track weld seam.
Fig. 2 is the system schematic of the obstacle recognition of laser type complicated track weld seam.
In figure：1- adaptively swings luffing performing module, 2- laser signal acquisition processing modules, the control of 3- obstacle recognitions
Device, 4- laser vision transceiver modules, 5- adaptively swings luffing and calculated and drive module, 6- walking mechanisms, 7- laser works position
Put a little, 8- barriers, 9- complicated track weld seams.
Fig. 3 is multivariate statistics distinguished number and the adaptive schematic diagram for swinging luffing.
Fig. 4 is operating diagram of the adaptive strain amplitude wig-wag on complicated track weld seam.
Embodiment
In order to preferably express the technical scheme and beneficial effect entirely invented, with reference to the accompanying drawings and examples to this hair
It is bright to be described in further details.But, the implementation of the present invention is not limited to this.
Embodiment 1, the angle welding obstacle recognition method based on scanning type laser, the differentiating obstacle is by 1-
Adaptive to swing luffing performing module, 2- laser signal acquisition processing modules, 3- obstacle recognition controllers, 4- laser visions are received
Module is sent out, 5- adaptively swings luffing and calculated and drive module, 6- walking mechanisms composition, wherein laser type obstacle recognition are sensed
Device is by 2,3,4 module compositions, and adaptive strain amplitude wig-wag is by 1,5 module compositions；Laser vision transceiver module 4 is driven in adaptively
Luffing performing module 1 is swung, reciprocally swinging is made perpendicular to complicated track weld seam 9 with certain frequency, complicated track work is initially located at
Part rise weldering side, walking mechanism start along welding direction at the uniform velocity advance when, laser vision transceiver module receive from laser work position
The a little 7 sampling characteristic value signal reflected is put, its position peak value is extracted and by barrier through laser signal acquisition processing module 2
Identification controller 3 carries out follow-up obstacle recognition algorithm and differentiated.Refering to Fig. 1 and Fig. 2.
Embodiment 2, the laser type obstacle recognition sensor and multivariate statistics distinguished number, its embodiment exists
In：Laser vision transceiver module receives the sampling characteristic value signal reflected from laser work location point, is adopted through laser signal
Collection processing module extracts its position peak value and preceding p peak volume is sequentially constituted into a p dimension successively by obstacle recognition controller
Eigenvalue matrix, is denoted as overall X(0), often increase p/n position peak volume newly afterwards and substitute p/n most preceding position peak volume, shape
Location matrix is tieed up into a new p, overall X is denoted as(1).By that analogy, form polynary with dimension dynamic matrix X(i),(i≥0).Root
Understand that clear operating position point reflection of the laser on workpiece is returned according to the output characteristics of laser-vision sensing device used
The p dimension location matrixs of position peak value formation obey p member normal distributions.Assuming that the peak volume of collection is designated as x successively11, x12...,
x1k, x21, x22..., x2k..., xn1..., xnk, wherein k=p/n then has
Totality X is understood by above formula(i), (i >=0) is one n × k dynamic data matrixes.
Two groups of sample matrix are gathered, respectively there is p dimension sample matrix totality O during barrierbP dimensions during with clear
Sample matrix totality No, calculate every group of matrix totality X(i), the mean vector of (i >=0)The poor battle array ∑ of associationi, totality ObWith overall No
Desired value μO、μNAnd X(i)With ObAnd NoMahalanobis distance D and discriminant function W (Xi), it is specific as follows：
Matrix totality X(i), the mean vector of (i >=0)
In order to obtain linear discriminant function W (X(i)), it is considered to polynary p dimension dynamic matrixs X(i)To sample matrix totality ObAway from
From square with to sample matrix totality NoSquare distance difference：
Known μO, μN, ∑i, order
Then W (X(i))=0.5D2(X(i),Ob)-D2(X(i),No)
The overall barrier attribute of every group of matrix be can determine whether out by the positive negativity of discriminant function, as W (X(i)) ＞ 0 when, D2
(X(i),Ob)-D2(X(i),No) ＞ 0, it should now judge X(i)∈No, then decision rule be represented by
Thus it can determine whether out that any time polynary p ties up the barrier attribute of dynamic matrix, meanwhile, by the value for changing p and n
The ageing accuracy of adjustable obstacle recognition.Refering to Fig. 3, remaining same above-described embodiment.
