WO2004109268A1

WO2004109268A1 - Weld quality evaluation

Info

Publication number: WO2004109268A1
Application number: PCT/SE2004/000870
Authority: WO
Inventors: Pavel Lemarinier; Dick Skarin; Kent Mak; Ahmed Kaddani
Original assignee: Abb Ab; RUAULT-KADDANI, Pascale
Priority date: 2003-06-06
Filing date: 2004-06-04
Publication date: 2004-12-16
Also published as: SE0301674L; SE526624C2; SE0301674D0

Abstract

A method and a system for evaluation of the quality of a weld from an image of a cross section of a welded joint (1), comprising an evaluation unit (4) including a processor for carrying out instruction from a computer program and memory means (6) for storing images and computer programs.

Description

Weld quality evaluation

TECHNICAL FIELD

The present invention relates generally to a device and a method for evaluation of the quality of a weld. More precisely the invention concerns a method and a system for determine a quality parameter from examining a cross section of a welded joint. By welding in this context should mean to unite parts of metal or plastic either by heating and allowing the parts themselves or by submitting melted material to flow together, or by hammering or press together. More specifically the present invention relates to a device and a method for analyzing an image of a welded joint to evaluate the quality of the weld.

BACKGROUND OF THE INVENTION

A quality evaluation of a welded joint cannot in general be made by an ocular inspection of the appearance of the weld. To find out the degree of penetration of the molten material, the presence of cracks or cavities, the degree of filling of added molten material and other parameters the cross section of a welded joint has to be inspected.

Electric arc welding is a complicated process and the resulting deposition of molten metal into a weld pool for performing the welding operation is determined by a tremendous number of interrelated and non-interrelated parameters. These parameters affect the depo- sition rate, the spatter and debris around the welding operation, the shape and appearance of the weld bead, and the location and quality of the protective slag, to name just a few. The welding process is controlled by the protective gas composition, its flow rate, torch design, the welding torch angle, welding tip design, the size and shape of the deposition groove, control apparatus used in the welding process, amount of stick-out, wire feed speed, speed of the torch along the workpiece, smoke extraction, type of grounding contact on the workpiece, atmospheric conditions, the composition of the workpiece and other variables. Consequently, arc welding is largely a trial and error procedure with the ability of the welder to use the appropriate settings for obtaining consistent welds. Each time one of the parameters is changed, the appearance, size, shape, contour, chemistry and mechanical properties of the resulting weld is affected. For this reason, arc welding is a very complex science. Today trained welding engineers are required to provide the desired results. There is no procedure in the art which controls an arc welding process ad hoc without the intervention of the welder or welding engineer.

Arc welding systems comprising a controllable manipulator or an industrial robot are widely used in the industry. In such a process the robot is programmed to follow a desired path to be welded with the welding torch being held at a specified distance. Before the welding process starts the electric circuit, the movement of the robot and the arc welding process parameters must be tuned to have optimum quality and productivity.

Arc welding techniques include seam and projection welding, and the type of weld includes fillet welds, butt welds and others. The exemplary embodiments of the present invention disclosed are particularly useful for arc welds of the fillet weld type, but may find applications in other welding techniques and other manufacturing processes.

According to welding standards a plurality of parameters, such as size of legs, throat thickness, penetration and face shape must be determined in order to evaluate the quality of the weld. This determination is currently performed manually by measuring the parameters with a bare ruler and a naked eye. Such measurements is time consuming and highly dependent on the operator who is performing the measurement. The outcome of the qual- ity evaluation is therefore sometimes doubtful and thus has a poor value. Statistic treatment of quality evaluations from such measurements or sorting of different production tunings based on their quality evaluations is laborious and practically lacking reliable value.

Accordingly, there is a need for an improved system and method for correctly and uniformly determining the parameters for a quality evaluation of an arc weld. Further, there is a need for a system and method for evaluating the quality of a plurality of welds quickly and efficiently also in a manufacturing environment. SUMMARY OF THE INVENTION

The object of the invention is to provide a systematic quality evaluation of a weld, where the influence of a human operator is minimized. Thus the evaluation process is to be performed without the need of a human operator. The quality evaluation should be performed quickly, uniformly and in an accurate way. All parameters for evaluating the quality shall thus be determined automatically from an image of a cross section of a weld. Thus the pa- rameters needed for the evaluation shall be automatically measured from the image.

These objects are achieved according to a first aspect of the invention by a quality evaluating method according to the features in the characterizing part of the independent claim 1 , and according to a second aspect of the invention by a quality evaluation system according to the features in the characterizing part of the independent claim 5. Preferred embodiments are described in the dependent claims.

