PIPE SURFACE TESTING EQUIPMENT, PIPE SURFACE TESTING METHODS, PIPE MANUFACTURING EQUIPMENT, PIPE MANUFACTURING METHODS, PIPE QUALITY CONTROL METHODS, AND PIPES

Abstract

The pipe surface inspection device of the invention is a pipe surface inspection device which inspects the inner surface of the pipe as the inspection target area. The device consists of an imaging unit which includes an image sensor with a field of view on the surface of the inner surface, and two distinguishable light sources, arranged on the axis of the pipe so that the image sensor is sandwiched between the two distinguishable light sources. The processing unit is configured to collect the images obtained by capturing the reflection of each light source, creating a differential image between the two collected images, and detecting surface defects in the inspection target area from the generated differential image. The scanning unit is configured to scan the imaging unit along the axis of the pipe, where the two distinguishable light sources irradiate the same predetermined inspection target area, and the image sensor records the image of the predetermined inspection target area.The invention also relates to a method for inspecting the surface of a pipe to inspect the inner surface of the pipe as the target area for inspection.

Description

Pipe surface inspection device, surface inspection method, manufacturing device, manufacturing method, quality control method, and pipe

[0001] The present invention relates to a pipe surface inspection device, a surface inspection method, a manufacturing device, a manufacturing method, a quality control method, and a pipe, which inspect the inner surface of the pipe as an inspection target area.

[0002] In modern society, pipe materials are used in a variety of applications, including construction, infrastructure such as pipelines, transportation equipment, and industrial machinery, and their quality is crucial from a safety perspective. Surface defects, in particular, can lead to reduced strength and fractures, potentially resulting in the collapse of structures and the leakage of pipeline contents. Among these surface defects, concave surface defects caused by the intrusion of foreign objects during manufacturing processes such as rolling directly lead to wall thinning. Furthermore, if the defect is caused by foreign objects burned onto rolling rolls, they can occur in large numbers and be highly harmful. Inspecting pipe materials for surface defects at the raw material stage and preventing their leakage into finished products is crucial to preventing serious accidents.

[0003] For example, steel pipes made from iron may undergo a process of visually inspecting the exterior of the pipe after molding to check for surface defects before shipping. Surface defects on the outer surface of the steel pipe are easily detected by the naked eye, and inspection can be performed under conditions that make it easy to detect surface defects. In contrast, surface defects on the inner surface of the steel pipe are inspected by looking into the pipe from the outside, which makes it very difficult to inspect under conditions that make it easy to detect surface defects due to factors such as the long distance to the surface defect and the need to visually inspect the innermost part at a very shallow angle. For this reason, submerged pipe inspections, which involve entering the interior of the steel pipe for inspection, are sometimes performed. However, this method presents many challenges, including dangerous work and the inability to be applied to small-diameter steel pipes, which are difficult to inspect in the first place.

[0004] Against this background, there is a strong need for technology that can measure the condition of the inner surface of a steel pipe and automatically detect concave surface defects, and various technological developments have been made in the past. Specifically, Patent Document 1 describes a method of capturing an image of the inner surface of a steel pipe with a camera and detecting surface defects from the obtained image.

[0005] JP 2011-069616 A JP 2015-125089 A JP 2015-210150 A

[0006] "Development of a Steel Pipe Surface Inspection Device Using Twin Projection and Differential Method," 35th Sensing Forum, 1B1-4, August 2018, pp. 105-110

[0007] However, with the method described in Patent Document 1, when patterns other than concave surface defects are present on the inner surface of a steel pipe, it is difficult to distinguish between the surface defects and the patterns in the image, and there is a concern that the detection performance of surface defects may be reduced. In particular, in black steel materials with an oxide film called black scale, the oxide film adheres unevenly due to uneven cooling, resulting in strong patterns. However, since such patterns often do not pose a quality problem, they must be distinguished from surface defects, which can lead to overdetection of surface defects.

[0008] The present invention has been made in consideration of the above-mentioned problems, and its object is to provide a pipe surface inspection device and a surface inspection method that can accurately detect concave surface defects that occur on the inner surface of a pipe having a pattern on its surface with a simple device configuration. Another object of the present invention is to provide a pipe manufacturing device, manufacturing method, quality control method, and pipe that can provide high-quality pipes.

[0009] [1] A pipe surface inspection device according to the present invention is a pipe surface inspection device for inspecting the inner surface of a pipe as an inspection target area, and is equipped with an imaging unit having an imaging element having a field of view on the surface of the inner surface and two distinguishable light sources, and arranged on the axis of the pipe so that the imaging element is sandwiched between the two distinguishable light sources, a processing unit that acquires images of reflected light from each light source, generates a difference image between the two acquired images, and detects surface defects in the inspection target area from the generated difference image, and a scanning unit that scans the imaging unit in the axial direction of the pipe, wherein the two distinguishable light sources irradiate illumination light onto the same predetermined inspection target area, and the imaging element captures an image of the predetermined inspection target area.

[0010] [2] The pipe surface inspection device of the present invention is the pipe surface inspection device of [1] above, wherein the imaging pitch of the imaging element in the axial direction of the pipe is smaller than a value calculated from the inner radius of the pipe, the maximum displacement of the imaging element expected when the imaging unit is scanned in the axial direction of the pipe, the axial overlap width of the two fields of view of the imaging elements adjacent in the axial direction of the pipe, and the angle of view of the imaging element in the axial direction of the pipe.

[0011] [3] The pipe surface inspection device of the present invention is the pipe surface inspection device of [1] or [2] above, wherein the imaging element is a full-circumferential imaging element capable of obtaining an image capturing the entire circumferential area of ​​the inner surface of the pipe in a single image capture.

[0012] [4] The pipe surface inspection device of the present invention is any of the pipe surface inspection devices described above in [1] to [3], wherein the angular difference between the angle formed by the direction of specular reflection of illumination light from an upstream light source, which is the light source located upstream of the imaging element in the axial direction of the pipe, and the imaging direction of the imaging element, and the angle formed by the direction of specular reflection of illumination light from a downstream light source, which is the light source located downstream of the imaging element in the axial direction of the pipe, and the imaging direction of the imaging element is greater than a value calculated from the inner radius of the pipe, the maximum displacement amount of the imaging element expected when the imaging unit is scanned in the axial direction of the pipe, the angle of view of the imaging element in the axial direction of the pipe, and the distance from the imaging element to the upstream light source or the downstream light source.

