CN104776799A - Cosmetic welding front seam detection device and method adopting lateral light to construct light and shadow characteristics - Google Patents

Abstract

A device and method for detecting a weld seam before welding on a cover surface using side light to construct light and shadow features, belonging to the field of welding automation. The invention uses a surface light source to irradiate the edge of the weld, and the imaging element is placed near the optical axis of the specular reflection on the side wall of the groove, and avoids the optical axis of the specular reflection on the base material and the surface of the weld, and collects the grayscale image of the weld surface, which is processed and controlled After image processing by the unit, the seam trajectory can be accurately extracted. The lateral illumination condition constructed by the invention can effectively highlight the feature information of the edge of the weld, and the gray level of the edge of the weld is close to saturation, which is convenient for quickly and accurately extracting the position of the weld trajectory; the detection accuracy can reach 0.04mm, the system structure is simple, and the cost is low. It has good real-time performance and is suitable for automatic detection of weld trajectory before cover welding where the height difference between the weld seam and the base metal surface is as low as 1mm or even lower.

Description

Apparatus and method for pre-weld weld detection on cap surface using side light to construct light and shadow features

technical field

The invention belongs to the field of welding automation, and in particular relates to a welding seam detection device and method for cover surface before welding by using side light to construct light and shadow features.

Background technique

On-line detection of weld trajectory before cap welding is of great significance in applications such as real-time tracking of the welding process. At present, the structured light method is widely used in the field of weld detection. However, in the case of weld trajectory detection before cover welding, especially when the height difference between the weld and the base metal surface is as low as 1 mm or even lower, the geometry of the weld surface The structural features become very weak, and the lack of distortion features of the structured light strip will make it impossible to effectively and accurately identify the weld trajectory. At present, there is no welding seam trajectory detection device and method applicable to the above occasions.

Contents of the invention

The purpose of the present invention is to address the deficiencies of the prior art, and to propose a device and method for detecting weld seams on the cover surface before welding using side light to construct light and shadow features. Geometric structural features, detection accuracy and limited applicability, etc., in order to realize automatic identification of weld trajectory before cap welding, especially for pre-weld welds with a height difference between the weld seam and the base metal surface as low as 1mm or even lower Track automatic detection occasions.

Technical scheme of the present invention is as follows:

A pre-weld inspection device for cover surface welding using side light to construct light and shadow features, characterized in that it includes a surface light source array, an imaging element and a processing control unit; the surface light source array includes at least two surface light sources, and the surface light source array includes at least two surface light sources. The light sources are respectively placed on both sides of the weld to be detected, and the light emitted by the surface light source on the same side is projected on the edge of the weld to be detected and the base material on the other side; the surface light source, the weld to be detected and the imaging element The relative position of is satisfied: the imaging element collects the specular reflection light of the main axis ray of each surface light source passing through the side wall of the groove, and the imaging element does not collect the specular reflection light of the main axis ray of each surface light source passing through the base material and the surface of the weld to be detected The surface light source array is connected to the processing control unit through wires; the imaging element is connected to the processing control unit through wires, or communicates through wireless transmission; the processing control unit processes the image collected by the imaging element;

The welding seam detection device before cover welding according to the present invention is characterized in that: the imaging element is a charge coupled device, a complementary metal oxide semiconductor imaging device, a position sensitive device or a charge injection device; the sensitive wavelength of the imaging element is The range is greater than or equal to the emission wavelength range of the surface light source;

The present invention provides a method for detecting a weld seam on a cover surface before welding using side light to construct light and shadow features, which is characterized in that the method includes the following steps:

1) Turn on all surface light sources located on both sides of the weld to be detected at the same time, or alternately light up surface light sources on different sides of the weld to be detected;

