JP2010223621A - Inner surface inspection method of tubular article - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an inner surface inspection method that minimizes the installation space of a camera and a light source when inner surface inspection of a tubular article is performed, certainly determines a harmful uneven defect and a harmless pattern, and can detect only the uneven defect. <P>SOLUTION: This inner surface inspection method includes a camera that has an optical center axis inclined with respect to the center axis of the tubular article and images a predetermined region of the inner surface of the tubular article from the open end of the tubular article, and a pair of light sources that are arranged symmetrically on both sides of a plane including the optical center axis of the camera and the center axis of the tubular article and illuminate the predetermined region from the open end. The same predetermined region is imaged by the camera while the illuminations from the light sources are switched sequentially, and the shading difference between the same pixels of each image is calculated, and shadows 5a, 5b, 6a and 6b appearing due to the uneven defects 5 and 6 are recognized and the uneven defects 5 and 6 are detected based on the information on the shading difference. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a method for inspecting an inner surface of a tubular article, in which an inner surface of the tubular article is imaged with a camera and image processing is performed to detect an irregularity defect.

  Tubular products such as steel pipes may have irregularities on the inner surface due to the manufacturing process, or may have irregularities such as scale unevenness or dirt on the inner surface. Unevenness is a defect that is harmful in quality, and the pattern is harmless because it does not have unevenness. For this reason, from the viewpoint of quality assurance, the inner surface inspection for detecting irregularities on the inner surface is indispensable for the tubular product.

  Conventionally, as a method for inspecting the inner surface of a tubular product, there are a method of visually observing the inner surface from one opening end and the other opening end, or a method of performing insertion by inserting an endoscope from the opening end. . An inner surface inspection method using an endoscope is roughly classified into a forward view and a lateral view.

  FIG. 1 is a schematic diagram for explaining an inner surface inspection method by forward viewing using an endoscope. As shown in the figure, in the front surface inspection method using the endoscope, the camera 101 on which the light source 102 is mounted is inserted into the tubular article 10 and the entire inner surface 11 illuminated by the light source 102 is inserted. The circumference is imaged by the camera 101 and inspected.

  FIG. 2 is a schematic view for explaining an inner surface inspection method by side view using an endoscope. As shown in the figure, in the method of inspecting the inner surface by side view using an endoscope, a reflecting mirror 103 is installed in front of a camera 101 on which a light source 102 is mounted, and these are centered inside a tubular article 10. A part of the inner surface 11 inserted along the axis and illuminated by the light source 102 via the reflecting mirror 103 is imaged and inspected by the camera 101 via the reflecting mirror 103. At this time, the inner surface 11 of the tubular article 10 can be inspected over the entire circumference by rotating the reflecting mirror 103 around the optical center axis of the camera 101.

  These conventional inner surface inspection methods have the following problems. In the method by visual observation, when inspecting the inner surface on the far side from the open end of the tubular product, the distance from the observation position to the inspection area is long, so it is difficult to visually recognize the irregular defect and the pattern. There is a possibility of oversight or misidentifying the pattern as a concave-convex defect.

  In the method using an endoscope, it is difficult to discriminate a concavo-convex defect from a pattern because a stereoscopic effect is hardly generated in an image captured by a camera in both forward and lateral views. Further, in the method using an endoscope, it is necessary to insert a camera into the tubular product, and the operation is extremely complicated.

  For these reasons, in the inner surface inspection, it is required to reliably discriminate harmful irregularities and harmless patterns and detect only the irregularities, but the above-mentioned conventional inner surface inspection methods meet the requirements. I cannot respond enough.

  Conventionally, the following methods have been proposed as a method for discriminating irregularities and patterns. Patent Document 1 discloses a method of inspecting the outer surface of a link plate such as a chain or a connection pin. In the inspection method disclosed in Patent Document 1, a defect candidate is extracted from an image obtained by imaging with specularly reflected light, and whether the defect is a concavo-convex defect or a pattern depending on the presence or absence of the defect candidate in the image obtained by imaging with irregularly reflected light. This is a method of determining.