[0013] [5] The pipe surface inspection device according to the present invention is any one of the pipe surface inspection devices according to [1] to [4] above, wherein the two distinguishable light sources are annular illumination.

[0014] [6] The pipe surface inspection device according to the present invention is any one of the pipe surface inspection devices described above in [1] to [5], wherein the scanning unit scans the imaging element along the central axis of the pipe.

[0015] [7] A pipe surface inspection method according to the present invention is a surface inspection method for inspecting the inner surface of a pipe as an inspection target area, and includes a processing step of acquiring images of reflected light from each light source using an imaging unit having an imaging element with a field of view on the surface of the inner surface and two distinguishable light sources, and arranged on the axis of the pipe so that the imaging element is sandwiched between the two distinguishable light sources, generating a difference image between the two acquired images, and detecting surface defects in the inspection target area from the generated difference image, and a scanning step of scanning the imaging unit in the axial direction of the pipe, wherein the processing step includes a step of irradiating the same predetermined inspection target area with illumination light using the two distinguishable light sources, and capturing an image of the predetermined inspection target area using the imaging element.

[0016] [8] The method for inspecting the surface of a pipe according to the present invention is the method for inspecting the surface of a pipe according to [7] above, wherein the imaging pitch of the imaging element in the axial direction of the pipe is smaller than a value calculated from the inner radius of the pipe, the maximum displacement of the imaging element expected when the imaging unit is scanned in the axial direction of the pipe, the axial overlap width of the two fields of view of the imaging elements adjacent in the axial direction of the pipe, and the angle of view of the imaging element in the axial direction of the pipe.

[0017] [9] The pipe surface inspection method of the present invention is the pipe surface inspection method of [7] or [8] above, wherein the imaging element is a full-circumferential imaging element capable of obtaining an image capturing the entire circumferential area of ​​the inner surface of the pipe in a single image capture.

[0018]

[10] A tube manufacturing apparatus according to the present invention includes the tube surface inspection apparatus according to any one of [1] to [6] above.

[0019]

[11] A method for manufacturing a tube according to the present invention inspects the inner surface of a tube using any one of the tube surface inspection methods [7] to [9] above, and manufactures the tube based on the inspection results.

[0020]

[12] The quality control method of the present invention inspects the inner surface of a pipe using any one of the pipe surface inspection methods [7] to [9] above, and controls the quality of the pipe based on the inspection results.

[0021]

[13] The quality of the pipe according to the present invention is guaranteed by any one of the pipe surface inspection methods [7] to [9] above.

[0022] The pipe surface inspection device and method according to the present invention can accurately detect concave surface defects that occur on the inner surface of a pipe that has a pattern on its surface with a simple device configuration. Furthermore, the pipe manufacturing device, manufacturing method, quality control method, and pipe according to the present invention can provide high-quality pipes.

[0023] FIG. 1 is a schematic diagram showing the overall configuration of a surface inspection apparatus according to one embodiment of the present invention. FIG. 2 is a schematic diagram showing the configuration of a modified scanning unit shown in FIG. 1. FIG. 3 is a schematic diagram for explaining the configuration of the imaging unit shown in FIG. 1. FIG. 4 is a diagram showing the configuration of a downstream optical system, an internal image, and a differential image obtained from the internal image. FIG. 5 is a schematic diagram showing the configuration of an upstream light source and a downstream light source. FIG. 6 is a schematic diagram showing the configuration of an omnidirectional camera. FIG. 7 is a schematic diagram showing the configuration of a modified omnidirectional camera. FIG. 8 is a schematic diagram showing the configuration of a modified omnidirectional camera. FIG. 9-1 is a schematic diagram for explaining the configuration of an omnidirectional camera. FIG. 9-2 is a schematic diagram for explaining the configuration of an omnidirectional camera. FIG. 10 is a flowchart showing the flow of image processing according to one embodiment of the present invention. FIG. 11 is a diagram for explaining the flow of image processing shown in FIG. 10. FIG. 12 is a diagram showing the light and dark patterns of concave surface defects and convex surface defects, where FIG. 12(a) shows the light and dark pattern of the concave surface defect and FIG. 12(b) shows the light and dark pattern of the convex surface defect. Fig. 13 is a diagram showing an example of a concave surface defect detected by the image processing shown in Fig. 10. Fig. 14 is a schematic diagram for explaining the configuration of a scanning unit. Fig. 15 is a schematic diagram for explaining the configuration of a scanning unit. Fig. 16 is a diagram showing an example of a defect map. Fig. 17 is a schematic diagram showing an example of the configuration of a marking device. Fig. 18 is a schematic diagram showing an example of the configuration of a marking device.

[0024] The configuration and operation of a surface inspection device according to one embodiment of the present invention will be described below with reference to the drawings.

[0025] [Overall Configuration] First, with reference to FIG. 1, the overall configuration of a surface inspection device according to an embodiment of the present invention will be described.

[0026] Fig. 1 is a schematic diagram showing the overall configuration of a surface inspection device according to one embodiment of the present invention. As shown in Fig. 1, the surface inspection device 1 according to one embodiment of the present invention is a device that inspects the entire circumference of the inner surface of a steel pipe P by scanning the inspection target area, which is the inner surface of the steel pipe P, along the axial direction of the steel pipe P, and is equipped with an imaging unit 2, a scanning unit 3, and a processing unit 4.

[0027] The imaging unit 2 includes an imaging element (hereinafter referred to as an omnidirectional camera) 21 capable of imaging the entire circumferential area (i.e., all directions) of the inner surface of the steel pipe P in one go, an upstream light source 22 provided upstream in the scanning direction, and a downstream light source 23 provided downstream in the scanning direction. The omnidirectional camera 21 corresponds to the imaging element according to the present invention, particularly the omnidirectional imaging element. The omnidirectional camera 21 is disposed on the axis of the steel pipe P between the upstream light source 22 and the downstream light source 23. The omnidirectional camera 21 captures two inner surface images by separately capturing the reflected light of illumination light irradiated onto the inner surface of the steel pipe P from the upstream light source 22 and the downstream light source 23 for each scanning position, and inputs the data of the two inner surface images to the processing unit 4 via an electric communication line.