2) The light emitted by the surface light source is projected on the edge of the weld to be detected, the processing control unit controls the imaging element to collect images, and the imaging element transmits the collected weld surface grayscale image to the processing control unit unit, the processing control unit preprocesses the grayscale image of the weld surface:

a) If all the surface light sources on both sides of the weld to be detected are turned on at the same time, the processing control unit controls the imaging element to collect a grayscale image of the surface of the weld when the surface light source array is turned on, denoted as I(x, y), where x, y are row coordinates and column coordinates of the weld surface grayscale image respectively, and the imaging element transmits the weld surface grayscale image to the processing control unit; the processing control unit uses a threshold segmentation method , convert the grayscale image I(x,y) of the weld surface into a binary image B(x,y); the processing control unit extracts the two connected elements with the largest area in the binary image B(x,y) domain, denoted as R ₁ and R ₂ ;

b) If the surface light sources on different sides of the weld to be detected are turned on alternately, the processing control unit sends a trigger signal, so that the imaging element collects the grayscale images of the weld surface when the surface light sources on different sides are turned on, respectively denoted as I ₁ (x, y) and I ₂ (x, y), where x, y are the row coordinates and column coordinates of the weld surface grayscale image respectively, and the imaging element transmits the weld surface grayscale image to The processing control unit; the processing control unit converts the weld surface grayscale images I ₁ (x, y) and I ₂ (x, y) into binary images B ₁ (x, y ) and B ₂ (x, y); the processing control unit respectively extracts the connected domains with the largest area in the binary images B ₁ (x, y) and B ₂ (x, y), denoted as R ₁ and R ₂ ;

3) Let the number of pixels contained in the connected domain R ₁ and R ₂ be #R ₁ and #R ₂ respectively, the i-th pixel coordinate in the connected domain R ₁ is (x _1i , y _1i ), and the connected domain R The coordinates of the jth pixel in ₂ are (x _2j , y _2j ), where both #R ₁ and #R ₂ are positive integers, i is a positive integer greater than zero and less than or equal to #R ₁ , and j is greater than zero and A positive integer less than or equal to #R ₂ ; the processing control unit processes the connected domains R ₁ and R ₂ to extract the edge coordinates of the weld seam and the base metal:

a) If the angle between the weld track and the coordinate axis of the gray image of the weld surface is greater than 45°, calculate the connected domains R ₁ and R ₂

Average column coordinates of all pixels and

$\begin{array}{c}{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; i</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\\ {\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; j</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}\end{array}$

like Then the connected domain R ₁ is recorded as the left edge area of the weld, and the connected domain R ₂ is recorded as the right edge area of the weld; if Then the connected domain _R1 is recorded as the right edge area of the weld, and the connected domain _R2 is recorded as the left edge area of the weld; the left edge area of the weld is recorded as _L1 , and the right edge area of the weld is _L2 ; Assume that the set of column coordinate values of the pixel points in the _mth row of the left edge area L of the weld is Q _1m , and the set of column coordinate values of the pixel points in the mth row of the right edge area of the weld _L2 is Q _2m , where m is a positive integer greater than zero and less than or equal to the total number of lines of the weld surface grayscale image; scan the weld left edge area L ₁ and weld right edge area L ₂ line by line, and calculate the mth line of weld Sew the column coordinate Y _1m of the left edge point and the column coordinate Y _2m of the right edge point:

Y _1m ＝max(Q _1m )

Y _2m ＝min(Q _2m )

Among them, max(Q _1m ) represents the largest element of the set Q _1m , and min(Q _2m ) represents the smallest element of the set Q _2m ;

b) If the angle between the weld trajectory and the column coordinate axis of the weld surface gray image is less than or equal to 45°, calculate the average row coordinates of all pixels in the connected domains R ₁ and R ₂ and

$\begin{array}{c}{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; i</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\\ {\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; j</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}\end{array}$

like Then the connected domain R ₁ is recorded as the upper edge area of the weld, and the connected domain R ₂ is recorded as the lower edge area of the weld; if Then the connected domain R ₁ is recorded as the lower edge region of the weld, and the connected domain R ₂ is recorded as the upper edge region of the weld; the upper edge region of the weld is recorded as U ₁ , and the lower edge region of the weld is U ₂ ; Assume that the set of row coordinate values of the pixel points in the kth column of the upper edge area _U1 of the weld is P _1k , and the set of row coordinate values of the pixel points in the _kth column of the lower edge area of the weld seam U2 is P _2k -, where k is a positive integer greater than zero and less than or equal to the total number of columns of the weld surface gray image; scan the weld upper edge area U ₁ and weld lower edge area U ₂ column by column, and calculate the kth The row coordinate X _1k of the upper edge point of the column weld and the row coordinate X _2k of the lower edge point:

X _1k ＝max(P _1k )

X _2k ＝min(P _2k )

Wherein, max(P _1k ) represents the largest element of the set P _1k , and min(P _2k ) represents the smallest element of the set P _2k .