  Patent Document 2 discloses a method of inspecting an outer surface of a thin steel plate or a thick steel plate as an inspection target. In the inspection method disclosed in Patent Document 2, the inspection area is imaged almost vertically from two directions while illuminating the inspection area with a single light source, the positional deviation of each image is corrected, and then both images are identical. In this method, a difference between pixels is determined to determine whether the defect is a concavo-convex defect or a pattern.

JP 2007-271510 A JP 2006-177852 A

  However, when the inspection method disclosed in Patent Document 1 is applied to the inner surface inspection of a tubular product, a camera is installed at one opening end of the tubular product, and two different emission angles of illumination light are provided at the other opening end. It is necessary to install a light source. For this reason, a sufficient space for installing the camera and the light source is required outside both open ends of the tubular product, and the inspection apparatus for realizing this method becomes large-scale.

  When the inspection method disclosed in Patent Document 2 is applied to the inner surface inspection of a tubular product, it is necessary to install two cameras and a light source inside the tubular product. For this reason, it is difficult to cope with a tubular product having a small inner diameter. In addition, in the process of discriminating between the concave and convex defects and the pattern, the positional deviation between the two images must be corrected before evaluating each image taken by the two cameras. For this reason, image processing becomes extremely complicated. In particular, when performing an inner surface inspection over the axial direction of a tubular product, image processing becomes more complicated. This is because since the distance from the camera position to the imaging region changes, the conversion amount used for correcting the positional deviation of the image must be changed.

  Therefore, it is difficult to employ the inspection method disclosed in Patent Documents 1 and 2 for the inner surface inspection of a tubular product mainly due to restrictions on the installation space of the camera and the light source.

  The present invention has been made in view of the above problems, and when performing the inner surface inspection of a tubular product, while minimizing the space required for installation of the camera and the light source, harmful concave and convex defects and harmless It is an object of the present invention to provide an inner surface inspection method capable of reliably discriminating a pattern and detecting only uneven defects.

  In order to achieve the above object, the present inventor has conducted extensive studies on the premise that the inner surface of the tubular product is inspected using image processing. In order to reliably detect defects and patterns, a camera and a pair of light sources are installed outside the open end of one of the tubular products, and images are captured by the camera using illumination from each light source, and both images are processed. It has been found that it is effective to evaluate the shadow appearing due to the irregular defect.

  The present invention has been completed based on the above findings, and the gist thereof is the following inner surface inspection method. That is, an inner surface inspection method for inspecting an inner surface of a tubular article, the optical article having an optical center axis inclined with respect to the central axis of the tubular article, and an inner surface of the tubular article from an open end of the tubular article Using a camera that captures an image of a predetermined area, and a pair of light sources that are arranged symmetrically across a plane that includes the optical central axis of the camera and the central axis of the tubular article, and that illuminates the predetermined area from the opening end, The illumination from the light source is switched in order and the same predetermined area is imaged by the camera, and a light / dark difference is calculated between the same pixels of each image. Based on the light / dark difference information, an irregular defect appears. An inner surface inspection method characterized by recognizing a shadow and detecting an uneven defect.

  In this inspection method, the camera and the light source are rotated relative to each other about the central axis of the tubular article, and the inclination angle of the camera and the light source with respect to the central axis of the tubular article is adjusted. It is possible to adopt a configuration in which imaging by the camera with switching is repeatedly performed to detect uneven defects.

  In addition, the above inspection method can be configured to discriminate between a convex defect and a concave defect based on light and shade in an original image of a pixel adjacent to a pixel recognized as a shadow. In the case of this inspection method, the size of the defect can be calculated based on the distribution area of the pixels recognized as a shadow.

  According to the method for inspecting the inner surface of a tubular article of the present invention, by performing imaging with a camera with illumination from each light source, performing image processing on both images, and evaluating the shadow that appears due to the irregular defect, the irregular defect And the pattern can be reliably discriminated and only the irregularity defect can be detected. In addition, since the camera and the pair of light sources are collectively installed outside the one opening end of the tubular product, the installation space for the camera and the light source can be minimized.

