CN101961819B - Device for realizing laser welding and seam tracking and control method thereof - Google Patents

Abstract

The invention relates to a device for realizing laser welding and seam tracking and a control method thereof. The device of the invention comprises an image collecting and processing unit for collecting and processing the image data of laser welding and seam characteristics and outputting the seam position information, a motion control unit for receiving the seam position information and transmitting a motion control command to a motion shaft actuating mechanism, and an upper parameter setting and control unit for setting the parameters and being respectively connected with the image collecting and processing unit and the motion control unit in a communication mode. The method of the invention comprises the following steps: setting the parameters by utilizing the upper parameter setting and control unit; collecting and processing seam images to obtain the seam image data; caching and calibrating the seam image data, and determining the offset of a welding gun; carrying out filtering, fitting and track planning on tracking tracks by utilizing a motion controller in the motion control unit; and transmitting the result to a motion shaft driver. The invention can be suitable for various seam types and techniques and has the advantages of good generality, strong flexibility, high independence, high precision, accurate positioning, strong adaptability, and the like.

Description

A kind of laser weld seam tracking device and control method thereof

Technical field

Vision measurement and intelligent motion control technology, specifically a kind of laser weld seam tracking device and control method thereof in the present invention and the robot building field.

Background technology

At present, in manufacturing machine people's welding procedure is produced, generally adopt the mode of manually control, robot teaching or off-line programing to carry out path planning and movement programming, just simply repeat predefined action in the welding process.But in laser weld, because it has the technological requirement of high-speed, high precision, therefore higher requirement has been proposed laser welding apparatus.

In existing weld joint tracking implementation, the general sensors such as machinery, electromagnetism, vision that adopt provide weld seam information, are processed by system, realize welding torch position control.But such seam tracking system generally seals, and is sensor and Processing Algorithm for particular system exploitation, lacks general, real-time seam tracking system flexibly and independently.External had the product of ripe laser weld seam tracking system of part, but the comparison system performance is each has something to recommend him, be difficult on the precision and satisfy simultaneously the requirement of laser weld high-speed, high precision on the speed, and system architecture is huge, it is complicated to consist of.

Summary of the invention

For above shortcomings part in the prior art, the technical problem to be solved in the present invention provides a kind of have precision height, location standard, adaptable laser weld seam tracking device and control method thereof.

For solving the problems of the technologies described above, the technical solution used in the present invention is:

A kind of laser weld seam tracking device of the present invention comprises:

IMAQ and processing unit, welding seam position information is processed and exported to the view data of collection laser weld characteristics of weld seam to this view data;

Motion control unit receives above-mentioned welding seam position information and carries out analyzing and processing, sends motion control instruction to kinematic axis executing agency;

Upper-layer parameters arranges and control module, and IMAQ and processing parameter, motion control parameter are set, and carries out communication and is connected with IMAQ and processing unit, motion control unit respectively.

Described IMAQ and processing unit comprise LASER Light Source, optical module and smart camera, and wherein a word linear laser of LASER Light Source generation reflects by welding work pieces, is gathered by smart camera through optical module; Described laser source wavelength is 660nm, and power is that 50mw～100mw is adjustable; Optical filter in the described optical module can see through the light wave of wavelength 660nm.

Described motion control unit comprises motion controller, kinematic axis driver and kinematic axis executing agency, wherein:

Motion controller links to each other by the CAN bus with the smart camera of IMAQ and processing unit take DSP as control core, and the positional information of its input is analyzed, calculated, and obtains the control position information of kinematic axis executing agency;

The kinematic axis driver, the control position information of reception motion controller exports kinematic axis executing agency to, and compensating motion is corrected in the tracking of final welding gun butt welded seam position of realizing.

The control method of a kind of laser weld seam tracking device of the present invention is characterized in that may further comprise the steps:

Beginning is by upper-layer parameters setting and control module parameters;

By IMAQ and processing unit Real-time Collection weld image, and carry out image algorithm and process and to obtain the weld image data;

With the weld image data buffer storage, the rower of going forward side by side is processed surely, determines the welding gun side-play amount that current collection image is corresponding;

Motion controller in the motion control unit carries out filtering, match and the trajectory planning of pursuit path according to above-mentioned welding gun side-play amount;

Above-mentioned result of calculation is given the kinematic axis driver kinematic axis executing agency is implemented control.