[0028] Note that this embodiment uses an all-around camera 21 that can image the entire circumferential surface of the inner surface of the steel pipe P in all directions in the circumferential direction in one go, but the present invention is not limited to this embodiment. The imaging element may image a partial circumferential area of ​​the inner surface of the steel pipe P in one go. Also, multiple imaging elements may be provided that image a partial circumferential area of ​​the inner surface of the steel pipe P in one go. In this case, depending on the condition of the steel pipe P, only a partial circumferential area of ​​the inner surface of the steel pipe P may be used as the inspection target region. Also, in this case, if a partial circumferential area is imaged in one image capture, and the image capture is repeated multiple times while changing the imaging positions of the multiple imaging elements along the circumferential direction, the same effect as the all-around camera 21 can be obtained. On the other hand, considering that only one image capture is required, resulting in high work efficiency, that there is no need to combine the obtained inner surface images, that there is no need to align the inner surface images in the circumferential direction, and that the entire circumferential area of ​​the inner surface of the steel pipe P can be captured without any omissions, it is most preferable to use a circumferential imaging element that can capture the entire circumferential area of ​​the inner surface of the steel pipe P in all directions in one capture.

[0029] The scanning unit 3 is configured as a cart and is equipped with the imaging unit 2. The scanning unit 3 scans the position of the imaging unit 2 along the axial direction of the steel pipe P by traveling along the inner surface of the steel pipe P along the axial direction of the steel pipe P. An encoder is installed on the axle of the cart constituting the scanning unit 3, and the position of the imaging unit 2 in the axial direction of the steel pipe P (scanning position) can be identified by using the encoder to count the number of rotations of the cart's wheels. It is preferable that the cart be configured so as not to obstruct the field of view of the imaging unit 2 or the illumination light. Furthermore, if the processing unit 4 and battery are lightweight, they may be mounted on the scanning unit 3 together with the imaging unit 2. Furthermore, the cart may be powered manually or mechanically. Note that the scanning unit 3 may not be a cart, but may have a configuration as shown in FIG. 2. In the configuration shown in FIG. 2, the scanning unit 3 includes a rod-shaped member 3a extending in the axial direction of the steel pipe P and having a 360° camera 21, an upstream light source 22, and a downstream light source 23 attached to its tip, and a scanning mechanism 3b that moves the rod-shaped member 3a along the axial direction of the steel pipe P. In this configuration, it is desirable that the rod-shaped member 3a has sufficient rigidity to support the imaging unit 2, and if necessary, wheels may be attached to the tip to support the imaging unit 2.

[0030] The processing unit 4 is configured with an information processing device such as a computer, and is connected to the 360-degree camera 21 via a telecommunications line. The processing unit 4 detects concave surface defects within the imaging area of ​​the inner surface images by performing image processing on the two inner surface images input from the imaging unit 2, and outputs information about the detected concave surface defects (e.g., circumferential and axial positions) as inspection results. As an image processing method, a method can be applied in which processing is performed in the order of shading correction, difference processing, defect candidate detection using light and dark patterns, and harmfulness determination using machine learning, and ultimately harmful concave surface defects are detected (see, for example, Patent Documents 2 and 3). Specific examples of image processing will be described later with reference to FIGS. 10 to 13.

[0031] [Configuration of Imaging Unit] Next, the configuration of the imaging unit 2 will be described in detail with reference to FIGS.

[0032] FIG. 3 is a schematic diagram illustrating the configuration of the imaging unit 2 shown in FIG. 1 . As shown in FIG. 3 , in this embodiment, the omnidirectional camera 21 is disposed on the central axis of the steel pipe P so as to view the inner surface of the steel pipe P from the normal direction. An upstream light source 22 and a downstream light source 23 are disposed upstream and downstream of the scanning direction of the omnidirectional camera 21, respectively. The upstream light source 22 and the downstream light source 23 irradiate the inspection target area within the field of view V of the omnidirectional camera 21 with illumination light. As described in Non-Patent Document 1, by making the projection and reception angles of the upstream optical system and the downstream optical system as similar as possible, the effect of removing pattern signals due to the difference between the two inner surface images can be enhanced. Hereinafter, the optical system realized by the upstream light source 22 and the omnidirectional camera 21 will be referred to as the upstream optical system, and the optical system realized by the downstream light source 23 and the omnidirectional camera 21 will be referred to as the downstream optical system.

[0033] Figure 4(a) shows the configuration of the downstream optical system, Figure 4(b) shows an internal surface image captured by the optical system shown in Figure 4(a), and Figure 4(c) shows a difference image obtained from the internal surface image shown in Figure 4(b). Note that the symbol Pa in Figure 4(a) indicates the internal surface of the steel pipe P. As shown in Figure 4(c), in the central part of the difference image, the projection and reception angles of the upstream optical system and the downstream optical system are equal, resulting in optically equivalent conditions. Therefore, the pattern disappears and noise is reduced by the difference. In contrast, in the right part of the difference image, the illumination light from the upstream light source 22 approaches specular reflection, while the illumination light from the downstream light source 23 approaches diffuse backward reflection. Therefore, the pattern remains even after the difference and the noise is large.

[0034] In order to make the projection and reception angles of the upstream optical system and the downstream optical system the same, it is preferable that the area to be inspected is near a plane that passes through the 360° camera 21 and is perpendicular to the axial direction of the steel pipe P, and that the axial position of the 360° camera 21 is the central position of both light sources. It is also preferable to inspect with the same resolution and projection and reception angles for each circumferential position on the inner surface of the steel pipe P. For this reason, it is preferable that the central positions of the 360° camera 21 and the light source in the plane perpendicular to the axial direction of the steel pipe P are the same as the central position of the cross section of the steel pipe P.

[0035] Furthermore, there is an optimal projection angle for enhancing the signal of concave surface defects and eliminating the signal of patterns. For example, in the case of black steel, the projection angle is approximately 70 degrees. If the inner diameter of the steel pipe P is large, attempting to irradiate at a low angle using a conventional spot light source requires increasing the distance between the light source and the 360-degree camera 21, which makes the device itself large and difficult to handle, and is also disadvantageous in terms of cost. Therefore, as shown in Figure 5, ring lights (also called annular lights) are used as the upstream light source 22 and the downstream light source 23. This allows for low-angle illumination while keeping the distance between the 360-degree camera 21 and the light source short, and allows the imaging unit 2 to be simple with fewer parts.