It is a schematic diagram for demonstrating the inner surface test | inspection method by the front view using an endoscope. It is a schematic diagram for demonstrating the inner surface inspection method by the side view using an endoscope. It is a figure which shows typically the structure of the test | inspection apparatus which can apply the inner surface test | inspection method of this invention, The figure (a) is a top view, The figure (b) is a side view, The figure (c) is the figure ( FIG. 4A is a view taken along the arrow S in FIG. 4A, and FIG. It is a schematic diagram for demonstrating the inner surface inspection method of this invention, the figure (a) shows the imaging state by illumination of a right light source, the figure (b) shows the imaging state by illumination of a left light source, respectively. FIG. 2C shows an example of the image in FIG. 2A, and FIG. 2D shows an example of the image in FIG. In order to explain the contents of the image processing in the inner surface inspection method of the present invention, each image is divided into convex defects, concave defects, and patterns. It is a schematic diagram for demonstrating the method of calculating the magnitude | size of a convex defect in the image process employ | adopted with the inner surface inspection method of this invention.

  As described above, the method for inspecting the inner surface of a tubular product according to the present invention has an optical central axis that is inclined with respect to the central axis of the tubular product, and a predetermined region on the inner surface of the tubular product from the open end of the tubular product. From the light source, a camera for imaging, and a pair of light sources that are arranged symmetrically across a plane that includes the optical central axis of the camera and the central axis of the tubular product, and that illuminates the predetermined region from the opening end The same predetermined area is imaged with the camera by sequentially switching the illumination, and a light / dark difference is calculated between the same pixels of each image, and a shadow appearing due to the uneven defect is recognized based on the light / dark difference information. This is an inner surface inspection method for detecting uneven defects. Hereinafter, an inner surface inspection method of the present invention and an inspection apparatus to which the inspection method can be applied will be described in detail.

1. FIG. 3 is a diagram schematically showing the configuration of an inspection apparatus to which the inner surface inspection method of the present invention can be applied. FIG. 3 (a) is a top view, FIG. 3 (b) is a side view, and FIG. c) is a view taken in the direction of arrow S in FIG. 4A, and FIG. 4D is a cross-sectional view taken along line TT in FIG. In addition, in the figure (a) and (b), the tubular article 10 which is a test object is shown with sectional drawing for convenience.

  As shown in FIGS. 3A to 3D, the inspection apparatus to which the inner surface inspection method of the present invention can be applied includes one camera 1 and a pair of light sources 2A and 2B. The camera 1 and the light sources 2 </ b> A and 2 </ b> B are collectively installed outside one open end of the tubular product 10.

  As shown in FIG. 3B, the camera 1 has an optical center axis CCL that is inclined downward at an angle θ with respect to the center axis CL of the tubular article 10. As shown in FIGS. 3A and 3B, the camera 1 images the inspection area 11 a illuminated by the light source 2 </ b> A or the light source 2 </ b> B in the inner surface 11 of the tubular article 10.

  As shown in FIG. 3A, the light sources 2A and 2B are arranged symmetrically across a plane including the optical center axis CCL of the camera 1 and the center axis CL of the tubular article 10. Specifically, in the light sources 2A and 2B, the optical center axis LCL of the illumination light emitted from each of the light sources 2A and 2B is angled with respect to the center axis CL of the tubular article 10 in a side view as shown in FIG. In addition to being inclined downward at α, as shown in FIG. 3D, it is inclined inward at an angle γ with respect to the optical center axis CCL of the camera 1.

  Further, as shown in FIGS. 3A, 3B, and 3D, the center position of the front surfaces (light exit ports) of the light sources 2A and 2B and the center position of the front surface (light receiving port) of the camera 1 are as follows. The optical center axes LCL of the light sources 2A and 2B intersect at the intersection point P between the optical center axis CCL of the camera 1 and the inspection region 11a. The light sources 2A and 2B can be turned off / on individually, and individually illuminate the inspection region 11a of the inner surface 11 of the tubular article by switching.