Described parameter comprises filtering, the match control algolithm parameter of laser instrument light intensity regulating parameter, workpiece parameter, image processing algorithm parameter and pursuit path computational process.

Described image algorithm is processed by FPGA hardware programming and dsp software programming and is combined to realize, wherein realizes the image preprocessing process by the FPGA hardware programming, may further comprise the steps:

Window to gathering weld image, obtain area-of-interest;

The region of interest area image is carried out medium filtering obtain image after the filtering;

Image after the filtering is carried out the binary picture image intensifying to be processed;

Image after the binaryzation is implemented the morphology processing of dilation and erosion;

Image behind the dilation and erosion is carried out edge extracting;

Image after edge extracts carries out central line pick-up;

By dsp software programming realization character point leaching process, may further comprise the steps:

Adopt least square method to carry out fitting a straight line to processed the center line that obtains by FPGA;

By finding maximum corner position to find characteristic point, realization character point extracts.

Described definite welding gun side-play amount corresponding to current collection image may further comprise the steps:

Optical plane uniformly-spaced is divided into rectangular mesh by density, and each grid vertex is the characteristic point of prior demarcation;

In as the plane, have and the corresponding grid of optical plane, carry out the actual measurement of grid each point by Position Control, obtain the corresponding optical plane physical coordinates of image coordinate (u, the v) (x on irregular grid summit in the picture plane _w, y _w, z _w), obtain corresponding relation database table;

Process the characteristics of weld seam dot information that obtains for above-mentioned image, be any to be calibrated some P (u in the picture plane, v), go out its physical coordinates by the little quadrilateral mesh interpolation calculation of surrounding this point to be calibrated in the above-mentioned related table of having set up, obtain the three-dimensional coordinate P (x of actual welds point _w, y _w, z _w).

Determine to be fixed in the tracking displacement of the welding gun butt welded seam in the kinematic axis executing agency by three-dimensional coordinate (x, y, z) and camera and welding gun physical location relation.

The filtering of described pursuit path, match are also calculated trace point and movement instruction planning function in real time, and implementation procedure may further comprise the steps:

Motion controller adopts the welding gun offset information of timer interrupt mode reading images acquisition and processing unit;

The welding gun offset information that apparent distance scope obtains before the motion controller pursuit path filtration module is as the filtering input message, filtering parameter (such as intermediate value or Mean Filtering Algorithm and 3 or 5 filtering parameters) by the host computer transmission carries out the selection of algorithm and parameter, 2 times of forward sight distances are set as the length of filter window, finish filtering and calculate, output is through the welding gun side-play amount after the filtering;

The pursuit path fitting module with the output quantity of filtration module as input, fitting parameter by host computer transmission carries out the selection of approximating method (such as least square or Cubic Spline Fitting) and match window size (such as parameter setting or forward sight apart from the multiple setting), finish the Fitting Calculation, and the division that the line segment after the match is controlled the welding gun side-play amount according to the time beat synchronous with gathering beat is calculated in real time;

The output quantity of tracking trajectory planning module after with the Fitting Calculation by leading deviation computational methods, calculated the rectification side-play amount that current welding gun need to be followed the tracks of compensating motion as input, and carries out movement instruction planning according to the control beat of 10 milliseconds of predetermined multiples.

Described leading deviation computational methods are as follows:

When welding gun forward travel distance during less than or equal to L, welding gun from the position " n " correction value of " n+1 " is to the position:

${\mathrm{Delta}}_{n+1}={\mathrm{Dist}}_{n+1}-\underset{p=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(n+1\≤M,n,p\⊆Z);$ The welding gun forward travel distance is less than or equal to L; Wherein:

M is the forward sight distance L length welding gun beat number that advances, and namely welding gun begins the processing picture number before the real-time tracking;

Welding gun when n is the leading deviate of current calculating advances and controls the last beat number of beat;

P is a certain beat before the current location;

Dist _N+1The distance that departs from initial position for the current location welding gun;

Delta _pCorrection value for the relatively last beat of a certain beat horizontal level before the current location;

Delta _N+1Correction value for the relatively last beat of current location position.

