JPH0120957B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Technical Field of the Invention] The present invention relates to an apparatus for automatically optically detecting the cross-sectional shape of a groove shoulder of a base material to be welded.

[Prior art]

A conventional device of this type is based on the detection principle shown in FIG. In the figure, 1 is an illumination device such as a laser, 2 is the base material to be welded, 3 is a light cutting line, 4 is an imaging device, Gw is the groove width direction, Gl is the groove line direction, and Ga is the left opening. Tip shoulder, Gb is right bevel shoulder,
L indicates the detected position of the left shoulder of the groove, R indicates the detected position of the right shoulder of the groove, and C indicates the position of the center of the groove. Next, an outline of the detection principle shown in FIG. 1 will be explained for each step.

Step 1: Using the illumination device 1, a linear beam (slit light) is irradiated onto the groove portion of the base material 2 to be welded from an appropriate angle so as to be perpendicular to the welding line.

Step 2: Using an imaging device 4 such as an ITV camera, the shape of the optical cutting line 3 developed in the groove portion, that is, the cross-sectional shape of the groove is detected.

Step 3: Based on the detected image information, detect the positions L and R of the left and right groove shoulders Ga and Gb (the intersection of the base material surface and the groove slope), and find the center position between the left and right groove shoulders. Calculate C((L+R)/2).

A device applying such a detection principle is proposed in Japanese Patent Laid-Open No. 55-20631, and a relatively similar device is proposed in Japanese Patent Laid-Open No. 52-133050.

However, the "welding groove" that is the object of the above-mentioned device refers to a shape that is generally referred to as a "V-shaped groove" or "V-shaped groove."
Therefore, the application of the above device is limited to only base materials to be welded where the left and right groove shoulders and the groove slopes in the vicinity thereof can be clearly observed.

However, as a practical matter, the manufacturing process of welded structures in the metal processing industry includes processes such as ``pressing'' and ``tack welding'' as pre-processes to ``main welding.'' The base metal to be welded may have a groove cross-sectional shape in which the groove shoulder was crushed during press working, or the groove shoulder may be covered and hidden by the weld bead during tack welding. Many of them have an ivy groove cross-sectional shape.
In addition, in the multi-layer welding method that is commonly used for main welding, weld beads are layered on the inner surface of the groove, so depending on the thickness of the weld bead, the shoulder of the groove may become a weld bead. Things that are covered up may occur during the process of reaching the final layer.

FIG. 2 shows an example of the cross-sectional shape of the groove of the base material to be welded during main welding. In the same figure, a is the weld bead WB
b is an example of a base metal to be welded in which the right groove shoulder Gb is hidden by the weld bead WB. , and c is an example of a base material to be welded in which the groove shoulder Gb on the right side has been crushed in the pre-welding process. In addition, in the figure, WBe is bead edge.

Therefore, during actual welding and at the finishing stage of actual welding, it is difficult to say that the groove shoulder of the base metal to be welded can be clearly observed, and the conventional technology has shortcomings in the detection principle itself. It was not possible to accurately detect the cross-sectional shape of the tip shoulder and the groove center position.

[Purpose of the invention]

The present invention solves the problems of the prior art described above, and provides an apparatus for detecting various groove cross-sectional shapes in which the groove shoulder is obscured or crushed. It is.

[Summary of the invention]

A cross-sectional shape detection device for a welding groove shoulder according to the present invention includes an illumination means, a photoelectric conversion means, a means for detecting an optical cutting position signal, a macro second-order difference circuit, a micro second-order difference circuit, and a rough shape of the groove shoulder. The apparatus includes a detection means, a groove shoulder detailed shape detection means, and a cross-sectional shape detection means.