[0036] Next, an embodiment will be described in which the reflected light of illumination light from the upstream light source 22 and the reflected light of illumination light from the downstream light source 23 are separately captured. As described in Patent Document 3, if there is a misalignment in the capturing positions of the two inner surface images when generating a differential image, the pattern signal cannot be removed and remains in the differential image as noise. On the other hand, in a method in which the imaging unit 2 is stopped to capture two inner surface images at the same position, repeated stopping and scanning takes time to inspect the entire length of the steel pipe P. Therefore, to efficiently inspect the inner surface of the steel pipe P along its axial direction, illumination light is irradiated while the imaging unit 2 is scanned along the axial direction of the steel pipe P to capture the inner surface images.

[0037] Therefore, one possible method is to reduce the misalignment in the imaging positions by shortening the time difference between imaging using the upstream light source 22 and imaging using the downstream light source 23 sufficiently relative to the moving speed of the imaging unit 2 and instantly obtaining two internal surface images. Another promising method is to separate the irradiated light by changing the color of the illumination light used for imaging using the upstream light source 22 and imaging using the downstream light source 23 and employing a color camera that can receive only the reflected light of each color in different channels. The former method requires a special device that captures two internal surface images at high speed and controls the light emission of the light source. In contrast, the color camera and colored light source used in the latter method are commonly used for general purposes, so it can be realized with a less expensive and simpler system than the former method. Furthermore, the former method has the advantage of being resistant to image blur because the irradiation time can be very short.

[0038] When using a color camera or colored light sources, it is preferable to prevent interference between their wavelength sensitivity bands. For example, consider a case where channel 1 and channel 2 of a color camera are used to receive illumination light from colored light source 1 and colored light source 2, which have different wavelengths. Here, channel 1 and channel 2 have different wavelength sensitivity bands, with the wavelength sensitivity band of channel 1 coinciding with the wavelength band of colored light source 1 and the wavelength sensitivity band of channel 2 roughly coinciding with the wavelength band of colored light source 2. The illumination light from colored light source 1 is received only by channel 1, and the illumination light from colored light source 2 is received only by channel 2. In this case, it is preferable that the wavelength band of colored light source 1 does not include the wavelength band to which channel 2 is sensitive. It is also preferable that the wavelength band of colored light source 2 does not include the wavelength band to which channel 1 is sensitive. Therefore, when using a general color camera, it is most preferable to use channels with wavelength sensitivity bands that are far apart, particularly the red and blue channels, and select red and blue for colored light sources 1 and 2, respectively. Furthermore, if the surface of the inspection area is covered with a coating or layer of a specific color, it is preferable to use a wavelength band that excludes that specific color. For example, in the case of steel material covered with a red scale, the red scale may strongly reflect the red illumination light and cause noise, so it is preferable to select blue and green illumination light even if it causes some interference. Here, green refers to the wavelength band in the wavelength range of 490 to 560 nm, blue refers to the wavelength band in the wavelength range of 430 to 490 nm, and red refers to the wavelength band in the wavelength range of 600 to 760 nm.

[0039] Next, the configuration of the omnidirectional camera 21 will be described. When using the omnidirectional camera 21 to image the inner surface of the steel pipe P, it is desirable to hold the omnidirectional camera 21 so that structures do not interfere with the field of view of the omnidirectional camera 21, while also imaging the entire circumferential area of ​​the inner surface of the steel pipe P at a predetermined position in the axial direction of the steel pipe P as thoroughly as possible. For this reason, as shown in FIGS. 6( a) and 6(b), multiple cameras 21a and lenses 21b for imaging the inner surface of the steel pipe P may be installed circumferentially around a support 24, and the fields of view of adjacent cameras 21a in the circumferential direction may be overlapped to image the entire inner surface of the steel pipe P. In this case, communication and power supply are possible by passing wiring inside the support 24. Furthermore, in this case, one set of cameras 21a and lenses 21b has a field of view of an area extending 90° or more in the circumferential direction of the inner surface of the steel pipe P, and four sets of cameras 21a and lenses 21b are configured to be able to image the entire circumferential area in one go. As shown in FIG. 7, the cameras 21a and the lenses 21b may be arranged alternately in the circumferential direction of the support 24 to avoid interference, thereby saving space.

[0040] In the examples shown in Figures 6 and 7, four pairs of cameras 21a and lenses 21b are installed, but any number of pairs of cameras 21a and lenses 21b can be installed as long as the entire circumference of the inner surface of the steel pipe P can be imaged. Wide-angle cameras capable of 180-degree imaging may be installed on both sides of a substrate to image the entire circumference of the inner surface of the steel pipe P. In this case, to avoid a decrease in defect detection performance due to distortion or aberration at the edge of the field of view, as shown in Figure 8, two substrates with wide-angle cameras 21c installed on both sides may be installed orthogonally on the structure, and the entire circumference of the inner surface of the steel pipe P may be imaged with four wide-angle cameras 21c. In this case, communication and power supply can be enabled by running wiring within the structure. Even without using a camera capable of omnidirectional imaging, the entire circumference of the inner surface of the steel pipe P may be imaged by scanning the camera multiple times within the steel pipe while changing the circumferential field of view and combining the images and detection results obtained from each scan.

[0041] Furthermore, in order to accurately capture the entire axial area of ​​the inner surface of the steel pipe P at a predetermined circumferential position, it is preferable to set several parameters, such as the size of the field of view V of the omnidirectional camera 21 and the axial overlap width h between the obtained axially adjacent inner surface images. Specific setting examples are described below using Figures 9-1 and 9-2. In Figures 9-1 and 9-2, the left-right direction indicates the horizontal direction and the axial direction of the steel pipe P, and the up-down direction indicates the vertical direction and the radial direction of the steel pipe P. The axis (pipe axis) of the steel pipe P is represented by β, and the central axis connecting the central positions of the upstream light source 22, the omnidirectional camera 21, and the downstream light source 23 is represented by γ. The central axis γ is also represented by the central axis γ of the imaging unit 2. In Figure 9-1(a), the central axis γ coincides with the pipe axis β, and the imaging direction ε of the omnidirectional camera 21 is aligned with the radial direction of the steel pipe P. In this state, the omnidirectional camera 21 captures an image of the inner surface of the steel pipe P along the circumferential direction at a predetermined position on the pipe axis β. The field of view of the omnidirectional camera 21 on the inner surface of the steel pipe P obtained at a predetermined position on the pipe axis β is defined as V. The field of view V of the omnidirectional camera 21 is assumed to be obtained at an angle of view φ in the axial direction of the steel pipe P. There are two possible angles of view, the axial direction and the circumferential direction of the steel pipe P, but the angle of view used in the following discussion is only the angle of view in the axial direction of the steel pipe P.