The correction value computing formula is behind the welding gun in-position " M ":

${\mathrm{Delta}}_{M+n+1}={\mathrm{Dist}}_{M+n+1}-\underset{p=n+2}{\overset{M+n}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(n,p\⊆Z);$

Wherein:

M, the p implication is the same;

N controls the last beat number of beat for beginning from initial position to advance apart from the current location welding gun that begins length after to calculate through first forward sight;

Dist _M+n+1The distance that departs from initial position for the current location welding gun;

Delta _pCorrection value for the relatively last beat of a certain beat horizontal level before the current location;

Delta _M+n+1Correction value for the relatively last beat of current location position.

The present invention has following beneficial effect and advantage:

1. use the inventive method and device can adapt to multiple weld seam type and technique, locate in the time of can realizing the high-precision real of position while welding, have the advantages such as versatility is good, flexibility is strong and independence is high and precision is high, location standard, strong adaptability.The present invention can adapt to V-arrangement interface, overlap joint and etc. the multiple weld seam types such as slab, unequal thickness plate docking, gather the laser welded seam image, extract the weld bead feature points position; Smart camera provides network interface to realize that host computer arranges the running parameter of camera, and the weld joint tracking position data that provides the CAN communication interface to carry out after image is processed is transmitted; Motion control DSP has to the simulated power control interface of linear laser device and to the signal interactive interface of PLC; After finishing parameter and arrange for a kind of operating mode, system can break away from the automatic weld joint tracking of host computer complete independently.

2. the present invention is directed to V-arrangement interface, overlap joint with etc. the multiple weld seam types such as slab, unequal thickness plate docking different operating condition designs different parameters be used for the weld image processing module, and according to smart camera and the connected mode that connects firmly of following the tracks of executing agency, developed leading deviation Weld Seam Tracking Control algorithm, to adapt to the weld joint tracking requirement of multiple welds types operating mode.

Description of drawings

Fig. 1 is the structured flowchart of laser weld seam tracking device of the present invention;

Fig. 2 is the inventive method control flow chart;

Fig. 3 is the structured flowchart of smart camera in the laser weld seam tracking device;

Fig. 4 is the functional flow diagram of smart camera in the laser weld seam tracking device;

Fig. 5 is motion control unit functional flow diagram in the laser weld seam tracking device;

Fig. 6 is the calibrating function schematic diagram of smart camera in the laser weld seam tracking device;

Fig. 7 is that the laser weld seam tracking device switch board is arranged schematic diagram;

Fig. 8 is embodiment of the invention structural representation;

Fig. 9～Figure 15 is treatment effect figure ()～(seven) that image of the present invention is processed each step;

Figure 16 follower and weld seam relative position schematic diagram.

The specific embodiment

Apparatus of the present invention structure comprises as shown in Figure 1:

IMAQ and processing unit, welding seam position information is processed and exported to the view data of collection laser weld characteristics of weld seam to this view data;

Motion control unit receives above-mentioned welding seam position information and carries out analyzing and processing, sends motion control instruction to kinematic axis executing agency;

Upper-layer parameters arranges and control module, and IMAQ and processing parameter, motion control parameter are set, and carries out communication and is connected with IMAQ and processing unit, motion control unit respectively.

Described IMAQ and processing unit comprise LASER Light Source, optical module and smart camera, and wherein a word linear laser of LASER Light Source generation reflects by welding work pieces, is gathered by smart camera through optical module.

Described motion control unit comprises motion controller, kinematic axis driver and kinematic axis executing agency, wherein:

Motion controller links to each other by the CAN bus with the smart camera of IMAQ and processing unit take DSP as control core, and the positional information of its input is analyzed, calculated, and obtains the control position information of kinematic axis executing agency;

The kinematic axis driver receives the control position information of motion controller, exports to kinematic axis executing agency, thereby finishes the tracking rectification compensating motion of the welding gun butt welded seam position that the present invention institute will finally realize.