The illumination means irradiates the welding groove with linear light, and the photoelectric conversion means captures a light sectioned image obtained by the illumination means to obtain an electrical signal corresponding to the reflected brightness. After the optical cutting position signal is detected based on the electrical signal, the macro second-order difference circuit performs a second-order difference calculation on the optical cutting position signal with a wide difference width. On the other hand, the micro second-order difference circuit uses a narrow difference width only in a predetermined range based on the measurement range data obtained from the macro second-order difference circuit.
Perform floor differences. The groove shoulder rough shape detection means is
The approximate shape of the groove shoulder is classified based on the positive or negative sign of the output waveform of the macro two-stage differential circuit, and the approximate position of the groove shoulder is detected. On the other hand, the groove shoulder detailed shape detection means classifies the detailed shape of the groove shoulder based on the positive or negative sign of the output waveform of the micro two-stage differential circuit, and detects the detailed position of the groove shoulder. The cross-sectional shape detection means determines the shape of the groove shoulder based on the combination of the classified rough shape of the groove shoulder and the detailed shape of the groove shoulder, and also determines the shape of the groove shoulder determined. The position of the groove shoulder is detected by selecting either the general position or the detailed position based on the shape.

〔Example〕

Next, embodiments of the present invention will be described based on the drawings.

FIGS. 3 and 4 are structural explanatory diagrams showing a part of an apparatus according to an embodiment of the present invention. In FIG. 3, 10 is an illumination/imaging device, 11 is a signal processing device, 12 is an image guide (optical fiber), 1
3 is a light guide (optical fiber), 14 is a light source,
15 is Tatsuchi Roll, 16 is a welding position control device,
And 17 is a welding torch. In FIG. 4, in the illumination/imaging device 10, 101 is a light projecting device, 102 is a light receiving device, 103 is an imaging device, and 10
4 is a lighting/imaging device housing case, and 105 is a disturbance light shielding material.

The laser light emitted from the light source 14 is transmitted to the floodlight 1 of the illumination imaging device 10 via the light guide 13.
01 to generate a linear light beam (slit light), which is irradiated onto the groove of the base material 2 to be welded (orthogonal to the welding line, with an appropriate irradiation angle) to illuminate the groove surface. The developed light section image is captured by the imaging device 103 using the light receiving device 102 and the image guide 12.
The resulting video signal is processed by a signal processing device 11 to detect the cross-sectional shape of the shoulder of the welding groove and the center position of the groove.

The welding position control device 16 performs positioning control so that the welding torch 17 always points to the welding groove center position based on the detected welding groove center position, and also performs positioning control of the illumination/imaging device 10. It is. Lighting/imaging device storage housing 104
is controlled by a welding position control device 16 using a sensor such as the Tatsuchiro 15 so that the imaging distance l ₁ and the illumination distance l ₂ are always constant. In addition, disturbance light such as welding arc is shielded by the disturbance light shielding material 105, so that a stable light section image can always be obtained.

The above is an example of the device configuration of this embodiment.
The signal processing device 11 will be explained based on the diagram.

This signal processing device 11 is roughly divided into an optical cutting position detection section 111 and a groove shoulder cross-sectional shape detection section 11.
5 and a groove center position calculation circuit 120.

The optical cutting position detection section 111 includes a video signal noise removal circuit 112, an optical cutting position detection circuit 113, and a disconnection/abnormal shape data correction circuit 114. A video signal (light-cut image) from the imaging device 103 is sent to a video signal noise removal circuit 112.
Moving average processing is performed only in the weld line direction by
Noise components included in the light-cut image are removed.
The output signal of the video signal noise removal circuit 112 is
The optical cutting position detection circuit 113 repeatedly detects the brightest point (position showing the maximum video signal level) in the welding line direction sequentially in the groove width direction to obtain optical cutting position information as shown in FIG. 6A. At this time, the optical cutting position information detected by the optical cutting position detection circuit 113 may include information about disconnection or abnormal shape of the optical cutting line due to adverse effects such as the surface properties of the base material to be welded. In such a case, the wire breakage/abnormal shape data correction circuit 114 performs interpolation processing using normal data on the left and right sides of the abnormal portion to create light cut position information as shown in FIG. 6B.

The groove shoulder cross-sectional shape detection unit 115 includes a moving average circuit 116 and a groove rough cross-sectional shape detection circuit 11.
7. Consists of a groove detailed cross-sectional shape detection circuit 118 and a groove shoulder cross-sectional shape detection circuit 120, which detects the left and right grooves based on the optical cutting position information detected by the optical cutting position detection section 111. The cross-sectional shape and position information of the shoulder, that is, the groove cross-sectional shape is detected.

Here, prior to a detailed explanation of the groove cross-sectional shape detection section 115, the cross-sectional shape of the groove shoulder will be explained.

The cross-sectional shape of the groove shoulder can be roughly classified into the following three types.

[1] Groove shoulder shape: Refers to the cross-sectional shape in which the intersection between the welding base material surface and the groove slope can be observed, as shown in Figure 2a.For convenience, the intersection between the welding base material surface and the groove slope is referred to as It is called a "grooved shoulder."

[2] Sagging shoulder shape: As shown in Figure 2c, this refers to the cross-sectional shape in which the groove shoulder has been crushed in a pre-welding process such as press working, and for convenience,
A crushed shoulder is called a "sagging shoulder."

[3] Bead shape: Refers to the cross-sectional shape in which the groove shoulder is covered and hidden by the weld bead, as shown in Figure 2b.For convenience, the intersection of the weld base metal surface and the weld bead is referred to as the bead edge. It is called.

The actual groove cross-sectional shape of the base material to be welded is a combination of the above three types of shapes at the left and right groove shoulders.

Further, the detection accuracy of the welding groove center position depends on the cross-sectional shape of the groove shoulder. That is, it can be evaluated based on the information used to detect the welding groove center position, and its reliability can be evaluated based on the following: [1] Groove shoulder position information > [2] Sagging shoulder position information > [3] Bead edge position Information is the highest.

Now, returning to FIG. 5 again, details of the groove shoulder cross-sectional shape detection section 115 will be described.

A moving average circuit 116 removes noise components from the light cutting position information from the light cutting position detecting section 111 through moving average processing. Here, the moving average width is set within a range that does not impair the position detection accuracy of the groove shoulder. The output signal of this moving average circuit 116 is as shown in FIG. 6C.

A groove rough cross-sectional shape detection circuit 117 and a groove detailed cross-sectional shape detection circuit 118, to which the output signal of the moving average circuit 116 is supplied, both detect the optical cutting position information that has been subjected to the moving average processing by By performing differential processing, the cross-sectional shape and position of the left and right groove shoulders are detected.

The principle of this detection is as follows. This is an application of the fact that large negative or positive waveforms appear at the groove shoulder position and bead edge position through second-order difference processing.By determining the negative or positive waveform pattern, the cross section of the groove shoulder The shape can be determined, and position information can be obtained by finding the peak position. here,
The parameter (difference width) when performing second-order difference processing has a very important meaning. In other words, when a wide differential width is used as a parameter, it has the advantage of high detection ability for macroscopic shapes such as bead shapes and good S/N ratio, but on the other hand, it has the advantage of having a good signal-to-noise ratio (S/N). The disadvantage is that the detection ability of so-called micro-shapes, which are only slightly observed, is low, and the position detection accuracy is poor. On the other hand, when a narrow difference width is used as a parameter, it has the advantage of high micro-shape detection ability and good position detection accuracy, but it has the disadvantage of low macro-shape detection ability and poor S/N. There is.

The present invention focuses on the characteristic that contradictory advantages and disadvantages as described above appear depending on the difference width when performing second-order difference processing, and uses only the mutual advantages to compensate for each other's disadvantages. To achieve this goal, two series of second-order differential circuits were provided. These are a groove portion rough cross-sectional shape detection circuit 117 and a groove portion detailed cross-sectional shape detection circuit 118.

First, the groove rough cross-sectional shape detection circuit 117 performs second-order difference processing using the wide difference width as a parameter to obtain groove rough cross-sectional shape information as shown in FIG. 6D. The large plus or minus waveform that appears in this information suggests the cross-sectional shape of the left and right groove shoulders, and the position information in the groove width direction that forms this waveform should be used to detect that range in more detail. This shows that highly accurate cross-sectional shape detection is possible. Here, the difference width is set to a width that can sufficiently detect a shape that shows a gradual change, such as a sagging shoulder shape or a bead shape.

The groove detailed cross-sectional shape detection circuit 118 performs second-order difference processing using the narrow difference width as a parameter. At this time, the groove width forming a large waveform obtained by the groove rough cross-sectional shape detection circuit 117 described above is By limiting the processing range of the second-order difference based on the positional information in the direction, the S/N ratio is improved and detailed cross-sectional shape information of the groove portion as shown in FIG. 6E is obtained. here,
The difference width is set within a range that can sufficiently detect a shape in which only a slight open shoulder is observed, as shown in FIG. 2a, and does not impair position detection accuracy.

The groove shoulder cross-sectional shape detection circuit 119 determines the waveform pattern appearing in the information obtained by the groove detailed cross-sectional shape detection circuit 118 and determines that the shapes of the left and right groove shoulders are "groove shoulders." It identifies whether it is a ``sagging shoulder'' or a ``bead edge'' and detects the position. However, for shapes that show gradual changes that cannot be detected by the groove detailed cross-sectional shape detection circuit 118 (sag shoulder shape and bead shape),
By supplementing with the detection results of the groove rough cross-sectional shape detection circuit 117, the cross-sectional shapes and positions of the groove shoulders having various cross-sectional shapes are detected.

FIGS. 7A and 7B are explanatory diagrams showing the relationship between the waveform pattern obtained by second-order difference processing and the cross-sectional shape of the groove shoulder.

The groove rough cross-sectional shape detection circuit 117 determines whether the rough shape waveform (second-order difference waveform with a wide difference width as a parameter) has a single peak in the positive direction or a single peak in the negative direction. The rough shapes are classified into three types depending on whether a peak in the positive direction and a peak in the negative direction appear together. The groove detailed cross-sectional shape detection circuit 118 determines whether the detailed shape waveform (second-order difference waveform with a narrow difference width as a parameter) has a single peak in the positive direction or a single peak in the negative direction. Detailed shapes are classified into three types depending on whether they appear or not.
The groove cross-sectional shape detection circuit 119 detects three types of general shapes and three types of detailed shapes as shown in FIGS. 7A and 7.
A matrix as shown in Figure B is created to detect whether the shape of the left and right groove shoulders is a groove shoulder, a bead edge, or a sag shoulder, and to determine the position of the shoulder according to the shape. B1,A
1, B2, and A2 are detected. The cross-sectional shape and position detection information of the groove shoulder are supplied to the next groove center position calculation circuit detection section 120, and are also used for other information processing as appropriate.

The groove center position calculation circuit 120 calculates the groove center position based on the detection result of the groove shoulder cross-sectional shape detection circuit 119 described above. The calculation method is as follows: First, the detected position of the left and right groove shoulders is the “groove shoulder position”.
If the left and right groove shoulders have the same cross-sectional shape, the center of the left and right detection positions is set as the groove center position, and the left and right groove shoulders are opened. If the cross-sectional shape of the tip shoulder is different, the detection position of the cross-sectional shape with high reliability ([1]
The groove center position is calculated using groove shoulder position>[2] sag shoulder position>bead edge position) and the groove width standard value inputted in advance. Figure 8 shows the calculation method for the weld groove center position. In Fig. 8, Δd is the standard groove width, and the numbers 1, 2, and 2 indicate reliability.
The smaller 3 is, the higher it is. Also, A ₁ in the same figure is the 7th A
A 2 corresponds to A ₁ or B ₁ in Figure 7B, and A ₂ corresponds to A ₂ or B 1 in Figure 7B.
Corresponds to B ₂ .

For example, as shown in FIG. 9, the macro detection signal on the right shoulder (output of 117) is a positive sine wave signal,
When there is no micro detection signal (output of 118), it is recognized in FIG. 7 that the bead shape has a gradual rise, and here, the address
Let A ₁ = 144. In addition, if the macro detection signal of the left shoulder is a positive and negative sine wave signal and the micro detection signal is a negative sine wave signal, it is recognized that the groove shoulder shape is clearly observable in FIG. Assume that address A ₂ =102 and Δd=46. Therefore, the groove center position can be determined from FIG. 8 as: groove center position=A ₂ +Δd/2=102+23=125.

This groove center position information is stored in the false detection prevention circuit 12.
1. This circuit 121 has a function of storing the groove center position information of the same welding base material, and compares the newly detected weld groove center position with the position information detected previously or previously. Then, it is determined whether the newly detected welding groove center position is appropriate or not, and appropriate position information is output to the welding position control device 16.

〔Effect of the invention〕

As is clear from the above description, the device according to the present invention is capable of detecting the cross-sectional shape and position of the welding groove shoulder without being restricted by the groove cross-sectional shape of the base material to be welded. .

In addition, in the above-mentioned example, the following matters supporting the usefulness of the present invention have been confirmed.

(i) Groove shoulders can be detected by micro second-order difference when the groove shoulder width (see Figure 10a) W0.5 mm or more.

(ii) For the bit edge part, sufficient detection ability cannot be obtained with micro second-order difference, but if macro second-order difference is used, at least the bead inclination (bead height/bead width/2=b/a, 10th (See figure b)
0.2 or more can be detected.

[Brief explanation of drawings]

Fig. 1 is an explanatory diagram of the conventional welding groove center position detection principle, Fig. 2 a, b, and c are explanatory diagrams of the groove cross-sectional shape of the base material to be welded, and Fig. 3 is an illustration of an embodiment of the present invention. 4 is an explanatory diagram of the structure of such a device, FIG. 4 is an enlarged cross-sectional diagram of the illumination imaging device of FIG. 3, FIG. 5 is a block diagram of the signal processing device of FIG. 3, and FIGS. Output signal waveform diagrams of each part in the figure, Figures 7A and 7B are explanatory diagrams of the waveform pattern by second-order difference processing and the cross-sectional shape of the groove shoulder, and Figure 8 is the cross-sectional shape of the groove shoulder and the groove. Fig. 9 is an explanatory diagram of the calculation of the center position, and Fig. 9 is a flowchart of an example of calculation of the groove detection position, Fig. 10.
Figures a and b are explanatory diagrams of the groove cross-sectional shape of the base material to be welded. 10... Illumination/imaging device, 11... Signal processing circuit,
12...Image guide, 13...Light guide, 1
4...Light source, 15...Tatsuchiroll, 16...Welding position control device, 17...Welding torch. 101... Light projecting device, 102... Light receiving device, 103... Imaging device. 11
DESCRIPTION OF SYMBOLS 1... Optical cutting position detection unit, 115... Groove shoulder cross-sectional shape detection means, 117... Groove rough cross-sectional shape detection circuit, 118... Groove detailed cross-sectional shape detection circuit, 11
9... Groove shoulder cross-sectional shape detection circuit, 120... Groove center position calculation circuit.

Claims (1)

[Scope of Claims] 1. Illumination means for irradiating linear light onto the welding groove; Photoelectric conversion means for capturing a light sectioned image obtained by the illumination means and obtaining an electrical signal corresponding to reflected brightness. means for detecting an optical cutting position signal using an electrical signal obtained by the photoelectric conversion means; a macro second-order difference circuit that performs a two-stage difference calculation with a wide difference width on the detected optical cutting position signal; ; a micro second-order difference circuit that performs second-order difference calculations with a narrow difference width on the optical cutting position signal; and a macro second-order difference circuit that classifies the approximate shape of the groove shoulder based on the sign or negative of the output waveform of the macro second-order difference circuit. , a groove shoulder rough shape detection means for detecting the rough position of the groove shoulder; classifying the detailed shape of the groove shoulder based on the positive/negative of the output waveform of the micro second-order difference circuit; a groove shoulder detailed shape detection means for detecting the position of the groove shoulder; determining the shape of the groove shoulder based on a combination of the classified rough shape of the groove shoulder and the detailed shape of the groove shoulder; , cross-sectional shape detection means for detecting the position of the groove shoulder by selecting either the approximate position or the detailed position based on the determined shape of the groove shoulder; A cross-sectional shape detection device for the groove shoulder.