  For the inner surface inspection, one light source 2A (hereinafter also referred to as “right light source”) disposed on the right side of the camera 1 and the other light source 2B (hereinafter also referred to as “left light source”) disposed on the left side are used. It switches in order, and the same test | inspection area | region 11a illuminated with each light source 2A, 2B is imaged with the camera 1 separately. Image processing to be described later is performed on the images obtained by the respective imaging, whereby the inspection region 11a can be inspected.

  In addition, the camera 1 is preferably configured so that the tilt angle θ can be adjusted with the intersection point P as a fulcrum, but the tilt angle θ can be easily adjusted with the center position of the front surface of the camera 1 as a fulcrum. May be. On the other hand, the light sources 2 </ b> A and 2 </ b> B are configured to be able to adjust the tilt angle γ in accordance with the adjustment of the tilt angle θ of the camera 1 and satisfy the following expression (1).

γ = tan ⁻¹ ((A × cos θ) / L) (1)
In the equation, A is the distance between the center position of the front surfaces of the light sources 2A and 2B and the optical center axis CCL of the camera 1, as shown in FIG. L is the distance between the intersection point P and the center position of the front surface of the camera 1 in the axial direction of the tubular article 10 as shown in FIG.

  By adjusting the tilt angle θ of the camera 1 and the tilt angle γ of the light sources 2A and 2B, the inspection region 11a can be sequentially moved in the axial direction of the tubular article 10. In each case, by repeatedly performing imaging by the camera 1 accompanied by switching between the light sources 2A and 2B and executing image processing, it is possible to perform an inner surface inspection in the axial direction.

  At this time, it is preferable that the camera 1 has a zoom function linked to the adjustment of the tilt angle θ. This is because the field of view of the camera 1 varies as the inspection region 11a moves in the axial direction of the tubular article 10.

  Furthermore, it is preferable that the light sources 2A and 2B have an illuminance adjustment function linked to the adjustment of the inclination angle γ. This is because as the inspection region 11a moves in the axial direction of the tubular article 10, the illuminance to the inspection region 11a varies, and the luminance of the image acquired by the camera 1 varies. In order to adjust the brightness of the image, the camera 1 may be provided with an aperture adjustment function instead of the illuminance adjustment function of the light sources 2A and 2B.

  In the present invention, the tubular article 10 is configured to be rotatable about its central axis CL. By rotating the tubular product 10, the inspection region 11 a can be sequentially moved in the circumferential direction of the tubular product 10. In each case, if the imaging by the camera 1 accompanied by switching between the light sources 2A and 2B is repeatedly performed and image processing is executed, the inner surface inspection can be performed over the entire circumference. In order to perform the inner surface inspection over the entire circumference, the camera 1 and the light sources 2A and 2B may be integrally rotated about the central axis CL of the tubular product 10 instead of the rotation of the tubular product 10.

2. Inspection Method FIGS. 4A and 4B are schematic diagrams for explaining the inner surface inspection method of the present invention. FIG. 4A shows an imaging state by illumination of the right light source, and FIG. 4B shows imaging by illumination of the left light source. FIG. 4C shows an example of the image in FIG. 4A, and FIG. 4D shows an example of the image in FIG. FIGS. 3A to 3D illustrate cases where convex defects and patterns exist in the inspection area.

  As shown in FIG. 4A, the inspection area is illuminated only by the right light source 2 </ b> A, and the inspection area is imaged by the camera 1. At this time, a shadow 5a extending from the left side of the convex defect 5 to the back side appears in the inspection region. In this case, as shown in FIG. 4C, a convex defect 5, a shadow 5 a that appears on the left side of the convex defect 5, and a pattern 7 appear in the image of the camera 1.

  Next, as shown in FIG. 4B, the light source is switched, the same inspection area is illuminated only by the left light source 2 </ b> B, and the inspection area is imaged by the camera 1. At this time, a shadow 5b extending from the right side of the convex defect 5 to the back side appears in the inspection region, contrary to the case of illumination by the right light source 2A. In this case, as shown in FIG. 4D, a convex defect 5, a shadow 5b that appears on the right side of the convex defect 5, and a pattern 7 appear in the image of the camera 1.