[0042] As shown in Figure 9-1(a), the inner radius of the steel pipe P is R, the distance from the omnidirectional camera 21 to the upstream light source 22 or the downstream light source 23 is L, and the imaging pitch of the omnidirectional camera 21 in the tube axis direction is d. In Figure 9-1(a), the field of view of the omnidirectional camera 21 obtained at a predetermined position on the tube axis β is V0, and the field of view of the omnidirectional camera 21 when the omnidirectional camera 21 moves leftward along the tube axis β by the imaging pitch d is V1. The fields of view V0 and V1 are adjacent with an overlap width h. The width of the fields of view V, V0, and V1 of the omnidirectional camera 21 in the tube axis direction is H. As shown in Figure 9-1(b), the maximum displacement amount, which is the maximum expected deviation of the omnidirectional camera 21 from the tube axis β when the imaging unit 2 scans in the tube axis direction, is ΔR, and the maximum tilt amount of the central axis γ in the vertical direction with respect to the tube axis β is Δθ. Here, the maximum displacement amount ΔR can also be said to be a value obtained by converting into distance the maximum amount of deviation of the omnidirectional camera 21 from the tube axis β, which is predetermined when the imaging unit 2 is scanned in the tube axis direction. As shown in FIG. 9-1(a), in an ideal state where the omnidirectional camera 21 is positioned on the tube axis β, the field of view width H of the omnidirectional camera 21 falls within the range shown in the following formula (1). Note that, in deriving formula (1), an approximation was used in which Δθ is sufficiently smaller than 1 rad so that its influence can be ignored, in order to simplify the calculation.

[0043]

[0044] If the field of view width H of the omnidirectional camera 21 is reduced, the field of view will become even smaller if the position of the omnidirectional camera 21 changes during scanning and approaches the inner surface of the steel pipe P, leading to missing fields of view. Furthermore, the axial overlap width h of the inner surface images of two adjacent fields of view V0 and V1 obtained by capturing images before and after the scan is preferably set so that the target surface defect is not separated in the image. Therefore, in order to ensure the axial overlap width h even when the omnidirectional camera 21 is displaced by the maximum displacement amount ΔR to minimize the field of view, the imaging pitch d preferably satisfies the following formula (2). That is, the imaging pitch d of the omnidirectional camera 21 in the pipe axial direction is preferably set smaller than a value calculated from the inner radius R of the steel pipe P, the maximum displacement amount ΔR of the imaging element expected when the imaging unit 2 scans the steel pipe P in the axial direction, the axial overlap width h of the two fields of view V of the omnidirectional camera 21 adjacent in the pipe axial direction, and the axial angle of view φ of the imaging element.

[0045]

[0046] Furthermore, when the 360-degree camera 21 is tilted as shown in Figure 9-1(b), the difference in the light projection and reception angles is greatest at the axial end vb of the steel pipe P in the field of view V. When the maximum tilt amount Δθ = 0, the difference in the light projection and reception angles at end va and end vb is approximately the same. However, when the tilt direction relative to the axial direction is reversed, i.e., when tilted downward to the right in Figure 9-1(b), the difference in the light projection and reception angles is greatest at end va. An indicator of the difference in the light projection and reception angles is the angle difference α between the specular reflection direction of the illumination light from the light source and the imaging direction ε of the imaging element. It is preferable to set the 360-degree camera 21 so that this angle difference α is as small as possible between the upstream optical system and the downstream optical system by tilting it. In Figures 9-1(a) and 9-1(b), the symbol vc indicates the midpoint between end va and end vb.

[0047] Here, the angular difference α at the position of the end vb will be described in detail. First, as shown in FIG. 9-2(c), the specular reflection direction of the illumination light irradiated from the upstream light source 22 toward the position of the end vb (a direction symmetrical with the surface normal ζ at the position of the end vb) is defined as η1. In this case, the angle formed by the specular reflection direction η1 and εb, which is the direction of the omnidirectional camera 21 as viewed from the position of the end vb, is defined as α1. Similarly, as shown in FIG. 9-2(d), the specular reflection direction of the illumination light irradiated from the downstream light source 23 toward the position of the end vb (a direction symmetrical with the surface normal ζ at the position of the end vb) is defined as η2. In this case, the angle formed by the specular reflection direction η2 and εb, which is the direction of the omnidirectional camera 21 as viewed from the position of the end vb, is defined as α2. Using these angles α1 and α2, the angle difference between the light projection and reception angle α is defined as |α1 - α2|.

[0048] For this reason, it is preferable to perform inspection so that the angular difference is equal to or less than the allowable angular difference at the position of the end vb where the angular difference is largest so that the signal can be sufficiently removed after subtraction, and it is preferable that the angular difference α satisfy the following mathematical formula (3): The angular difference α between the angle α1 formed by the direction of specular reflection η1 of the illumination light from the upstream light source 22 and the imaging direction εb of the image sensor and the angle α2 formed by the direction of specular reflection η2 of the illumination light from the downstream light source 23 and the imaging direction εb of the image sensor is preferably set to be greater than a value calculated from the inner radius R of the steel pipe P, the predetermined maximum displacement ΔR of the image sensor when the image sensor 2 scans in the tube axial direction, the angle of view φ of the image sensor in the axial direction, and the distance L from the image sensor to the upstream light source 22 or the downstream light source 23.