As shown in Figure 8, the smart camera 3 in the system connects firmly with following the tracks of executing agency 10, and welding welding-gun 9 connects firmly with executing agency 10.IMAQ and processing unit 4 interior semiconductor one word linear laser device 1, optical filter 2 and the smart cameras 3 of having encapsulated.The 1st perforate 6 of outer cover top, be used for drawing the connection of power line, laser power analog control signal line and smart camera CAN Bus and the ethernet lan of laser instrument and smart camera, the 1st, 2 perforates 7,8 are as the going out of IMAQ and processing unit 4 air cooling systems, air intake.Single-piece frame and welding gun 9 are rigidly connected, mounting distance accurately guarantees by installing hole, be and integrated connection in executing agency 10, thereby executing agency 10 can realize that welding gun 9 is with respect to the adjustment of position while welding, in the present embodiment with what direction perpendicular to the bead direction motion? be defined as Y-direction.Executing agency 10 is fixed on the welding equipment by support 11 integral body.As shown in Figure 7, motion control unit and upper-layer parameters setting and control module integrated installation are in switch board, wherein the kinematic axis driver in the motion control unit is that servo-driver is installed in the switch board below, the industrial control computer of motion controller DSP2812 and employing Windows system is the top that upper-layer parameters setting and control module are installed on respectively switch board, the spare interface connection layout is as smart camera cable in the sampling and processing unit 4, outside PLC (realizing external control for linking to each other with other system) and system power supply etc., the switch board guidance panel is equipped with display and various indicator lamp, button and emergency stop switch.

The laser beam that semiconductor one word linear laser device 1 sends produces the structured light plane, shines the welding work pieces surface, forms the feature laser stripe of weld seam 5, laser focus is 110mm, the focal position live width is 0.044mm, and power is 50～100mw, and wavelength is 660nm.Optical filter 2 can be through the light wave of wavelength 660nm.Smart camera 3 gathers the laser stripe image that is formed by laser instrument 1.

Laser weld seam tracking system total working flow process may further comprise the steps as shown in Figure 2:

Beginning is by upper-layer parameters setting and control module parameters;

By IMAQ and processing unit Real-time Collection weld image, and carry out image algorithm and process and to obtain the weld image data;

With the weld image data buffer storage, the rower of going forward side by side is processed surely, determines the welding gun side-play amount that current collection image is corresponding;

Motion controller in the motion control unit carries out filtering, the match of pursuit path and calculates in real time the planning of trace point and movement instruction according to above-mentioned welding gun side-play amount;

Above-mentioned result of calculation is given the kinematic axis driver kinematic axis executing agency is implemented control.

At first, by upper-layer parameters setting and control module parameters, comprise 1) laser instrument light intensity regulating parameter, be transferred to slave computer DSP2812 by serial ports, by its simulation control interface the laser instrument light intensity is being controlled; 2) selection of the interface shapes such as V word interface, docking, overlap joint and etc. the selection of slab, unequal thickness plate welding condition, and the relevant parameter of the size of windowing, filtering scheduling algorithm is transferred to smart camera by the lan network interface; 3) set the filtering of tracking system pursuit path, fitting parameter, be transferred to the DSP2812 motion control unit by serial ports.

Then, by IMAQ and processing unit Real-time Collection weld image, and carry out image algorithm and process and to obtain the weld image data; With the weld image data buffer storage, the rower of going forward side by side is processed surely, determines the welding gun side-play amount that current collection image is corresponding, and detailed process is as follows:

IMAQ and processing unit Real-time Collection also carry out the weld image processing capacity and are finished by smart camera, the composition structure of smart camera as shown in Figure 3, the course of work as shown in Figure 4, now specifically describe as follows: the optical pickocff of camera carries out IMAQ according to certain beat, rhythm control, time for exposure and the size of windowing wait relevant parameter to be finished by the LAN LAN interface by host computer and preset, FPGA reads and gathers the original image that is stored in " frame buffer ", and carry out the hard-wired image processing function of FPGA, comprise and windowing, filtering, binaryzation, dilation and corrosion, edge extracting and axis of a weld extract, image processing effect is seen Fig. 9～15, wherein Fig. 9 is the original weld image with the sign of windowing, Figure 10 is the region of interest area image after windowing, Figure 11 is the image after filtering is processed, Figure 12 is the image after the binary conversion treatment, Figure 13 is the image after dilation and corrosion is processed, Figure 14 is the image after edge extracting is processed, the image after line drawing is processed centered by Figure 15.The axis of a weld view data of extracting stores in " image buffer storage ", magazine DSP takes out the axis of a weld view data and is stored in " synchronous DRAM " associated with it from " image buffer storage ", and it is carried out the position while welding feature point extraction process (as shown in figure 15), data are stored in " read-only storage ".Simultaneously, the DSP butt welded seam view data in the smart camera is demarcated processing, determines the side-play amount of the welding gun phase butt welded seam that current collection picture position is corresponding, sends the welding gun side-play amount to motion control unit by the CAN Bus interface at last.

Wherein, calibration principle figure as shown in Figure 6, process is as follows:

Optical plane uniformly-spaced is divided into rectangular mesh by certain density, and each grid vertex is the characteristic point of prior demarcation, has in as the plane and the corresponding grid of optical plane, because lens distortion, these grids become irregular quadrangle; By carrying out a series of accurate experiment measurings, can obtain the corresponding optical plane physical coordinates of image coordinate (u, the v) (x on irregular grid summit in the picture plane _w, y _w, z _w), obtain corresponding relation database table; Process the characteristics of weld seam dot information that obtains for above-mentioned image, be any to be calibrated some P (u in the picture plane, v), can go out its physical coordinates by the little quadrilateral mesh interpolation calculation of surrounding this point to be calibrated in the above-mentioned related table of having set up, obtain the three-dimensional coordinate P (x of actual welds point _w, y _w, z _w).

Before system starts working, need to carry out Zero calibration by accurate measurement, namely demarcate certain any physics corresponding relation on the dead-center position executing agency of regulation and picture plane.Because smart camera is fixed on executing agency and welding gun, therefore by this corresponding relation, can will correspond on the welding gun physical location as certain some position on the plane.

Then, the DSP2812 motion controller in the motion control unit carries out filtering, the match of pursuit path and calculates in real time the planning of trace point and movement instruction according to above-mentioned welding gun side-play amount;

The amount that kinematic axis executing agency implements control is the rectification side-play amount that current welding gun need to be followed the tracks of compensating motion, be that welding gun advances to the correct amount that next position is needing perpendicular to bead direction from a position, to obtain image constantly to be fixed in welding gun in the kinematic axis executing agency and the relative displacement of weld seam and processing result image embodies, be not the motion control correct amount of above-mentioned needs, so must carry out motion planning to the result of IMAQ and processing unit gained.

Before motion planning, in order to guarantee motion stabilization, need first the pursuit path data to be carried out filtering, match and calculated in real time trace point.

The DSP motion controller adopts the welding gun offset information of timer interrupt mode reading images acquisition and processing unit, and the timing beat is designed to 10 milliseconds;

Apparent distance scope is (in Fig. 8 before the DSP motion controller pursuit path filtration module (as shown in Figure 5), laser stripe 5 and the vertical range D of welding gun 9 between the projection on the workpiece) the welding gun offset information that obtains is as the filtering input message, filtering parameter (such as intermediate value or Mean Filtering Algorithm and 3 or 5 filtering parameters) by host computer (being upper-layer parameters setting and control module) transmission carries out the selection of algorithm and parameter, 2 times of forward sight distances are set as the length of filter window, finish filtering and calculate, output is through the welding gun side-play amount after the filtering.

The output quantity of pursuit path fitting module (as shown in Figure 5) is as input, fitting parameter by host computer transmission carries out the selection of approximating method (such as least square or Cubic Spline Fitting) and match window size (such as parameter setting or forward sight apart from the multiple setting), finish the Fitting Calculation, and the division that the line segment after the match is controlled the welding gun side-play amount according to the time beat synchronous with gathering beat is calculated in real time.