Compared with the prior art, the present invention does not rely too much on the macroscopic geometric structure characteristics of the weld surface, and the side lighting conditions constructed by it can effectively highlight the feature information of the weld edge, and the gray scale of the side wall of the groove is close to saturation, which is convenient Quickly and accurately extract the position of the weld trajectory; the detection accuracy can reach 0.04mm or even higher, the system structure is simple, the cost is low, and the real-time performance is good. It is suitable for the cover surface where the height difference between the weld seam and the base metal surface is as low as 1mm or even lower The occasion of weld track detection before welding.

Description of drawings

FIG. 1 is a schematic diagram of the structure and principle of an embodiment of a welding seam detection device for cover surface before welding using side light to construct light and shadow features proposed by the present invention.

Fig. 2 is a schematic diagram of the reflection of the main axis ray of the first surface light source ray on the surface of the workpiece in the embodiment of the welding seam inspection device before welding of the cover surface using side light to construct the light and shadow feature proposed by the present invention.

Fig. 3 is a flow chart of the first embodiment and the second embodiment of a welding seam inspection method for cover surface before welding using side light to construct light and shadow features proposed by the present invention.

Fig. 4 is a grayscale image of the weld surface collected by the imaging element when all the surface light sources are turned on simultaneously in the first embodiment of the method for detecting the weld seam before welding using side light to construct light and shadow features proposed by the present invention.

Fig. 5 is a processing result of extracting the maximum connected domain of a binary image in the first embodiment of a method for detecting a pre-weld weld seam on a cover surface using side light to construct light and shadow features proposed by the present invention.

Fig. 6 is the final extraction result of the weld track in the first embodiment of the method for detecting the weld seam on the cover surface before welding using side light to construct light and shadow features proposed by the present invention.

Fig. 7 is a grayscale image of the weld surface collected by the imaging element when the first surface light source is turned on in the second embodiment of the method for detecting the weld seam before welding on the cover surface using side light to construct light and shadow features proposed by the present invention.

Fig. 8 is a grayscale image of the weld surface collected by the imaging element when the second surface light source is turned on in the second embodiment of the method for detecting a weld seam before welding using side light to construct light and shadow features proposed by the present invention.

Fig. 9 is a processing result of extracting the maximum connected domain of a binary image in the second embodiment of a method for detecting a pre-weld weld seam on a cover surface using side light to construct light and shadow features proposed by the present invention.

Fig. 10 is the final extraction result of the weld track in the second embodiment of the method for detecting the weld seam before welding on the cover surface using side light to construct light and shadow features proposed by the present invention.

In Figures 1 to 10:

1—surface light source array; 11—first surface light source; 12—second surface light source; 2—imaging element; 3—processing control unit; 4—weld seam to be detected; The angle between the side wall of the groove at the edge of the base metal and the surface of the base metal; β—the angle between the main axis ray of the first surface light source and the surface of the base metal; R ₁ , R ₂ —the two connected domains with the largest areas in the binary image .

Detailed ways

The structure, principle and working process of the present invention will be further described below in conjunction with the accompanying drawings.

Fig. 1 is a schematic diagram of the structure and principle of an embodiment of a pre-weld weld detection device using side light to construct light and shadow features proposed by the present invention, including a surface light source array 1, an imaging element 2 and a processing control unit 3; The light source array 1 includes two surface light sources, which are respectively called the first surface light source 11 and the second surface light source 12; the first surface light source 11 and the second surface light source 12 use LED array strip light sources with a luminous power of 5.7W. The luminous wavelength is within the visible light wavelength range, and they are respectively placed on both sides of the weld 4 to be detected, and the light emitted by the surface light source on the same side is projected on the edge of the weld 4 to be detected on the other side and the base material; Mega Ethernet complementary metal oxide semiconductor imaging device, the sensitive wavelength range is greater than the visible light wavelength range, the image size is 1600×1200, the field of view is about 55.5mm×41.0mm, the detection accuracy can reach 0.04mm, and the image grayscale range is 0 to 255; the weld 4 to be detected is the weld before the cover welding, the height difference between the weld and the base metal surface is about 1mm, and the melting width is about 14mm; the processing control unit 3 uses an industrial computer with a CPU frequency of 2.3GHz and a memory of 4GB; The surface light source array 1 is connected to the processing control unit 3 through a wire; the imaging element 2 is connected to the processing control unit 3 through a wire; the processing control unit 3 processes the image collected by the imaging element 2 .