[0049]

[0050] In addition, as the position of the omnidirectional camera 21 shifts circumferentially from the pipe axis β, the resolution decreases when imaging the inner surface in the direction opposite to the shift. Therefore, it is preferable that the resolution be equal to or greater than the minimum resolution per pixel required by the inspection object at all positions within the field of view of the omnidirectional camera 21. For this reason, if the resolution per pixel when the position and / or tilt of the omnidirectional camera 21 does not change with respect to the pipe axis β is r and the allowable minimum resolution per pixel is r', it is preferable that the following mathematical formula (4) be satisfied. In other words, it is preferable that the resolution per pixel r of the imaging element be smaller but higher than the value calculated from the inner radius R of the steel pipe P, the maximum displacement ΔR of the imaging element expected when the imaging unit 2 scans in the axial direction of the steel pipe P, and the minimum resolution per pixel r' required by the inspection object. For example, if the required minimum resolution r' per pixel is 0.6 mm / pixel, the inner radius R of the steel pipe P is 50 mm, and the maximum displacement ΔR is 10 mm, it is preferable to set the resolution per pixel to 0.5 mm / pixel or less.

[0051]

[0052] It is preferable not to make the field of view width H of the omnidirectional camera 21 too large. If the axial angle of view φ of the omnidirectional camera 21 is increased to increase the field of view width H of the omnidirectional camera 21, the effect of the difference tends to decrease at the axial ends va, vb of the field of view V of the omnidirectional camera 21. This is because the difference between the projection and reception angles of the upstream optical system and the downstream optical system increases as the distance from the irradiation position set on the inner surface of the steel pipe P increases, as at the axial ends va, vb of the field of view V. The decrease in the effect of the difference mentioned here means that the signal of a harmless flat surface pattern cannot be sufficiently reduced.

[0053] Furthermore, it is preferable that the depth of field D satisfies the following formula (5) in order to perform inspection under conditions where the image is in focus even if the position and / or angle of the omnidirectional camera 21 changes. In other words, it is preferable that the depth of field D of the imaging element be smaller than the value calculated from the maximum displacement ΔR of the imaging element expected when the imaging unit 2 scans the steel pipe P in the axial direction.

[0054]

[0055] Based on the above considerations, it is desirable for the central axis γ of the imaging unit 2 to stably scan along the axis β of the steel pipe P so that the maximum displacement ΔR and maximum tilt Δθ of the omnidirectional camera 21 during scanning are as small as possible. Furthermore, by considering the optical system while taking into account changes in the position and tilt of the omnidirectional camera 21, it is possible to accurately inspect the axial area of ​​a predetermined circumferential position on the inner surface of the steel pipe P as closely as possible. Note that, in order to perform the above-described inspection, optical adjustments are made in advance to obtain images free of saturation and out-of-focus. Furthermore, the above considerations were made using the omnidirectional camera 21 as an example, and the fields of view V, V0, and V1 were considered to have an axial width H of the steel pipe P and a circumferential width of 360°. However, this is not limiting. The circumferential width may be less than 360° and greater than 0°. Similar effects can be obtained even with, for example, 45°, 30°, 5°, or 1°.

[0056] [Image Processing] Finally, a specific example of image processing executed by the processing unit 4 will be described with reference to FIGS.

[0057] 10 is a flowchart showing the flow of image processing according to one embodiment of the present invention. In this embodiment, a 360-degree camera having fisheye cameras on both sides of the substrate is used as the omnidirectional camera 21, a ring light that emits green illumination light is used as the upstream light source 22, and a ring light that emits blue illumination light is used as the downstream light source 23. As shown in FIG. 10 , in the image processing according to this embodiment, first, the processing unit 4 divides the channels of the color image captured by the omnidirectional camera 21 (step S1), and then extracts images of the same inspection range (see FIG. 11( a)) from the obtained blue channel image and green channel image (steps S2a and S2b).

[0058] Next, the processing unit 4 performs shading correction on the cut-out blue channel image and green channel image (steps S3a and S3b, see FIG. 11(b)), and then generates a difference image between the blue channel image and the green channel image (step S4, see FIG. 11(c)). Note that the order of channel division and cut-out of the image of the inspection range, and the order of cut-out of the image of the inspection range and shading correction may be reversed. Cutting out the image of the inspection range first has the advantage of reducing the amount of calculation, and performing shading correction before cutting out the image of the inspection range has the advantage of being less susceptible to the influence of non-stationary parts that appear at the edges of the image during calculation.

[0059] Next, the processing unit 4 extracts bright and dark areas from the difference image by binarizing the difference image using a predetermined brightness threshold (step S5), and then performs an expansion / contraction process toward dark areas for pixels corresponding to bright areas and toward bright areas for pixels corresponding to dark areas (step S6). The processing unit 4 then extracts bright and dark patterns indicating concave surface defects as defect candidates from the binarized image after the expansion / contraction process (step S7). As shown in Figures 12(a) and 12(b), the bright and dark patterns for concave surface defects are reversed from those for convex surface defects, so that only the bright and dark patterns indicating concave surface defects can be efficiently extracted from the difference image. The bright and dark patterns may also be extracted by comparing the positions of the bright and dark areas after the binarization process.

[0060] Finally, the processing unit 4 uses machine learning to determine whether the defect candidate portion is a defect and the defect type and grade. The determination rules may be manually defined or automatically generated using a common machine learning method using feature quantities, such as a regression model (linear regression, logistic regression, multiple regression, support vector machine, nonlinear kernel, etc.), a decision tree model, a random forest, a Bayesian estimation model, a Gaussian mixture model, or a boosted version of these models. Multiple machine learning methods may also be combined. Figure 13 shows an example of a concave surface defect detected by the above process. This is an example of a defect called a slag indentation defect, shown in Figure 13(a), on the inner surface of a steel pipe. Figure 13(b) shows the image processing process. In the image processing shown in Figure 13(b), the color image channels are divided in step S1, and blue and green raw images are obtained by extracting the inspection range in steps S2a and S2b. Then, blue-corrected and green-corrected images are obtained by shading correction in steps S3a and S3b, and a difference image is calculated using both images in difference processing in step S4. The difference image shown in Figure 13(c) clearly shows the slug defects as a light-dark pattern. Figure 13(e) is a graph of the horizontal intensity change relative to the approximate vertical center position in Figure 13(c). As shown in Figure 13(e), defective areas appear as stronger signals than non-defective areas. The light-dark pattern is then extracted from the difference image through steps S5 to S7. The extracted light-dark pattern is shown as the detection result in the binary image of Figure 13(d). The detection result in the binary image shown in Figure 13(d) has light and dark areas at both ends. Feature values ​​are calculated from each, and the judgment process in step S8 determines whether or not the defect is actually harmful. Information about the detected surface defects is then output, providing guidance to the operator to prevent the release of surface defects. Information about the detected surface defects can also be sent to a server device for collection and use in determining whether or not the product can be shipped or whether maintenance is required. Although a color camera and color illumination are used in this embodiment, the same effect can be obtained by combining a camera capable of high-speed imaging with pulsed illumination to obtain two images of the inner surface.