Tracking trajectory planning module (as shown in Figure 5) with the output quantity after the Fitting Calculation as input, calculate the rectification side-play amount that current welding gun need to be followed the tracks of compensating motion by leading deviation computational methods, and carry out movement instruction planning according to the control beat of 10 milliseconds of predetermined multiples.Leading deviation is calculated Method And Principle figure as shown in figure 16, and detailed process is as follows:

Because camera is positioned at the place ahead of welding gun, so the tracking lag of welding gun butt welded seam is in image processing process, the picture number that begins to process the real-time tracking to welding gun from camera collection the first width of cloth image, it is the picture number that gathers in the forward sight distance range, can adopt following formula to calculating, as shown in figure 16, be the deviation schematic diagram of weld seam and welding gun.

$M=\frac{L}{v\·t_p}---\left(1\right)$

In the formula: M is for processing picture number; L is distance between welding gun and the camera, i.e. the forward sight distance; V is speed of welding; t _pFor processing the frame weld image time.

The meaning of following formula: because welding gun just begins to carry out real-time tracking after will arriving the camera initial position, namely pass through the length L of forward sight distance, so at first calculate from the welding gun begin column and go to welding gun arrival camera initial position required time (L/v), process the used time t of a two field picture with this time divided by smart camera again _p, i.e. (L/ (vt _p)), obtain the processing picture number (M) before welding gun begins real-time tracking.

In actual applications, what the present invention was concerned about is that welding gun advances to the correct amount that next position needs in the horizontal direction from a position, and the processing result image embodiment is the relative position that obtains this moment welding gun of image and weld seam, be not the correct amount that the present invention needs, so must the result that image is processed gained be adjusted.The welding gun of the processing result image of position if " n " and the distance Dist of weld seam center _nExpression, from the position " n-1 " to the position horizontal direction correction value Delta of " n " _nExpression.Dist _nAnd Delta _nBe signed number, the Dist when the definition weld seam is positioned at the welding gun right side _nAnd Delta _nFor just, corresponding Dist when being positioned at the welding gun left side _nAnd Delta _nFor negative.

By primary condition as can be known, to advance to camera initial position (being made as a reset) front when welding gun, do not carry out the position adjustment of any horizontal direction, when arriving the camera initial position, welding gun is over against weld seam, and welding gun advances to the required horizontal level correction value (Delta of a set by the position reset ₁) calculate the deviate Dist of gained before equaling ₁No longer be the processing result image Dist that obtains and advance to position " 2 " horizontal direction correction value from the position set ₂, must correct this result, it is as follows to correct formula:

Delta ₂＝Dist ₂-Delta ₁ (2)

" 2 " correction value of " 3 " is welding gun to the position from the position:

Delta ₃＝Dist ₂-Delta ₁-Delta ₂ (3)

The rest may be inferred, from the position " n " correction value of " n+1 " is to the position:

${\mathrm{Delta}}_{n+1}={\mathrm{Dist}}_{n+1}-\underset{p=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(n+1\≤M,n,p\⊆Z)---\left(4\right)$

Top formula is applicable to the welding gun forward travel distance less than or equal to the situation of L.

When welding gun from the position " M " (position " M " expression welding gun begins first forward sight distance and position of beginning from initial position before the real-time tracking here) when advancing to position " M+1 ", situation changes.The camera weld image that " M+1 " locates to obtain in the position is corresponding to welding gun (being the camera center) translation Delta in the horizontal direction ₁Distance, namely this moment image horizontal reference be the abscissa of position set, rather than the abscissa of reset before this, so the welding gun correction value that " M " advances to position " M+1 " from the position is:

${\mathrm{Delta}}_{M+1}={\mathrm{Dist}}_{M+1}-\underset{p=2}{\overset{M}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(p\⊆Z)---\left(5\right)$

The processing result image of namely using position " M+1 " to locate deduct welding gun from the position set to the position " M " displacement sum in the horizontal direction be exactly from the position " M " to the position horizontal direction correction value of " M+1 ".