According to the first aspect of the invention the objects are achieved by a method for evaluating the quality of a weld from an image of a cross section of a welded joint. In a first step the shape and position of the jointed plates and the shape and position of the weld is determined. When concerning an arc welded joint the weld in this context should include the weld bead itself as well as the melted portion of the jointed plates. In a next step a template representing a predetermined shape of a desired weld is superimposed onto the weld. In a further step the areas of mismatch is calculated and used for the quality evaluation.

The contours of the jointed plates and their intersection point are determined from the image. The shape of the weld comprising the weld bead and the melted portions of the jointed plates is determined from the image by contrast comparison of a plurality of points on each side of the contour of the weld. Points representing the contour line are determined by its position in relation to a known point. Thus such a position is denoted by its distance and orientation from the intersection point or by its cartesian coordinates. In a next step the determined shape of the weld is compared with a template of a predeter- mined shape of a wanted weld. The determined weld bead is superimposed by the template. The weld quality is then evaluated by calculating the areas of mismatch between the shapes of the weld and the template. If the template shape is the greater than the weld it would indicate a lack of penetration. If the weld shape is the greater it would indicate an excess of weld material, or excessive penetration.

According to a second aspect of the invention the objects are achieved by an evaluation system comprising a camera and an evaluation unit including a processor and memory means for storing images and a computer program. The camera may be of the video type and configured to acquire a set of image data representing an image of the weld. The picture is transferred into the memory means by the processor. The computer program is embodied on a computer-readable medium and is configured to control the system in performing the quality evaluation method. Thus the system as controlled by the computer program receive the image data, identify the jointed plates data and the weld data from the image data, compare the weld data with a preferred template data by superimposing the weld data by the template, calculate along the template contour line the contour mismatch between the template and the weld, and by a scoring system evaluate the quality of the weld.

A weld is normally described by the length of its two legs from the intersection point

(root) along each of the jointed plates and the thickness of the throat. From the image the form of the jointed elements are determined in dimension, orientation and position. From this the position of the intersection point or root of the weld is determined. The surface and the dimension of the melted portion of the weld is then determined by comparing con- trasts of pixels in the image. In order to raise the contrast between the different areas of interest of the weld the system may comprise optical filtering means. Filtering effects may also be accomplished by image processing means performed by the processor.

In a preferred embodiment the shape of the weld is determined by a plurality of points around the weld shape. Firstly the root and the end positions of the legs are determined. Then the throat is determined by length and inclination from the root of points from each of a plurality of equally spread angles between the legs. Then the penetration of the weld along the first jointed plate is determined from positions along spaced apart lines normal to the first jointed plate. Equally the penetration along the second jointed plate is determined from spaced apart lines normal to the second jointed plate. In a preferred embodiment the determination of the weld is defined by the root and the two leg end positions, a plurality of points along the reinforcement, the first jointed plate and along the second jointed plate.

From an image of an arc weld there are mainly three areas of interest. That is the shape of the jointed plates, the shape of the weld bead and its melting properties into the jointed plates and the background. The distinction of these areas is made by contrast comparison of point on both sides of contour lines around these areas. To raise the contrast between the areas the image may be filtered. In resolving the contour line of an area, points along a line crossing the contour line are determined by position and by luminance. These points may advantageously be the pixels of the image. By comparing adjacent pixels along the crossing line in terms of luminance a contrast value between each pair of pixels are deter- mined. The position of the area contour line is then determined from the highest contrast value. In cases of uncertainty the luminance of each pixel is plotted in a diagram and the contour line is determined from the center of the steepest part of the curve.

The template is an image of an ideal weld image for the present welding operation. A plu- rality of such templates are stored in a memory means in the evaluation unit. In many cases the template must be scaled to correctly fit the image of the welded joint. This scaling is made automatically by the evaluation unit. However the scaling and orientation of the template is also made available to the operator.

The evaluation of the quality of the performed weld is accomplished by a score system. The areas of overlap and the areas of non-overlap of the template superimposing the weld are calculated by the evaluation unit. These areas are compared and evaluated by a set of rules, which are changeable by the operator. A maximum score will be achieved by the template completely matching the weld. A minimum score will result if there is a com- plete mismatch between the template and the weld.

According to one exemplary embodiment, a method of determining a measure of the quality of the arc weld based on a set of image data representing an image of the arc weld in- eludes identifying the image data into a first data set and a second data set, the first data set taken to be weld data and the second data set taken to be background data, and estimating a shape of the arc weld based on the weld data. The image data includes weld data and background data. The estimation provides a measure of the quality of the weld.