Embodiment 3, the adaptive strain amplitude wig-wag, the signal acquisition model for adjusting laser vision signal pickup assembly
Enclose, to ensure adaptivity of the laser on complicated track workpiece.It is characterized in that：The device is adjusted by swing performing module and the amplitude of oscillation
Section control computing module and drive module composition.According to the output characteristics of obstacle recognition sensor, the position peak that must can be gathered
The fitting function relation l=f (x of the physical length of value amount and light pathij), wherein l is the physical length of light path, xijFor any bit
Put peak volume；Furthermore, the multiple feedback signal that multivariate statistics distinguished number is analyzed in a upper embodiment, including mean vector
And desired value μiDeng；Know there is relation againμi=h (xij), constructorThenRefering to Fig. 4, when the seam track of workpiece changes, p=4 is taken, can be obtained such as figure a, b, c, d corresponding 4
Individual position peak volume, is designated as (x11, x12, x21, x22), obtainBy l=f (F-1) 4 l values can be tried to achieve, it is designated as (l1,
l2, l3, l4), i.e., the distance set of laser and operating position point in one group dynamic matrix, fitting SIN function is obtained
Lj=L0(j)sin(ωt+θ)+ΔL(j)
Wherein,For apart from a reference value,For distance increment value.The amplitude of oscillation
Regulation control module can calculate wig-wag according to the variation tendency of above-mentioned two amount and swing up and down angular adjustment amount, such as following formula
Wherein, ψ=G (L) is the corresponding triangle changed with swinging up and down angle of laser and the distance measurements of operating position point
Shape solves relation.
Wig-wag, which swings up and down angular adjustment amount and can drawn, in above formula swings up and down pulsed quantity, swings luffing drive module
Control swings luffing executing agency and carries out real-time intelligent adjustment, and now laser can be adjusted adaptively on complicated track workpiece
Position, as illustrated, curve L1And L2Respectively operating position point high-low limit.Refering to Fig. 4, remaining same above-described embodiment.
Described above is only the preferred embodiment of the present invention, it is noted that under the premise without departing from the principles of the invention
Made some improvement, are all considered as protection scope of the present invention.
Claims (3)
1. a kind of obstacle recognition method of laser type complicated track weld seam, during complicated track workpiece Automation of Welding
Obstacle recognition, it is characterized in that：Laser type obstacle recognition sensor is by laser vision signal pickup assembly, adaptive strain amplitude
Wig-wag, obstacle recognition controller composition, are fixed on walking mechanism and preposition in welding torch；This method utilizes laser vision signal
Harvester and obstacle recognition controller carry out data sampling and signal transacting to laser system output signal；To sampled data
The amplitude of oscillation above and below the barrier attribute of polynary dynamic matrix, adaptive strain amplitude wig-wag is analyzed using multivariate statistics distinguished number to realize
The accurate identification of barrier.
2. a kind of obstacle recognition method of laser type complicated track weld seam according to claim 1, has invented a kind of sharp
Light formula obstacle recognition sensor, it is characterized in that：Laser vision signal pickup assembly carries out characteristic sampling to workpiece, adaptive
Strain amplitude wig-wag adjusts the signal acquisition range of laser vision signal pickup assembly, the signal for the identification controller that breaks the barriers
Processing module carries out feature locations peak extraction to output signal, the dynamic matrix of certain sample size is formed, using polynary system
Count the barrier attribute that distinguished number analyzes each dynamic matrix.