In like manner, the weld image that the position " M+2 " that obtains when welding gun in-position " 2 " is located, and obtain corresponding Dist by image processing algorithm _M+2, for obtain from the position " M+1 " to the position correction value Delta of " M+2 " _M+2, need to be from Dist _M+2Deduct from the position " 2 " to the position " M+1 " welding gun displacement sum in the horizontal direction, that is:

${\mathrm{Delta}}_{M+2}={\mathrm{Dist}}_{M+2}-\underset{p=3}{\overset{M+1}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(p\⊆Z)---\left(6\right)$

So the general correction value computing formula after welding gun in-position " M " is:

${\mathrm{Delta}}_{M+n+1}={\mathrm{Dist}}_{M+n+1}-\underset{p=n+2}{\overset{M+n}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(n,p\⊆Z)---\left(7\right)$

Above (4), (7) two formula be applicable to the various position relationships that welding gun and weld seam exist, be above-mentioned leading deviation computational methods.

At last, above-mentioned movement instruction is given the kinematic axis driver by the pwm signal of motion controller DSP2812 kinematic axis executing agency is implemented the motion of control realization weld joint tracking.

Claims (6)

1. laser weld seam tracking device is characterized in that comprising:

IMAQ and processing unit, the view data of collection laser weld characteristics of weld seam is carried out this view data

Process and the output welding seam position information;

Motion control unit receives above-mentioned welding seam position information and carries out analyzing and processing, sends motion control instruction to kinematic axis executing agency;

Upper-layer parameters arranges and control module, and IMAQ and processing parameter, motion control parameter are set, and carries out communication and is connected with IMAQ and processing unit, motion control unit respectively;

Described IMAQ and processing unit comprise LASER Light Source, optical module and smart camera, and wherein a word linear laser of LASER Light Source generation reflects by welding work pieces, is gathered by the smart camera with DSP and FPGA through optical module;

Described laser source wavelength is 660nm, and power is that 50mw～100mw is adjustable;

Optical filter in the described optical module can see through the light wave of wavelength 660nm;

Described motion control unit comprises motion controller, kinematic axis driver and kinematic axis executing agency, wherein:

Motion controller links to each other by the CAN bus with the smart camera of IMAQ and processing unit take DSP as control core, and the positional information of its input is analyzed, calculated, and obtains the control position information of kinematic axis executing agency;

The kinematic axis driver, the control position information of reception motion controller exports kinematic axis executing agency to, realizes that finally compensating motion is corrected in the tracking of welding gun butt welded seam position.

2. the control method of a laser weld seam tracking device is characterized in that may further comprise the steps:

Beginning is by upper-layer parameters setting and control module parameters;

By IMAQ and processing unit Real-time Collection weld image, and carry out image algorithm and process and to obtain the weld image data;

With the weld image data buffer storage, the rower of going forward side by side is processed surely, determines the welding gun side-play amount that current collection image is corresponding;

Motion controller in the motion control unit carries out filtering, match and the trajectory planning of pursuit path according to above-mentioned welding gun side-play amount;

Above-mentioned result of calculation is given the kinematic axis driver kinematic axis executing agency is implemented control;

Described image algorithm is processed by FPGA hardware programming and dsp software programming and is combined to realize, wherein realizes the image preprocessing process by the FPGA hardware programming, may further comprise the steps:

Window to gathering weld image, obtain area-of-interest;

The region of interest area image is carried out medium filtering obtain image after the filtering;

Image after the filtering is carried out the binary picture image intensifying to be processed;

Image after the binaryzation is implemented the morphology processing of dilation and erosion;

Image behind the dilation and erosion is carried out edge extracting;

Image after edge extracts carries out central line pick-up;

By dsp software programming realization character point leaching process, may further comprise the steps:

Adopt least square method to carry out fitting a straight line to processed the center line that obtains by FPGA;

By finding maximum corner position to find characteristic point, realization character point extracts.