  Then, the image processing described with reference to FIG. 5 below is performed on the two images respectively obtained by illumination of the right light source 2A and the left light source 2B, and a defect is detected.

  FIG. 5 is a diagram in which each image is divided into a convex defect, a concave defect, and a pattern in order to explain the contents of the image processing in the inner surface inspection method of the present invention. In the figure, images of a convex defect, a concave defect, and a pattern are shown in order from the top, and an image by the right light source, an image by the left light source, and a difference image between these two images are shown in order from the left column.

  For two images respectively obtained by illumination of the right light source and the left light source, the difference in shading between the pixels at the same position is sequentially calculated based on the luminance information of each pixel. At this time, since the two images are images obtained by switching only the illumination using the same camera, there is no positional deviation between the two images, and the complicated correction processing that is essential in the image processing of Patent Document 2 Is absolutely unnecessary. When the absolute value of the grayscale difference of each pixel obtained by the calculation is represented as an image, the image shown in the right column of FIG. 5 is obtained.

  In the case of the convex defect 5, as shown in the upper part of FIG. 5, the pixels of the shadows 5a and 5b appearing due to the convex defect 5 are greatly different in density between the right illumination image and the left illumination image. Therefore, the absolute value of the light / dark difference is high and dark. Since the pixels of the convex defect 5 itself existing between the shadows 5a and 5b are equally thin in the right illumination image and the left illumination image, the absolute value of the density difference is low and thin.

  For this reason, by setting the threshold value of the light and shade difference, the portion exceeding the threshold value can be recognized as the shadows 5a and 5b, and the convex defect 5 in which the shadows 5a and 5b appear can be detected. At this time, for the pixels adjacent to the shadows 5a and 5b in the difference image, that is, the pixels existing on the near side between the shadows 5a and 5b (the part indicated by the reference numeral “5c” in FIG. 5), the shading in the original image is light. If so, the defect can be reliably determined to be a convex defect.

  In the case of the concave defect 6, as shown in the middle of FIG. 5, a shadow 6a appears from the front side to the right side inside the concave defect 6 by the right side illumination, and a shadow from the near side to the left side inside the concave defect 6 by the left side illumination. 6b appears.

  Since the pixels of the shadows 6a and 6b appearing due to the concave defect 6 are greatly different in density between the right side illumination image and the left side illumination image except for the near side part, the absolute value of the density difference Is high and dark. Since the pixels in the front side portion 6c of the shadows 6a and 6b are equally dark in the right side illumination image and the left side illumination image, the absolute value of the density difference is low and light. Since the pixels of the portion where the shadows 6a and 6b do not exist inside the concave defect 6 are equally thin in the right illumination image and the left illumination image, the absolute value of the density difference is low and thin.

  For this reason, the part exceeding a threshold value can be recognized as shadow 6a, 6b, and it becomes possible to detect the concave defect 6 which makes shadow 6a, 6b appear. At this time, with respect to the pixels adjacent to the shadows 6a and 6b in the difference image, that is, the pixels existing in the lower part of the shadows 6a and 6b in the original image (portions indicated by reference numeral “6c” in FIG. 5), If it is dark, it can be reliably determined that the defect is a concave defect.

  In the case of the pattern 7, no shadow appears as shown in the lower part of FIG. Since the pixels of the pattern 7 have the same shade in the right illumination image and the left illumination image, the absolute value of the shade difference is low and does not exceed the threshold value. For this reason, the pattern 7 is not determined as a defect.

  As described above, in the inner surface inspection method of the present invention, imaging with a camera is performed with illumination from each light source, both images are image-processed, and the shadow appearing due to the irregular defect is evaluated, thereby determining the irregular defect. The pattern can be reliably identified.

  In addition, since the camera and the pair of light sources are collectively installed outside the one open end of the tubular product, the installation space for the camera and the light source can be minimized.