Fig. 2 is a schematic diagram of the reflection of the main axis light of the first surface light source 11 on the surface of the workpiece in the embodiment of the welding seam inspection device before welding of the cover surface using side light to construct the light and shadow features proposed by the present invention. Taking the first surface light source 11 as an example, assuming that the angle between the main axis ray of the first surface light source 11 and the surface of the base material is β, the groove side wall and the base material edge at the edge of the weld projected by the first surface light source 11 and the base material The angle between the surfaces is α. According to the law of light reflection, the angle between the main axis ray of the first surface light source 11 and the surface of the base material through the specular reflection of the groove side wall is β+2α; the main axis ray of the first surface light source 11 is reflected by the mirror surface of the base material surface The angle between the light and the surface of the base metal is β; since the surface of the weld is generally flat, the angle between the specularly reflected light of the main axis light of the first surface light source 11 and the surface of the base metal is also approximately β. In order to highlight the edge features of the weld, the imaging element 2 should be able to collect the specular reflection light of the side wall of the groove, and try to avoid the specular reflection light of the base material and the surface of the weld seam. Therefore, the main axis light of the first surface light source 11 should pass through The angle β+2α between the specular reflection light on the side wall of the groove and the surface of the base metal is as close to 90° as possible, and the angle β between the specular reflection light on the surface of the base metal and the weld surface and the surface of the base metal is as close to 0° as possible. In the embodiment of the present invention, the angle between the groove side wall and the surface of the base material at the edge of the weld projected by the first surface light source 11 and the base material is α=60°, so the main axis ray of the first surface light source 11 and the base material The angle β on the surface of the material should be as close to -30° as possible and as close to 0° as possible. Considering that the main axis ray of the first surface light source 11 must be projected on the edge of the weld seam and the base material, β should be greater than or equal to zero degrees. In this embodiment, β is set to be 10°. Similarly, the arrangement of the second surface light source 12 should also make the imaging element 2 collect the specularly reflected light at the sidewall of the groove, and the imaging element 2 does not collect the specularly reflected light on the base material and the weld seam surface.

Fig. 3 is a flow chart of the first embodiment and the second embodiment of a welding seam inspection method for cover surface before welding using side light to construct light and shadow features proposed by the present invention. In the first embodiment, all the surface light sources located on both sides of the weld seam 4 to be detected are turned on at the same time; The light source 12 lights up alternately at a frequency of 60 Hz.

For the first embodiment of the method for detecting a weld seam before welding using side light to construct light and shadow features proposed by the present invention, firstly, all the surface light sources located on both sides of the weld seam 4 to be inspected are simultaneously lit, and then the processing control Unit 3 controls the imaging element 2 to collect the weld surface grayscale image I(x, y) when the surface light source array 1 is turned on, where x, y are the row coordinate values and column coordinate values of the weld surface grayscale image respectively, and the imaging Component 2 transmits the weld seam surface grayscale image I(x,y) to processing control unit 3 . Fig. 4 is the weld surface grayscale image I collected by the imaging element 2 when all the surface light sources are turned on simultaneously in the first embodiment of the method for detecting the weld seam before welding using side light to construct light and shadow features proposed by the present invention. (x,y). It can be seen from Figure 4 that the side wall of the groove presents a bright area, and the gray value of the base metal and the surface of the weld is significantly lower than that of the side wall of the groove, which is consistent with the optical path analysis results in Figure 2, and it is expected to detect the weld quickly and accurately edge. After obtaining the grayscale image I(x,y) of the weld surface, the processing control unit 3 converts the grayscale image I(x,y) of the weld surface into a binary image B(x,y) using a threshold segmentation method. The segmentation threshold used here is 0.75 times the maximum gray value of 255 in the gray image of the weld surface. The processing control unit 3 extracts the two connected domains R ₁ and R ₂ with the largest areas in the binary image B(x,y).