3. by the control method of laser weld seam tracking device claimed in claim 2, it is characterized in that:

Described parameter comprises filtering, the match control algolithm parameter of laser instrument light intensity regulating parameter, workpiece parameter, image processing algorithm parameter and pursuit path computational process.

4. by the control method of laser weld seam tracking device claimed in claim 2, it is characterized in that: described definite welding gun side-play amount corresponding to current collection image may further comprise the steps:

Optical plane uniformly-spaced is divided into rectangular mesh by density, and each grid vertex is the characteristic point of prior demarcation;

In as the plane, have and the corresponding grid of optical plane, carry out the actual measurement of grid each point by Position Control, obtain the corresponding optical plane physical coordinates of image coordinate (u, the v) (x on irregular grid summit in the picture plane _w, y _w, z _w), obtain corresponding relation database table;

Process the characteristics of weld seam dot information that obtains for above-mentioned image, be any to be calibrated some P (u in the picture plane, v), go out its physical coordinates by the little quadrilateral mesh interpolation calculation of surrounding this point to be calibrated in the above-mentioned related table of having set up, obtain the three-dimensional coordinate P (x of actual welds point _w, y _w, z _w);

Determine to be fixed in the tracking displacement of the welding gun butt welded seam in the kinematic axis executing agency by three-dimensional coordinate (x, y, z) and camera and welding gun physical location relation.

5. by the control method of laser weld seam tracking device claimed in claim 3, it is characterized in that: the filtering of described pursuit path, match are also calculated trace point and movement instruction planning function in real time, and implementation procedure may further comprise the steps:

Motion controller adopts the welding gun offset information of timer interrupt mode reading images acquisition and processing unit;

The welding gun offset information that apparent distance scope obtains before the motion controller pursuit path filtration module is as the filtering input message, filtering parameter by the host computer transmission carries out the selection of algorithm and parameter, 2 times of forward sight distances are set as the length of filter window, finish filtering and calculate, output is through the welding gun side-play amount after the filtering;

The pursuit path fitting module with the output quantity of filtration module as input, fitting parameter by the host computer transmission carries out the selection of approximating method and match window size, finish the Fitting Calculation, and the division that the line segment after the match is controlled the welding gun side-play amount according to the time beat synchronous with gathering beat is calculated in real time;

The output quantity of tracking trajectory planning module after with the Fitting Calculation by leading deviation computational methods, calculated the rectification side-play amount that current welding gun need to be followed the tracks of compensating motion as input, and carries out movement instruction planning according to the control beat of 10 milliseconds of predetermined multiples.

6. by the control method of laser weld seam tracking device claimed in claim 5, it is characterized in that: described leading deviation computational methods are as follows:

When welding gun forward travel distance during less than or equal to L, welding gun from the position " n " correction value of " n+1 " is to the position:

$\mathrm{Delt}{a}_{n+1}={\mathrm{Dist}}_{n+1}-\underset{p=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(n+1\≤M,n,p\⊆Z);$ The welding gun forward travel distance is less than or equal to L; Wherein:

M is the forward sight distance L length welding gun beat number that advances, and namely welding gun begins the processing picture number before the real-time tracking;

Welding gun when n is the leading deviate of current calculating advances and controls the last beat number of beat;

P is a certain beat before the current location;

Dist _N+1The distance that departs from initial position for the current location welding gun;

Delta _pCorrection value for the relatively last beat of a certain beat horizontal level before the current location;

Delta _N+1Correction value for the relatively last beat of current location position;

The correction value computing formula is behind the welding gun in-position " M ":

${\mathrm{Delta}}_{M+n+1}={\mathrm{Dist}}_{M+n+1}-\underset{p=n+2}{\overset{M+n}{\mathrm{\Σ}}}{\mathrm{Delta}}_p(n.p\⊆Z);$

Wherein:

M, the p implication is the same;

N controls the last beat number of beat for beginning from initial position to advance apart from the current location welding gun that begins length after to calculate through first forward sight;

Dist _M+n+1The distance that departs from initial position for the current location welding gun;

Delta _pCorrection value for the relatively last beat of a certain beat horizontal level before the current location;

Delta _M+n+1Correction value for the relatively last beat of current location position.

Images