  FIG. 6 is a schematic diagram for explaining a technique for calculating the size of the convex defect in the image processing employed in the inner surface inspection method of the present invention. For example, the height h of the convex defect 5 can be simply calculated by the following equation (2).

h = l / tan α (2)
In the equation, l is the actual length of the shadows 5a and 5b caused by the convex defect 5, and is obtained by the following equation (3). As shown in FIG. 3B, α is an inclination angle of the optical center axis LCL of the light sources 2A and 2B in side view.

l = ((X / 2) ² + (Y / sin θ) ² ) ^1/2 (3)
In the equation, as shown in FIG. 6, X is the horizontal length of the whole area where the pixels of the shadows 5a and 5b are distributed in the difference image, and Y is the whole area where the shadows 5a and 5b are distributed in the difference image. It is the length in the vertical direction. Here, the lateral length means a length corresponding to the circumferential direction of the tubular product, and the longitudinal length means a length corresponding to the axial direction. θ is an inclination angle of the optical center axis CCL of the camera 1 as shown in FIG.

  From the above equations (2) and (3), the height h of the convex defect 5 can be easily obtained, and the size of the convex defect 5 can be calculated.

  The image processing described above is executed by a computer connected to the camera 1.

  In addition, the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the spirit of the present invention. For example, the height position of the camera is not limited as long as the inner surface of the tubular product can be viewed. However, in order to accurately inspect the inner side of the tubular product, it is effective to secure a large tilt angle of the optical center axis of the camera. Therefore, the height position of the camera is higher than the central axis of the tubular product. It is preferable to set a high position.

  Further, when inspecting the inner surface in the vicinity of the opening end of the tubular article, there is a possibility that the difference in density does not occur in the image in the inspection region due to the reflection of the illumination at the opening end surface. In order to avoid this, a trumpet-shaped cover can be attached to the open end of the tubular product for inspection.

  According to the method for inspecting the inner surface of a tubular article of the present invention, by performing imaging with a camera with illumination from each light source, performing image processing on both images, and evaluating the shadow that appears due to the irregular defect, the irregular defect And the pattern can be reliably discriminated and only the irregularity defect can be detected. In addition, since the camera and the pair of light sources are collectively installed outside the one opening end of the tubular product, the installation space for the camera and the light source can be minimized.

1: camera, 2A, 2B: light source,
5: convex defect, 5a, 5b: shadow of convex defect, 5c: near side of convex defect,
6: concave defect, 6a, 6b: shadow of concave defect, 6c: front side of shadow of concave defect,
7: Pattern, 10: Tubular product, 11: Inner surface, 11a: Inspection area,
CL: the central axis of the tubular product, CCL: the optical central axis of the camera,
LCL: Optical center axis of the light source

Claims (4)

An inner surface inspection method for inspecting an inner surface of a tubular article,
A camera having an optical central axis inclined with respect to the central axis of the tubular article, and imaging a predetermined region of the inner surface of the tubular article from an open end of the tubular article;
A pair of light sources that are arranged symmetrically across a plane including the optical central axis of the camera and the central axis of the tubular product, and illuminate the predetermined area from the opening end,
The illumination from the light source is switched in order and the same predetermined area is imaged by the camera, and a light / dark difference is calculated between the same pixels of each image. Based on the light / dark difference information, an irregular defect appears. An inner surface inspection method characterized by recognizing a shadow and detecting an uneven defect.   The camera and the light source are switched by rotating the camera and the light source relative to each other about the central axis of the tubular product, and adjusting an inclination angle of the camera and the light source with respect to the central axis of the tubular product. The inner surface inspection method according to claim 1, wherein the imaging is repeatedly performed to detect a concavo-convex defect.   3. The inner surface inspection method according to claim 1, wherein a convex defect and a concave defect are discriminated based on shading in an original image of a pixel adjacent to a pixel recognized as a shadow.   The inner surface inspection method according to claim 3, wherein the size of the defect is calculated based on a distribution area of pixels recognized as a shadow.

Images