According to yet another exemplary embodiment, a computer program for analyzing a set of image data representing an image of a weld includes means for separating the weld data from the background data and means for calculating the surface brightness of the arc weld based on the weld data. The image data includes weld data and background data. The sur- face brightness indicates the quality of the weld.

BRIEF DESCRIPTION OF THE DRAWINGS

Fig. 1 is a block diagram of a system for analyzing a weld image to determine weld quality according to the invention,

Fig. 2 is an image of a performed weld of a welded joint where the contour of the weld is determined by a plurality of points,

Fig. 3 is an image of a template representing a predetermined weld shape, and

Fig. 4 is an image of a welded joint on which a system of computer related determination techniques is used in order to resolve the contours of the different areas is used.

Fig. 5 is a picture of an arc welded joint with definitions of parameters of interest for the quality evaluation

DETAILED DESCRIPTION OF EXEMPLARY EMBODIMENTS

A computer based system for evaluating the quality of a weld of a cross section of a welded joint 1 is shown in fig 1. The cross section may either be represented by the natural cross section or by an image, such as a bit map image or a jpeg picture, of the welded joint. The system comprises a camera 2, which may comprise a light source 3 for illuminating the image. The camera may conveniently comprise a video camera such as a DV camera. The camera is connected to a evaluation unit 4 including a processor or computer means. The evaluation unit comprises display means 5 for communication with an opera- tor and memory means 6 for storing templates, images and computer programs. The evaluation unit comprises also a communication link 7 for connection to a communication network including printer means, data bases and such.

According to the invention the weld profile is represented by 26 points as shown in fig 2. The image shows a first jointed plate 8, in the lower part of the fig, and a second jointed plate 9, directing upwardly in the fig, represented by a gray color. Further the image shows a weld 10 comprising the weld bead and its melted portions into the jointed plates represented by a white color. Finally the image shows the background portion 11 in black color.

Firstly the weld root 12 must be determined as the intersection point between the two jointed plates. Along the first jointed plate a first leg 13 is denoted from the root to the left most point 14 of the weld. A second leg 15 is denoted from the root along the second jointed plate to the up most point 16 of the weld in the fig. The reinforcement of the weld is defined by points on lines at 0, 15, 30, 45, 60, 75, and 90 degrees from the root between the two jointed plates. The melted portion of the first jointed plate is defined by eight points on perpendicular lines 17, of which six is equally spaced apart along the first leg. The melted portion of the second jointed plate is defined by eight points on perpendicular lines 18 equally spaced apart along the second leg.

Video camera 2 is configured to acquire a digital image of weld 1. The digital image including image data comprising weld data, jointed plates data and background data. A method of identifying the image data into a first data set, and a second data set will be described herein below, wherein the first data set is taken to be weld data and the second data set is taken to be jointed plates data. Video camera 2 may be any type of video or im- age capturing device, such as, a digital camera or digital video camera. In this exemplary embodiment, a digital black and white video camera is used. Alternatively, a color digital video camera may be used. The acquired digital image represents the weld 10 and portions of the first jointed plate 8 and the second jointed plate 9. The image also contains informa- tion of the background data 11. Optionally, a light source 3, such as a light bulb, image processing light, or other light source is provided in the vicinity of weld 1 to provide light in addition to the ambient light to the image or the cross section of the welded joint, so that a better image may be acquired by the video camera.

The video camera is configured to provide a digital image file or digital image data to the computer means in the evaluation unit for further processing. The computer means comprises any type of processor or signal processing device, such as, a microprocessor, a micro controller, a laptop computer, a personal computer, a personal digital assistant, etc. The evaluation unit includes the necessary memory, such as random access memory, readonly memory, programmable read-only memory, a hard drive, a floppy disk, some combination thereof, etc., for storing a software program suitable for operating the weld analysis steps. The memory means is also suitable for storing data, such as, the digital image file or digital image data received from the video camera. The computer further includes in- put/output means needed to communicate with the display and user input device. The display may be any type of display for displaying digital image data, charts, text and other data to a human operator, and may be a cathode ray tube (CRT), a liquid crystal display (LCD), a touch-pad interface, etc. The user input device may be a keyboard, a mouse, or other user input device. The computer is configured to process image data received from video camera and provide results via display or by saving the results to memory.