Fig. 5 is the maximum connected region R ₁ and R ₂ of the binary image B(x, y) extracted in the first embodiment of a method for detecting a pre-weld weld seam using side light to construct light and shadow features proposed by the present invention. Processing results, where the connected domain on the left side of Figure 5 is R ₁ , and the connected domain on the right side is R ₂ . Let the number of pixels contained in the connected domain R ₁ and R ₂ be respectively #R ₁ and #R ₂ , the i-th pixel coordinate in the connected domain R ₁ is (x _1i , y _1i ), in the connected domain R ₂ The coordinates of the jth pixel of is (x _2j , y _2j ), where #R ₁ and #R ₂ are both positive integers, i is a positive integer greater than zero and less than or equal to #R ₁ , and j is greater than zero and less than or A positive integer equal to #R ₂ . In this embodiment, the included angle between the weld track and the coordinate axis of the weld surface grayscale image I(x, y) is greater than 45°. In order to distinguish the left edge region and the right edge region of the weld, the average column coordinates of the connected domains _R1 and _R2 are calculated:

$\begin{array}{c}{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; i</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\\ {\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; the\; y</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; j</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}\end{array}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

The average column coordinates of the region at the left edge of the weld should be smaller than the average column coordinates of the region at the right edge of the weld, so if Then the connected domain R ₁ is recorded as the left edge area of the weld, and the connected domain R ₂ is recorded as the right edge area of the weld; if Then the connected domain _R1 is recorded as the right edge area of the weld, and the connected domain _R2 is recorded as the left edge area of the weld. In this embodiment, the connected domain R ₁ is the region of the left edge of the weld, and the connected domain R ₂ is the region of the right edge of the weld.

Note that the left edge area of the weld is L ₁ , and the right edge area of the weld is L ₂ ; let the set of column coordinate values of the pixel points in the mth row of the left edge area L ₁ be Q _1m , so The set of column coordinate values of the pixel points in the mth row of the right edge area _L2 of the weld is Q _2m , where m is a positive integer greater than zero and less than or equal to the total number of rows of the grayscale image of the weld surface; line by line scanning For the left edge area L ₁ of the weld and the right edge area L ₂ of the weld, calculate the column coordinate Y _1m of the left edge point of the m-th row of the weld and the column coordinate Y _2m of the right edge point:

Y _1m =max(Q _1m )(2)

Y _2m ＝min(Q _2m )

Wherein, max(Q _1m ) represents the largest element of the set Q _1m , and min(Q _2m ) represents the smallest element of the set Q _2m . Fig. 6 is the final extraction result of the weld track in the first embodiment of the method for detecting the weld seam on the cover surface before welding using side light to construct light and shadow features proposed by the present invention. It can be seen that the weld edge is accurately extracted.

For the second embodiment of the method for detecting a pre-weld weld seam on a cover surface using side light to construct light and shadow features proposed by the present invention, the first surface light source 11 and the second surface light source 12 located on both sides of the weld seam 4 to be inspected are as follows: The frequency of 60Hz lights up alternately. First, the processing control unit 3 sends a trigger signal to turn on the first surface light source 11, turn off the second surface light source 12, and make the imaging element 2 synchronously collect the grayscale image I ₁ (x, y) of the weld surface; then, the processing The control unit 3 sends a trigger signal to turn on the second surface light source 12, turn off the first surface light source 11, and make the imaging element 3 synchronously collect the grayscale image I ₂ (x, y) of the weld surface; The surface grayscale images I ₁ (x, y) and I ₂ (x, y) are transmitted to the processing control unit 3, where x, y are row coordinate values and column coordinate values of the weld surface grayscale image respectively. Fig. 7 and Fig. 8 respectively show the images when the first surface light source 11 and the second surface light source 12 are turned on in the second embodiment of a method for detecting weld seams before welding on top surface using side light to construct light and shadow features proposed by the present invention. The grayscale image of the weld surface collected by component 2. From Figure 7 and Figure 8, it can be seen that the side wall of the groove presents a bright area, and the gray value of the base metal and the surface of the weld is significantly lower than that of the side wall of the groove, which is consistent with the optical path analysis results in Figure 2, and it is expected to quickly and accurately Detect weld edges. After obtaining the weld surface grayscale images I ₁ (x, y) and I ₂ (x, y), the processing control unit 3 uses the threshold segmentation method to respectively convert the weld surface grayscale images I ₁ (x, y) and I ₂ (x, y) is converted into binary images B ₁ (x, y) and B ₂ (x, y), and the segmentation threshold used here is 0.75 times the maximum gray value of 255 in the gray image of the weld surface. The processing control unit 3 extracts the connected domains R ₁ and R ₂ with the largest areas in the binary images B ₁ (x,y) and B ₂ (x,y) respectively.