[0061] [Configuration of Scanning Unit] Next, the configuration of the scanning unit 3 will be described. As described above, the scanning unit 3 scans the imaging unit 2 along the axial direction of the steel pipe P so as to minimize alignment variations, such as changes in the circumferential and axial positions and inclination of the imaging unit 2. While both manual and automatic scanning methods are possible, automatic operation is preferably controlled to minimize alignment variations. For example, when the scanning unit 3 is self-propelled, such as with a wheel-type or crawler-type traveling mechanism, it is preferable to scan a fixed circumferential position on the inner surface of the steel pipe P parallel to the axial direction. However, as shown in Figures 14(a) and 14(b), there is a concern that the position of the scanning mechanism 3b may fluctuate in the circumferential and axial directions. Therefore, it is preferable to use an acceleration sensor 31 to detect the vertical downward direction and control the scanning mechanism 3a so that the vertical downward direction is always constant. This enables the scanning unit 3 to self-propel stably.

[0062] 15(a) and 15(b), the camera 32 may be installed in the scanning mechanism 3b so that its optical axis coincides with the axial direction of the steel pipe P during normal running, and the camera 32 may run while capturing an image of the pipe end. In this case, if the running direction deviates from the axial direction of the steel pipe P, the center position of the pipe end in the image captured by the camera 32 will change. Therefore, by controlling the scanning mechanism 3b so that the center position of the pipe end in the captured image is always constant, stable scanning parallel to the axial direction is possible. Instead of the center position of the pipe end, a target may be installed outside the pipe, and deviations in the running direction may be detected or the scanning mechanism 3b may be controlled based on the position of the target in the captured image.

[0063] [Mapping] A defect map may be created based on the obtained defect position, image, defect type, and grade information. A defect map clearly visualizes defect occurrence information and shape, and is useful for quality control of the pipe inner surface, such as providing guidance to operators. An example of a defect map is shown in Figure 16. In the defect map shown in Figure 16, the circumferential and axial directions of the steel pipe P are expanded into a rectangle, and the defect occurrence positions are clearly indicated within it, making it possible to grasp defect occurrence information. In addition, coordinates, defect type, grade, image, etc. may be written at the defect occurrence positions. The reference position in the circumferential direction is preferably a position that serves as a landmark in the vertical downward direction or in the axial direction of the steel pipe P. For example, in the case of a welded steel pipe, using the weld bead position as the reference position makes it easier to identify the defect occurrence position.

[0064] [Marking] When a defect is detected while scanning the inner surface of the steel pipe P, the occurrence of the defect may be visualized by marking the defective portion. In this case, for example, as shown in FIG. 17, a marking device 33 may be provided in addition to the scanning mechanism 3b, or, for example, as shown in FIG. 18, a scanning mechanism 3b1 may be provided in addition to the outside of the steel pipe P, which is capable of scanning the marking device 33 in the axial or circumferential direction. Furthermore, the defective portion may be marked using information on the defect map obtained after the inner surface inspection process. Marking the defective portion makes it possible to reliably grasp the defect position, thereby reducing the risk of the defect being released into the market.

[0065] [Secondary Inspection] In the present invention, it is difficult to obtain accurate information on the depth of defects. Therefore, the defect locations obtained by the present invention may be precisely re-inspected using another defect detection method. Re-inspection methods include a sensory inspection method in which a person inspects the pipe visually, a three-dimensional shape measurement method using a laser, or a magnetic measurement method such as magnetic flux leakage or eddy current flaw detection. Applying these methods alone to the entire inner surface of the steel pipe P requires a significant amount of inspection time, thereby reducing inspection efficiency. However, by limiting inspection to defects detected by the present invention, precise inspection can be performed without reducing inspection efficiency. Furthermore, the re-inspection location may be identified using the defect map described above or by detecting the markings described above.

[0066] [Maintenance] Detected defects on the inner surface of the steel pipe P may be repaired using a grinder or the like. In this case, the repair may be performed by an operator entering the steel pipe P, or by a self-propelled inner surface repair mechanism. The repair position is preferably determined based on the defect position on the defect map, the marking position, the position determined to require repair based on the results of the secondary inspection, or a combination of these pieces of information. Note that a secondary inspection may be performed again after the repair.

[0067] While this embodiment applies the present invention to steel pipes, it goes without saying that the present invention can be applied to tubular products other than steel pipes. According to this embodiment, by scanning the inspection range, which is the field of view of the imaging element, parallel to the axial direction of the steel pipe P while maintaining a certain width in the axial direction of the steel pipe P, even if the sensor head tilts due to vibration or the like and the field of view changes, by pre-setting an overlap width between each field of view, it is possible to inspect the inner surface of the steel pipe P in the axial direction as accurately as possible. Furthermore, if the imaging element is a circumferential imaging element (i.e., a circumferential camera 21), an image capturing the entire circumferential area of ​​the inner surface of the steel pipe P can be obtained in a single image capture. This allows for high work efficiency by capturing only one image capture, eliminates the need for processing to combine the obtained inner surface images, eliminates the need for circumferential image alignment, and enables the circumferential direction of the inner surface of the steel pipe P to be captured without any gaps. Furthermore, by using annular illumination as illumination, compact, low-angle illumination can be achieved even if the pipe diameter is large, improving the detection performance of concave surface defects. Furthermore, the omnidirectional camera 21 and the light source are small devices that consume little power, which provides the advantage of excellent portability of the sensor head.

[0068] The surface inspection device according to an embodiment of the present invention may also be applied to a pipe manufacturing apparatus. In this case, the inner surface of a pipe is inspected using the surface inspection device according to an embodiment of the present invention, and the pipe is manufactured based on the inspection results. The surface inspection method according to an embodiment of the present invention may also be applied to a pipe manufacturing method. In this case, the inner surface of a pipe is inspected using the surface inspection method according to an embodiment of the present invention, and the pipe is manufactured based on the inspection results. The surface inspection method according to an embodiment of the present invention may also be applied to a pipe quality control method. In this case, the inner surface of a pipe is inspected using the surface inspection method according to an embodiment of the present invention, and the quality of the pipe is controlled based on the inspection results. Furthermore, a pipe whose quality is guaranteed by the surface inspection method according to an embodiment of the present invention may also be obtained. Here, "guaranteed quality" means, in other words, "information is attached that the pipe satisfies predetermined conditions for defects detected by the surface inspection method." Alternatively, it may mean "information is attached that the pipe is an acceptable product." Specifically, this means that the inspection results obtained by the surface inspection method according to an embodiment of the present invention are described in an inspection certificate (e.g., a mill sheet) attached to the pipe. In this case, for example, if predetermined conditions are met for detected defects, the pipe is treated as an acceptable product. Preferably, a pipe that does not have any defects with a width of 2 mm or more and a depth of 0.2 mm or more is treated as an acceptable material.

[0069] Although the present invention has been described above as an embodiment, the present invention is not limited to the descriptions and drawings that form part of the disclosure of the present invention. In other words, other embodiments, examples, and operational techniques that can be made by those skilled in the art based on the present invention are all included in the scope of the present invention.

[0070] According to the present invention, it is possible to provide a pipe surface inspection device and a surface inspection method that can accurately detect concave surface defects that occur on the inner surface of a pipe having a pattern on the surface with a simple device configuration. Furthermore, according to the present invention, it is possible to provide a pipe manufacturing device, a manufacturing method, a quality control method, and a pipe that can provide high-quality pipes.

[0071] REFERENCE SIGNS LIST 1 Surface inspection device 2 Imaging unit 3 Scanning unit 3a Rod-shaped member 3b, 3b1 Scanning mechanism 4 Processing unit 21 All-around camera 21a Camera 21b Lens 21c Wide-angle camera 22 Upstream light source 23 Downstream light source 24 Support 31 Acceleration sensor 32 Camera 33 Marking device P Steel pipe

Claims

1. A pipe surface inspection device for inspecting the inner surface of a pipe as an inspection target area, comprising: an imaging unit having an imaging element having a field of view on the surface of the inner surface and two distinguishable light sources, the imaging element being positioned on the axis of the pipe so that the two distinguishable light sources sandwich the imaging element; a processing unit which acquires images of reflected light from each light source, generates a difference image of the two acquired images, and detects surface defects in the inspection target area from the generated difference image; and a scanning unit which causes the imaging unit to scan in the axial direction of the pipe, wherein the two distinguishable light sources irradiate illumination light onto the same specified inspection target area, and the imaging element captures an image of the specified inspection target area.

2. A pipe surface inspection device as described in claim 1, wherein the imaging pitch of the imaging element in the axial direction of the pipe is smaller than a value calculated from the inner radius of the pipe, the maximum displacement of the imaging element expected when the imaging unit is scanned in the axial direction of the pipe, the axial overlap width of the two fields of view of the imaging elements adjacent to each other in the axial direction of the pipe, and the angle of view of the imaging element in the axial direction of the pipe.

3. A pipe surface inspection device as described in claim 1 or 2, wherein the imaging element is a full-circumference imaging element capable of obtaining an image containing the entire circumferential area of ​​the inner surface of the pipe in a single imaging operation.

4. A pipe surface inspection device as described in any one of claims 1 to 3, wherein the angle between the angle formed by the direction of specular reflection of illumination light from an upstream light source, which is the light source arranged upstream of the imaging element in the axial direction of the pipe, and the imaging direction of the imaging element, and the angle formed by the direction of specular reflection of illumination light from a downstream light source, which is the light source arranged downstream of the imaging element in the axial direction of the pipe, and the imaging direction of the imaging element, is greater than a value calculated from the inner radius of the pipe, the maximum displacement amount of the imaging element expected when the imaging unit is scanned in the axial direction of the pipe, the angle of view of the imaging element in the axial direction of the pipe, and the distance from the imaging element to the upstream light source or the downstream light source.

5. A tube surface inspection apparatus as claimed in any one of claims 1 to 4, wherein the two distinguishable light sources are annular illumination.

6. A pipe surface inspection device according to any one of claims 1 to 5, wherein the scanning unit scans the imaging element along the central axis of the pipe.

7. A pipe surface inspection method for inspecting the inner surface of a pipe as an inspection target area, comprising: a processing step of acquiring images of reflected light from each light source using an imaging unit having an imaging element having a field of view on the surface of the inner surface and two distinguishable light sources, and arranged on the axis of the pipe so that the imaging element is sandwiched between the two distinguishable light sources, generating a difference image between the two acquired images, and detecting surface defects in the inspection target area from the generated difference image; and a scanning step of scanning the imaging unit in the axial direction of the pipe, wherein the processing step includes a step of irradiating the same predetermined inspection target area with illumination light using the two distinguishable light sources, and capturing an image of the predetermined inspection target area using the imaging element.

8. A method for inspecting the surface of a pipe as described in claim 7, wherein the imaging pitch of the imaging element in the axial direction of the pipe is smaller than a value calculated from the inner radius of the pipe, the maximum displacement of the imaging element expected when the imaging unit is scanned in the axial direction of the pipe, the axial overlap width of the two fields of view of the imaging elements adjacent to each other in the axial direction of the pipe, and the angle of view of the imaging element in the axial direction of the pipe.

9. A method for inspecting a pipe surface as claimed in claim 7 or 8, wherein the imaging element is a full-circumference imaging element capable of obtaining an image containing the entire circumferential area of ​​the inner surface of the pipe in a single imaging operation.

10. A tube manufacturing apparatus comprising the tube surface inspection apparatus according to any one of claims 1 to 6.

11. A method for manufacturing a tube, comprising inspecting an inner surface of the tube using the method for inspecting tube surface according to any one of claims 7 to 9, and manufacturing the tube based on the inspection results.

12. A pipe quality control method, comprising inspecting the inner surface of a pipe using the pipe surface inspection method according to any one of claims 7 to 9, and controlling the quality of the pipe based on the inspection results.

13. A pipe, the quality of which is assured by the method for inspecting the surface of a pipe according to any one of claims 7 to 9.

Images