According to a preferred embodiment of the invention the video camera is configured to provide the digital image data via a network to computer for further processing. The network may be any type of communication medium, such as, a local area network (LAN), a wide area network (WAN), an internet protocol network, such as an intranet, the Internet, etc., or other networks. Thus, the video camera may be configured with a network card suitable for communicating over the network to the computer. Alternatively, the video camera may provide digital image data to a local computer, the local computer being configured with a network interface means suitable for communicating via the network with the computer in the evaluation unit. In an embodiment where a local computer is used, some or all of the processing steps may be performed on the local computer, with the remaining steps being performed by the computer in the evaluation unit. A template 12 representing a predetermined image shape of a desired weld is shown in fig 3. A plurality of such templates are stored in the memory means of the evaluation unit. Each template is scaleable by the operator to fit the weld of penetration.

An image of a weld to be examined for quality evaluation is shown in fig4. The image shows the principle of defining cross lines for the contour determination. The image also the core of the weld A as an equally legged triangle inscribed in the weld.

The method of quality evaluation is a score system. Firstly the image of evaluation is compared with the template of a predetermined weld image. If the image match the template completely there is a maximum score. If there is not a complete match there are a plurality of other parameters that have to come into consideration. In fig 5 a main part of these parameters are defined. The core A of the weld area has two equal legs and a throat size a. Along the first plate there is a first melted portion length zl and along the second plate there is second a melted portion length z2. Between the two plates there is a gap h4. There is a first side penetration hi into the first plate and a second side penetration h2 into the second plate. Between the plates there is a penetration depth h3. The weld reinforcement is denoted by its height bl and width b2. All these parameters are defined by the evaluating system controlled by the computer program and are involved in the scoring sys- tem for determine the quality score of the weld to be examined.

While the exemplary embodiments illustrated and described above are presently preferred, it should be understood that these embodiments are offered by way of example only. For example, other methods may be used to separate the weld data from the background data. Further, the methods disclosed herein may be used for other applications in which the brightness or size of pixel areas associated with a digital image are useful in detecting a characteristic of the system. Also, not all of the steps of the exemplary embodiments need be performed in all embodiments, nor need they be performed in the specific order recited. Accordingly, the present invention is not limited to a particular embodiment, but extends to various modifications that nevertheless fall within the scope of the appended claims. According to the invention the method and system described is also useable for determining the quality of joints other than arc welded joints. A gas welded joint is likewise an issue for such an evaluation as also are soldered joints.

Claims

1. A method for automatically evaluating the quality of a weld from an image of a cross section of a welded joint (1), which image comprises jointed plates data, weld data and background data, c h a r a c t e r i z e d i n , determining from the image the shape and position of the first (8) and second (9) jointed plates, determining from the image the shape and position of the weld (10) including the weld bead and the melted portion into the first and second jointed plate, superimposing the weld by a template (12) of a predetermined weld, calculating the areas of mismatch between the weld and the template.

2. A method according to claim 1, wherein the contour of a shape is determined by measuring the luminance of points inside and outside the shape, defining pairs of points that has the highest contrast and locating the contour line between those points.

3. A method according to claim 2, wherein the points to be measured are arranged equally spaced apart on a line crossing the border of the shape, plotting the luminance value of the measures points in a curve, defining the steepest part of the curve between a high luminance and a low luminance and locating the contour line on the steepest part half way between the high and low luminance.

4. A method according to claim 3, wherein the lines crossing the border of the shape are equally spread around the weld.

5. A method according to any of the preceding claims, wherein the acquiring of the image data is made by using a digital video camera (2).

6. A method according to any of the preceding claims, wherein the image data is provided across an internet protocol network.

6. A system for evaluation of the quality of a weld from an image of a cross section of a welded joint (1), comprising an evaluation unit (4) including a processor for carrying out instruction from a computer program and memory means (6) for storing images and computer programs, c h a r a c t e r i z e d i n that the evaluation unit comprises means for determining the shapes and positions of the jointed plates (8, 9) and the weld (10), a plu- rality of templates (12) of predetermined welds, means for superimposing a template onto the weld, and means for calculating the area difference between the weld and the template.

7. A system according to claim 6, wherein the evaluation unit comprises a video camera (2) configured to acquire a set of image data representing an image of the weld (1).

8. A computer program product for evaluating the quality of a weld presented as an image of a cross section of a welded joint, c h a r a c t e r i z e d i n that the computer program comprises instructions to affect a processor of a computer to perform the method of any of claim 1 - 5.

9. A computer program product according to claim 8, provided at least partly over a network (7) such as the Internet.

10. Computer readable medium containing a computer program product according to claim 8.