Fig. 9 is the extraction of binary images B ₁ (x, y) and B ₂ (x, y) in the second embodiment of a method for detecting a pre-weld weld seam before welding using side light to construct light and shadow features proposed by the present invention Processing results of the largest connected domains R ₁ and R ₂ , where the lower connected domain in Figure 9 is R ₁ , and the upper connected domain is R ₂ . Let the number of pixels contained in the connected domain R ₁ and R ₂ be respectively #R ₁ and #R ₂ , the i-th pixel coordinate in the connected domain R ₁ is (x _1i , y _1i ), in the connected domain R ₂ The coordinates of the jth pixel of is (x _2j , y _2j ), where #R ₁ and #R ₂ are both positive integers, i is a positive integer greater than zero and less than or equal to #R ₁ , and j is greater than zero and less than or A positive integer equal to #R ₂ . In this embodiment, the included angle between the welding seam track and the coordinate axes of the weld surface grayscale images I ₁ (x, y) and I ₂ (x, y) is less than 45°, in order to distinguish the upper edge area of the weld seam from the lower one. For edge regions, calculate the average row coordinates of connected domains _R1 and _R2 :

$\begin{array}{c}{\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; i</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}}}\\ {\stackrel{<span\; class="notranslate">\; \&OverBar;</span>}{\mathrm{<span\; class="notranslate">\; x</span>}}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}<span\; class="notranslate">\; =</span>\frac{\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}{\mathrm{<span\; class="notranslate">\; x</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; j</span>}}}{<span\; class="notranslate">\; \#</span>{\mathrm{<span\; class="notranslate">\; R</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}}}\end{array}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 3</span>}<span\; class="notranslate">\; )</span>$

The average row coordinate of the upper edge area of the weld should be smaller than the average row coordinate of the lower edge area of the weld, so if Then the connected domain R ₁ is recorded as the upper edge area of the weld, and the connected domain R ₂ is recorded as the lower edge area of the weld; if Then the connected domain _R1 is recorded as the lower edge region of the weld, and the connected domain _R2 is recorded as the upper edge region of the weld. In this embodiment, the connected domain R ₁ is the lower edge region of the weld, and the connected domain R ₂ is the upper edge region of the weld.

Note that the upper edge area of the weld is U ₁ , and the lower edge area of the weld is U ₂ ; let the set of the row coordinate values of the pixel points in the kth column of the upper edge area U ₁ be P _1k , so The set of row coordinate values of the pixel points in the kth column of the weld lower edge area U ₂ is P _2k , where k is a positive integer greater than zero and less than or equal to the total number of columns of the weld surface grayscale image; column by column Scan the upper edge area U ₁ of the weld and the lower edge area U ₂ of the weld, and calculate the row coordinate X _1k of the upper edge point of the k-th column of the weld and the row coordinate X _2k of the lower edge point of the kth column:

X _1k ＝max(P _1k )(4)

X _2k ＝min(P _2k )

Wherein, max(P _1k ) represents the largest element of the set P _1k , and min(P _2k ) represents the smallest element of the set P _2k . Fig. 10 is the final extraction result of the weld track in the second embodiment of the method for detecting the weld seam before welding on the cover surface using side light to construct light and shadow features proposed by the present invention. It can be seen that the weld edge is also accurately extracted.

Experimental results show that, in the embodiment of the present invention, the detection error of the device and method of the present invention does not exceed 0.04 mm, and the processing time of a single image does not exceed 15 ms, meeting the requirements of high precision and high real-time performance.

It should be noted that the above embodiments are only used to illustrate the present invention rather than limit the solution described in the present invention; therefore, although the specification has described the present invention in detail with reference to the above embodiments, those of ordinary skill in the art should understand , the present invention can still be modified or equivalently replaced, such as the number of surface light sources used can be 3 or more, the processing control unit and imaging elements can communicate with wireless transmission, etc.; and everything does not depart from the spirit and scope of the present invention The technical solutions and their improvements shall be included in the scope of the claims of the present invention.

The present invention adopts a surface light source array, an imaging element, and a processing control unit to realize the automatic detection of the weld track before the cover surface welding; the side illumination condition constructed by the present invention can effectively construct the light and shadow features at the edge of the weld, and the imaging element is placed on the groove The optical axis of the specular reflection of the side wall is near, and the optical axis of the specular reflection of the base material and the weld surface is avoided, the grayscale image of the weld surface is collected, and the weld track is accurately extracted after image processing by the processing control unit. Compared with the traditional structured light detection method, the device and method proposed by the present invention do not rely too much on the macro-geometric structural characteristics of the weld, and the gray features at the edge of the weld are obvious, which is convenient for fast and accurate detection of the position of the weld track, and detection The accuracy can reach 0.04mm; the system has simple structure, low cost, and good real-time performance. It is suitable for the automatic detection of the welding seam trajectory before the welding seam and the base metal surface height difference as low as 1mm or even lower.

Claims (3)

1. use lateral light to build a detection device before the cosmetic welding of shadow feature, it is characterized in that: comprise area source array (1), image-forming component (2) and processing and control element (PCE) (3); Described area source array (1) comprises at least two area sources, described area source is placed on weld seam to be detected (4) both sides respectively, the to be detected weld seam (4) of ray cast at opposite side that the area source of homonymy sends and the edge of mother metal; The relative position of described area source, weld seam to be detected and described image-forming component (2) meets: image-forming component (2) gathers the specular light of axial principal ray through groove sidewall of each area source, and image-forming component (2) axial principal ray that do not gather each area source is through the specular light on mother metal and weld seam to be detected (4) surface; Described area source array (1) is connected by wire with processing and control element (PCE); Described image-forming component (2) is connected by wire with processing and control element (PCE) (3), or by wireless transmission method communication; The image that described processing and control element (PCE) (3) process image-forming component (2) gathers.

2. a kind of lateral light that uses as claimed in claim 1 builds detection device before the cosmetic welding of shadow feature, it is characterized in that: described image-forming component is charge-coupled image sensor, complementary metal oxide semiconductor (CMOS) image device, position sensitive detector or charge injection device; The sensitive wave length scope of described image-forming component is more than or equal to the emission wavelength range of described area source.

3. a kind of lateral light that uses adopting device as claimed in claim 1 or 2 builds weld inspection method before the cosmetic welding of shadow feature, it is characterized in that the method comprises the following steps:

1) all area sources being positioned at weld seam both sides to be detected are lighted simultaneously, or the area source being positioned at weld seam to be detected not homonymy is alternately lighted;

2) ray cast that sends of described area source is at weld edge place to be detected, described processing and control element (PCE) controls described image-forming component and gathers image, the face of weld gray level image of collection is transferred to described processing and control element (PCE) by described image-forming component, and processing and control element (PCE) carries out pre-service to described face of weld gray level image:

If a) all area sources being positioned at weld seam both sides to be detected are lighted simultaneously, described processing and control element (PCE) controls image-forming component and gathers face of weld gray level image when described area source array is lighted, be designated as I (x, y), wherein x, y is row-coordinate value and the row coordinate figure of face of weld gray level image respectively, and face of weld gray level image is transferred to described processing and control element (PCE) by described image-forming component; Processing and control element (PCE) uses threshold segmentation method, and described face of weld gray level image I (x, y) is converted to bianry image B (x, y); Processing and control element (PCE) extracts two connected domains that in described bianry image B (x, y), area is maximum, is designated as R ₁and R ₂;

If the area source b) being positioned at weld seam to be detected not homonymy is alternately lighted, described processing and control element (PCE) sends trigger pip, makes image-forming component gather face of weld gray level image when the area source of not homonymy is lighted respectively, is designated as I respectively ₁(x, y) and I ₂(x, y), wherein x, y are row-coordinate value and the row coordinate figure of face of weld gray level image respectively, and face of weld gray level image is transferred to described processing and control element (PCE) by described image-forming component; Processing and control element (PCE) uses threshold segmentation method, respectively by described face of weld gray level image I ₁(x, y) and I ₂(x, y) is converted to bianry image B ₁(x, y) and B ₂(x, y); Described processing and control element (PCE) extracts described bianry image B respectively ₁(x, y) and B ₂the connected domain that in (x, y), area is maximum, is designated as R ₁and R ₂;

3) connected domain R is established ₁and R ₂the pixel number comprised is respectively #R ₁and #R ₂, connected domain R ₁in i-th pixel coordinate be (x _1i, y _1i), connected domain R ₂in a jth pixel coordinate be (x _2j, y _2j), wherein #R ₁and #R ₂are all positive integers, i is for being greater than zero and being less than or equal to #R ₁positive integer, j is for being greater than zero and being less than or equal to #R ₂positive integer; Described processing and control element (PCE) is to described connected domain R ₁and R ₂process, extract the edge coordinate of weld seam and mother metal:

If a) seam track and face of weld gray level image row coordinate axis angle are greater than 45 °, then calculate connected domain R ₁and R ₂the average row coordinate of all pixels with

${\stackrel{\&OverBar;}{y}}_1=\frac{\underset{i=1}{\overset{\#R_1}{\mathrm{\&Sigma;}}}{y}_{1i}}{\#R_1}$

${\stackrel{\&OverBar;}{y}}_2=\frac{\underset{j=1}{\overset{\#R_2}{\mathrm{\&Sigma;}}}{y}_{2j}}{\#R_2}$

If then by connected domain R ₁be designated as weld seam left hand edge region, by connected domain R ₂be designated as weld seam right hand edge region; If then by connected domain R ₁be designated as weld seam right hand edge region, by connected domain R ₂be designated as weld seam left hand edge region; Remember that described weld seam left hand edge region is L ₁, described weld seam right hand edge region is L ₂; If described weld seam left hand edge region L ₁the set of the row coordinate figure composition of the capable pixel of m is Q _1m, described weld seam right hand edge region L ₂the set of the row coordinate figure composition of the capable pixel of m is Q _2m, wherein m is the positive integer being greater than zero and being less than or equal to the total line number of described face of weld gray level image; Line by line scan described weld seam left hand edge region L ₁with weld seam right hand edge region L ₂, calculate the row coordinate Y of the capable weld seam left hand edge point of m _1mwith the row coordinate Y of right hand edge point _2m:

Y _1m＝max(Q _1m)

Y _2m＝min(Q _2m)

Wherein, max (Q _1m) represent set Q _1mgreatest member, min (Q _2m) represent set Q _2mleast member;

If b) seam track and face of weld gray level image row coordinate axis angle are less than or equal to 45 °, then calculate connected domain R ₁and R ₂the average row coordinate of all pixels with

${\stackrel{\&OverBar;}{x}}_1=\frac{\underset{i=1}{\overset{\#R_1}{\mathrm{\&Sigma;}}}{x}_{1i}}{\#R_1}$

${\stackrel{\&OverBar;}{x}}_2=\frac{\underset{j=1}{\overset{\#R_2}{\mathrm{\&Sigma;}}}{x}_{2j}}{\#R_2}$

If then by connected domain R ₁be designated as weld seam upper edge region, by connected domain R ₂be designated as weld seam lower edge margin; If then by connected domain R ₁be designated as weld seam lower edge margin, by connected domain R ₂be designated as weld seam upper edge region; Remember that described weld seam upper edge region is U ₁, described weld seam lower edge margin is U ₂; If described weld seam upper edge region U ₁the set of the row-coordinate value composition of kth row pixel is P _1k, described weld seam lower edge margin U ₂the set of the row-coordinate value composition of kth row pixel is P _2k, wherein k is the positive integer being greater than zero and being less than or equal to the total columns of described face of weld gray level image; Scan by column described weld seam upper edge region U ₁with weld seam lower edge margin U ₂, calculate the row-coordinate X of kth row weld seam up contour point _1kwith the row-coordinate X of down contour point _2k:

X _1k＝max(P _1k)

X _2k＝min(P _2k)

Wherein, max (P _1k) represent set P _1kgreatest member, min (P _2k) represent set P _2kleast member.