KR100684630B1 - Image processing method for tracking welding line - Google Patents

Abstract

An image processing method for tracking a welding line is provided to acquire a clearer welding line by removing a noise incapable of being removed through an existing binary image process. A laser light source is irradiated on a groove of a pipe formed as a stainless steel. A reflected beam is passed through a BPF(Band Pass Filter), and raw data are acquired through a gray CCD(Charge Coupled Device) camera(S100). A certain threshold value is set up in a gray level of the raw data, and a background and a laser stripe are separated to be converted into a binary image(S110). An average width of the laser stripe is calculated(S120). A center point of the laser stripe is extracted(S130). A welding line region of the average width or more is treated as a noise and is removed on the basis of the center point of the laser stripe(S140). An image in which the noise is removed is acquired(S150). Welding line information is extracted(S160).

Description

Image processing method for tracking welding line}

1 is a view showing a welding line tracking sensor for a pipe of a conventional liquefied natural gas carrier.

2 is a view showing a state of the data photographed the weld line of the pipe through a conventional laser vision system.

3 and 4 illustrate weld line tracking data obtained through a conventional image processing process.

5 is a view showing an embodiment of a laser vision system for welding seam tracking according to the present invention.

FIG. 6 is a side view showing one embodiment shown in FIG. 5; FIG.

7 is a flow chart illustrating an image processing process in accordance with the teachings of the present invention.

8 and 9 illustrate weld line tracking data obtained through an image processing process according to the inventive concept.

<Explanation of symbols for the main parts of the drawings>

100: camera 102: optical mirror

104: protective glass 110: welding torch

120: laser diode 122: laser stripe

130: welding wire 140: pipe

The present invention relates to an image processing method for welding seam tracking, and more particularly, to an image processing method for welding seam tracking that can obtain a sharper weld seam by removing noise that cannot be removed through conventional binarization image processing. .

Due to the advancement of industrial technology and the improvement of the working environment, many phenomena have been avoided. For this reason, a variety of automation equipments are emerging to allow robots to work on behalf of humans in harsh environments. In particular, in the shipbuilding industry, welding is dangerous in a position to be fused, and the environment is also poor, and more realization of automation is required. Therefore, in order to automate the hull assembly process of the ship, the function of automatically tracking the weld line must be provided, and by this function, the welding torch can be accurately positioned on the weld line to obtain good welding quality.

In general, weld line tracking sensors are mounted at the front end of the welding torch and are divided into contact and non-contact types. Among them, the laser vision sensor is widely used as a non-contact welding seam tracking sensor. The laser vision sensor is generally composed of a laser diode that emits a laser and a camera that detects the weld line by receiving the emitted laser reflected from the welded object and returning.

Hereinafter, the prior art will be described in more detail with reference to the drawings.

1 is a view showing a welding line tracking sensor for a pipe of a conventional liquefied natural gas carrier.

As shown in FIG. 1, in the conventional laser vision sensor, a welding wire 50 is injected from a front surface of the welding torch 20, and a predetermined forward direction is provided on a straight line between the welding torch 20 and the welding wire 50. Equipped with a laser diode 30 that emits laser light spaced apart by a distance of. A charge coupled device (CCD) camera 10 having a predetermined tilt is mounted in front of the laser diode 30. The laser diode 30 is scanned by the laser diode 30 in the direction of the welding line 60, is reflected by the welding line 60, and then photographed by the camera 10.

Liquefied natural gas pipes are made of stainless steel, so most of them use TIG welding. TIG welding is a tungsten inert gas welding, in which an arc is generated between a tungsten electrode having a highest melting point and a base metal, and the welding is protected by argon gas to prevent oxidation and nitriding during welding.

The welding machine according to the spirit of the present invention is a TIG welding machine using a welding wire injector, and the welding machine has a structural feature of injecting the welding wire 50 from the outside.

However, in the TIG welding using the welding wire injector, the separation between the welding torch and the laser vision system is large due to the interference of the welding wire, and thus, there is a problem in that it is impossible to trace the weld line in near real time and the possibility of welding failure increases.

FIG. 2 is a diagram illustrating a state of data photographing a weld line of a pipe through a conventional laser vision system.

Referring to FIG. 2, a convex protruding portion is a bead region in which welding is performed, and a horizontally parallel portion except for the bead region is a region in which welding is not performed. Therefore, the boundary between the bead area and the non-contact area must be clearly recognized so that accurate data on the welded portion can be obtained. However, the weld line data which has not undergone the image processing process is not clearly recognized due to the noise due to the noise, and thus the probability of welding defects is increased accordingly.

In more detail, FIG. 2 is a view illustrating an image obtained by irradiating a laser light source to a groove of a pipe made of stainless steel and reflecting the light through a band pass filter to a monochrome CCD camera. The black and white image is used to extract information about the groove shape and the weld line. When extracting the information about the weld line tracking using only the black and white image, external noise such as a welding arc and a diffused reflection of the laser strip can be sufficiently removed. none. Therefore, when the black and white image is used without any processing, wrong welding line tracking information is extracted, so image preprocessing is required.

3 and 4 illustrate weld line tracking data obtained through a conventional image processing process.

Referring to FIG. 3, a conventional image processing process for welding seam tracking may include setting a threshold value at a gray level and separating a background from a laser stripe. Although this method is used, since the pipe is made of stainless steel, it is not possible to prevent the diffuse reflection of the laser strip, which leads to the possibility of extracting false weld tracking information.

Referring to FIG. 4, when the center of the weld line illustrated in FIG. 3 is referred to as 0 and the left and right distances of the beads are represented by the Y axis, boundaries of the welded and non-welded regions appear at left and right 5 and -5 positions, respectively. Able to know. The X axis is the time axis in which the weld seam tracking process takes place. When the boundary points of welded areas and non-welded areas are clearly distinguished according to the change of time, sharper weld line tracking is possible. However, in the case of the data shown in FIG. 4, incorrect weld line information appears at the wrong point in addition to the boundary point of the actual bead. It can be seen. In order to more accurately weld seam tracking, there is a need to prepare a method for removing the wrong weld seam information.

The present invention has been made to solve the above problems of the prior art, and provides an image processing method for welding seam tracking that can obtain a clearer weld seam by removing noise that cannot be removed through conventional binarization image processing. For the purpose of

In order to achieve the above technical problem, the present invention uses a laser vision sensor for welding seam tracking weld line to remove the image noise due to the diffuse reflection of the laser light source and to remove inappropriate welding line information from the raw data through the image processing process Image processing method for tracking, which irradiates a laser light source to a groove of a stainless steel pipe, passes the reflected light through a band pass filter, and then uses raw data with a black and white CCD camera. Obtaining; Setting an original threshold value at a gray level and converting the original data into a binary image by separating a background and a laser stripe; Calculating an average width of the laser stripe; Extracting a center point of the laser strip; Treating and removing a weld line area of an average width or more based on the center point of the laser strip as noise; Obtaining an image from which the noise is removed; And extracting weld line information. The method provides an image processing method for welding line tracking.

The present invention has the effect of providing an image processing method for welding seam tracking that can obtain a clearer seam by removing noise that cannot be removed through conventional binarization image processing.

In order to fully understand the present invention, the advantages of the operability of the present invention, and the objects achieved by the practice of the present invention, reference should be made to the accompanying drawings which illustrate preferred embodiments of the present invention and the contents described in the accompanying drawings.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Like reference numerals in the drawings denote like elements.

5 is a view showing an embodiment of a laser vision system for welding seam tracking according to the present invention.

As shown in FIG. 5, two light generators 120 are positioned at both sides of the welding line 142 and spaced apart from the torch 110 in the welding direction. In this case, a laser diode is used as the light generator 120. The light output from the laser diode has its own divergence and thus is difficult to use directly. For this reason, collimated beams are made by using aspherical lenses and various lenses, and line beams are generated by using them. The two laser diodes 120 scan the laser light 122 into the pipe 140 to be welded. At this time, the two laser light 122 has a portion overlapping with respect to the welding wire 130 is injected along the welding line 142. This makes it possible to avoid the interference caused by the welding wire in a single light source. The camera 100 is spaced apart from the two laser diodes 120 in the welding direction and positioned in parallel with the torch 110. The optical mirror 102 is inclinedly mounted on the lower portion of the camera 100. In another embodiment, the camera may be mounted obliquely.

The separation angle refers to the angle between the laser light emitted from the laser diode 120 and the path through which the laser light is reflected by the pipe. As the separation angle is larger, the shape of the groove is more pronounced, thereby improving image processing of the camera 100. However, in order to increase the separation angle without the optical mirror 102, the installation angle of the camera 100 must be increased. However, this increases the size of the laser vision system, it is effective to use the optical mirror 102 for the vertical installation of the camera 100 and to change the incident angle.

The protective glass 104 is mounted on the front surface of the optical mirror 102. This serves to prevent the optical mirror 102 or the lens of the camera 100 from being damaged during welding, and uses a normal transparent glass through which the laser light can pass. You may use it.

FIG. 6 is a side view illustrating one embodiment shown in FIG. 5.

As shown in FIG. 6, the laser vision system based on the torch 110 has a form that includes another laser vision system symmetrically opposite.

In order to weld the circular liquefied natural gas pipe 140, the welding path of the welder does not consist of only one rotation in one direction. It is recommended to weld from bottom to top to improve the quality of the weld. This is because, in the case of welding while moving from the top to the bottom, the melted material at high temperature flows downward, and there is a high possibility of defects in the beads. For this reason, the welding is moved from the bottom to the top, because the round pipe 140 as described above does not end the welding by only one revolution. That is, in the circular pipe 140, it is divided into the lowest end and the highest end, and the welding is performed by moving the semicircle from the lowest end to the highest end in the clockwise direction first, and then moving back in the counterclockwise direction to return to the lowest end point. Then, we weld with the semicircle moving from the lowest end to the highest end in the counterclockwise direction, and then moving back clockwise to return to the lowest end. As another embodiment, welding may be performed in the reverse order to the above.

In order to have the welding path as described above, two laser vision systems are mounted in both directions of the torch 110 so that the welding line can be sensed even when moving in any direction in both directions.

7 is a flowchart illustrating an image processing process according to the spirit of the present invention.

Referring to FIG. 7, first, a laser light source is irradiated to a groove of a pipe made of stainless steel, and the reflected light is passed through a band pass filter, and then raw data is transferred to a black and white CCD camera. Obtain (S100). Thereafter, the raw data is set to a threshold value at a gray level, and the background and the laser stripe are separated and converted into a binary image (S110). Next, the average width of the laser stripe is calculated (S120), and the center point of the laser stripe is extracted (S130). Next, the weld line region having an average width or more based on the center point of the laser strip is removed as noise (S140). Thereafter, the image from which the noise is removed is obtained (S150), and the weld line information is extracted (S160).

8 and 9 illustrate weld line tracking data obtained through an image processing process according to the inventive concept.

Referring to FIG. 8, as described with reference to FIG. 7, weld line data having a distinct boundary between a welded area and a non-welded area as compared with the conventional weld line tracking data shown in FIG. 3 through an image processing process for weld line tracking according to the spirit of the present invention. Can be obtained. In other words, the noise caused by the diffuse reflection of the laser strips can be largely removed and information about the weld line can be extracted with a clearer image.

Referring to FIG. 9, when the center of the weld line illustrated in FIG. 8 is referred to as 0 and the left and right distances of the beads are represented by the Y axis, boundaries of the welded and non-welded regions appear at the left and right 5 and -5 positions, respectively. Able to know. The X axis is the time axis in which the weld seam tracking process takes place. It can be seen that the data shown in FIG. 9 clearly distinguishes the boundary point between the welded area and the non-welded area according to the change of time compared to the data shown in FIG. This means that a sharper weld seam tracking is possible when using the welding data shown in FIG. 9 than using the data shown in FIG.

Although the present invention has been described with reference to one embodiment shown in the drawings, this is merely exemplary, and those skilled in the art will understand that various modifications and equivalent other embodiments are possible therefrom. . Therefore, the true technical protection scope of the present invention will be defined by the technical spirit of the appended claims.

The present invention has the effect of providing an image processing method for welding seam tracking that can obtain a clearer seam by removing noise that cannot be removed through conventional binarization image processing.

Claims (1)

An image processing method for welding seam tracking that removes image noise due to diffuse reflection of a laser light source and removes inappropriate welding seam information from raw data by using a laser vision sensor for welding seam tracking. Irradiating a laser light source to a groove of a pipe made of stainless steel, passing the reflected light through a band pass filter, and obtaining raw data with a monochrome CCD camera; Setting an original threshold value at a gray level and converting the original data into a binary image by separating a background and a laser stripe; Calculating an average width of the laser stripe; Extracting a center point of the laser strip; Treating and removing a weld line area of an average width or more based on the center point of the laser strip as noise; Obtaining an image from which the noise is removed; And Extracting weld line information.

Images