3. a kind of obstacle recognition method of laser type complicated track weld seam according to claim 1, has invented one kind certainly
Adapt to luffing wig-wag, the signal acquisition range for adjusting laser vision signal pickup assembly, to ensure laser in complicated rail
Adaptivity on mark workpiece；It is characterized in that：Adaptive strain amplitude wig-wag adjusts control module by swing performing module and the amplitude of oscillation
Composition, during the seam track change of workpiece, according to the output characteristics of obstacle recognition sensor and multivariate statistics distinguished number point
The multiple feedback signal of analysis, amplitude of oscillation regulation control module calculates amplitude of oscillation regulated quantity, and is adjusted by swing performing module real-time intelligent
It is whole.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610311753.3A CN105728903B (en) | 2016-05-12 | 2016-05-12 | A kind of obstacle recognition method of laser type complicated track weld seam |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610311753.3A CN105728903B (en) | 2016-05-12 | 2016-05-12 | A kind of obstacle recognition method of laser type complicated track weld seam |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105728903A CN105728903A (en) | 2016-07-06 |
| CN105728903B true CN105728903B (en) | 2017-10-13 |
Family
ID=56288325
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610311753.3A Active CN105728903B (en) | 2016-05-12 | 2016-05-12 | A kind of obstacle recognition method of laser type complicated track weld seam |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105728903B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106446962B (en) * | 2016-10-12 | 2019-10-18 | 湘潭大学 | A kind of crossbeam barrier dynamic clustering recognition methods |
| JP7028658B2 (en) * | 2018-01-30 | 2022-03-02 | 株式会社神戸製鋼所 | Weaving control method and weaving control system |
| CN112488201A (en) * | 2020-11-30 | 2021-03-12 | 湖南艾克机器人有限公司 | Weld obstacle identification method based on concave-convex radius complex function |
| CN113385779B (en) * | 2021-07-05 | 2022-09-27 | 湘潭大学 | Large-break-angle welding seam tracking and obstacle prediction system based on binocular four-line vision sensing |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPS55144383A (en) * | 1979-04-26 | 1980-11-11 | Shin Meiwa Ind Co Ltd | Automatic welding robot |
| JPS60130473A (en) * | 1983-12-16 | 1985-07-11 | Mitsubishi Heavy Ind Ltd | Nc automatic welding method |
| JPH02205266A (en) * | 1989-01-31 | 1990-08-15 | Kobe Steel Ltd | Detection of groove route gap by memory and reproduction type arc welding robot |
| CN101554672B (en) * | 2009-05-21 | 2011-05-04 | 山东大学 | Detection and control system for container corrugated plate welding track based on laser ranging |
| CN101745765B (en) * | 2009-12-15 | 2012-02-08 | 哈尔滨工业大学 | Man-machine collaboration shared control remote welding method |
| CN103551710A (en) * | 2013-11-13 | 2014-02-05 | 上海工业自动化仪表研究院 | Welding seam tracking method in membrane wall welding process |
-
2016
- 2016-05-12 CN CN201610311753.3A patent/CN105728903B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN105728903A (en) | 2016-07-06 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105728903B (en) | A kind of obstacle recognition method of laser type complicated track weld seam | |
| CN102615390B (en) | Swing arc-based multi-layer and multi-channel weld tracking system and identification method thereof | |
| CN108982546A (en) | A kind of intelligent robot gluing quality detecting system and method | |
| CN113427168A (en) | Real-time welding seam tracking device and method for welding robot | |
| CN104985289B (en) | Laser sensor-based welding seam automatic tracking test device and test method thereof | |
| CN102279190B (en) | Image detection method for weld seam surface defects of laser welded plates of unequal thickness | |
| Xue et al. | Autonomous agricultural robot and its row guidance | |
| Fang et al. | Visual seam tracking system for butt weld of thin plate | |
| CN102455171B (en) | Method for detecting geometric shape of back of tailor-welding weld and implementing device thereof | |
| CN204524618U (en) | A kind of automatic recognition and tracking welder of weld seam applying laser scanner | |
| CN110202273A (en) | A kind of automatic height regulating device and its control method based on laser displacement sensor | |
| CN203484784U (en) | Welding system for corner welding lines and lap welding lines of corrugated plate of container | |
| CN105328304A (en) | Welding seam starting point automatic position searching method based on statistics | |
| CN104493336A (en) | Welded joint detecting and tracking system and method based on video analysis | |
| Hou et al. | A teaching-free welding method based on laser visual sensing system in robotic GMAW | |
| CN101216288A (en) | Crop plant height detection method | |
| CN104942490B (en) | Method based on the identification of ultrasonic sensor angle welding starting point with location | |
| CN106425025A (en) | Automatic fillet weld welding control method and device for laser displacement sensor | |
| CN102721390A (en) | Speed optimization method for complex curved surface contact type tracking scan and measurement | |
| CN207077080U (en) | Weld Seam Tracking Control device based on cross laser | |
| CN106964875A (en) | A kind of gun welder space gesture recognition method based on arc sensor | |
| CN109226936B (en) | Rotary arc type self-adaptive complex curved surface surfacing method | |
| CN107449383A (en) | A kind of section of jurisdiction automatic identification grabbing device and method | |
| CN205437436U (en) | Mechanical arm moving orbit adjustment system | |
| CN103817450A (en) | Automatic plate boundary detecting and cutting device and cutting method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |