JP2001141664A - Inspection support method in surface inspection - Google Patents

Abstract

PROBLEM TO BE SOLVED: To enhance the reliability of inspection while shortening an inspection time by displaying the visual inspection of a penetration flaw detection test result performed heretofore by the naked eye as an image. SOLUTION: A flaw inspecting apparatus takes in a plurality of images within a range to be inspected at the same time and stores them in a memory medium.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention relates to a visual inspection which is one of a plurality of nondestructive inspections, and facilitates a visual inspection through an image taken by a CCD camera or the like, instead of performing the visual inspection directly with the naked eye. An imaging method that simultaneously uses a plurality of video devices as one component of the devices.

[0002] With this, the field of view range has been restricted in the past, but it is possible to simultaneously perform a plurality of imagings, thereby contributing to shortening of inspection time and improvement of reliability.

In addition, by storing inspection results and the like in a storage medium, the reliability of the inspection results is improved and the inspection time is reduced.

[0004]

2. Description of the Related Art Penetration testing is one of the methods for inspecting material flaws of industrial materials.

The principle will be described with reference to FIG.

FIG. 1A shows a test material 1 having a weld 2, and the weld 2 has a material flaw 3 opened on the surface thereof.
When the material flaw 3 has a small opening width and is not directly visible to the naked eye or is very difficult to see, a penetrant inspection method, which is one of the non-destructive inspection methods, is applied to make the material flaw easily visible. Hereinafter, an outline of the method will be described.

FIG. 1A shows a state before the start of the penetrating inspection test process. In the first step of the process, a red penetrating liquid 4 is applied to the test surface as shown in FIG. Part of the permeating liquid 4 penetrates into the fine material flaws 3. FIG. 1C shows a second step in which the permeated liquid 4 on the test surface is wiped off. Most of the permeate on the test surface is removed, leaving only the permeate in the material flaw 3. Next, in a third step of FIG. 1 (d), a developer in which white fine powder is suspended in a volatile solvent is applied to the test surface.

[0008] The volatile solvent evaporates immediately, after which a developed coating film 6 of fine powder is formed.

The flaw indication pattern 7 oozes out into the developed coating film and is displayed with its width enlarged, especially from the original material flaw 3, and the red filing pattern is displayed on the white developed coating film. Are formed, and the flaws which were not visible to the naked eye before the penetrant test can be easily seen from the good contrast of the hues.

The fifth step is to visually observe the flaw indication pattern and detect the flaw, and this visual inspection by the naked eye has been a factor that hinders the recordability and automation of the penetrant inspection test. .

In the above description, an example of a penetrant test using a red penetrant has been described, but there is a method called a fluorescence penetrant test using a penetrant that emits fluorescence. In this case, the fluorescent indication pattern is observed by illuminating with ultraviolet rays, but the visual inspection by the naked eye is the same.

Thus, the visual inspection by the naked eye is
In addition to visual inspection, it is widely used as a final step in penetration testing to check for flaws.

However, one of the drawbacks of this method is that it is performed with the naked eye of the inspector, so that recording properties are poor and automation of the work is difficult.

[0014]

The problem to be solved by the present invention is that in a visual inspection performed by a person, a defect may not be found completely or an inspection result may be different due to individual differences. It also included the problem of test reliability that the result was left only as characters.

In view of this, a plurality of images taken by a camera or the like are captured at the same time, and the inspection range necessary for the image is secured, including the color of the subject and the state of the test surface as the background. To save reliability in a storage medium, reduce the amount of work by simultaneously capturing a plurality of images, and improve efficiency by automation.

[0016]

[Means for Solving the Problems] The inspection results are displayed in a color image, and the portions judged to be defective by the automatic inspection are surrounded by rectangles, and the inspector confirms the rectangles one by one in the original image. Determine if it is a real defect. The original image and the inspection result are stored in a storage medium as a record so that they can be confirmed at a later date, thereby improving reliability. In addition, by using a plurality of cameras and capturing a plurality of images at once, the inspection time can be reduced.

[0017]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 2 is a configuration diagram of a defect inspection support apparatus for magnetic particle flaw detection according to the present invention. The test piece 21 has a crack defect 22
Is present, and it is assumed that the fluorescent magnetic powder has been applied and magnetized. This is the UV illumination connected to the illumination power supply 26.
By illuminating with 24, the fluorescent magnetic powder collected on the crack defect 22 emits green light. For this reason, color TV cameras 23
, The crack defect 22 appears as a dark green line. Further, for example, if the test body 21 has a welded portion, the fluorescent magnetic powder gathers along the weld bead, so that the pseudo defect 31 may be seen as a light green line. Color tv camera 23
In order to prevent the ultraviolet rays used for illumination from being incident on the color television camera 23 due to the reflection on the surface of the test piece 21 or a foreign substance, an ultraviolet cut filter 25 is attached in front of the lens.

RGB from a plurality of color television cameras 23
The color signal is supplied to an image memory 27,
By combining multiple image data taken at the same time using a pattern matching method, etc., and combining the images into one,
To analyze. The analysis result and the original image are displayed on the color monitor 29, and the inspector distinguishes the true crack defect 22 and the pseudo defect 31 from the defect candidates automatically displayed on the color monitor. When the inspection is completed, the analysis result and the original image are stored in the data storage device 30.

FIG. 3 is a diagram showing the effect of the ultraviolet cut filter 25. (a) is without filter, (b) is
This is when the filter is attached. In (a), the reflection by the foreign matter 32 such as lint and the specular reflection 33 from the test body 21 are photographed by the color television camera 23, making the defect inspection difficult.
In (b), these reflected lights are cut off, and an image of only light emission by the fluorescent magnetic powder can be captured as in the case where a person views.

FIG. 4 shows an example of an image processing algorithm for analyzing the contents of the image memory 27.

A color image is fetched, and then a differentiation process is performed on a green image containing the most light emission information of the fluorescent magnetic powder. As a result, a portion having a large linear luminance change such as a crack defect is emphasized, and a portion having a high luminance but a small luminance change such as a magnetic powder pool is not emphasized. Next, a threshold value for binarization is determined with reference to the average value of the green differential image, and binarization 43 is performed. Then, after removing image noise 44 such as an isolated point, the length, contrast, and the like of the remaining area are calculated, and if these values are larger than a specified value, they are determined as defect candidates.

FIG. 5 is a flowchart showing a process of confirming a defect. First, after synthesizing the captured image, a marker display 51 of a defect candidate is performed on the defect candidate portion. Next, the computer 28 requests the inspector to determine defect candidates one by one. The inspector looks at the original color image and determines whether it is a true defect. If the defect is recognized as a true defect, the position, length, contrast, etc. of the defect calculated when determining the defect candidate are data. The marker is stored in the storage medium 30, and the marker turns red. When the inspector determines that the defect candidate is a pseudo defect, the marker is erased. If the defect candidate still remains, the marker is displayed on the next defect candidate. When all the defect candidates have been confirmed, the color original image is stored in the data storage medium 30.

FIG. 6 shows an example of a defect candidate marker generation method. A center line 62 connecting the starting point P1 and the ending point P2 of the defect candidate 61 is obtained, and the long sides AB and CD of the defect candidate marker 63 are set at a distance of a fixed value d in parallel with the center line 62. The short sides AD and BC are determined in the same manner. The length of the defect is the distance between P1 and P2. The contrast related to the crack thickness is from P1 to P2
The contrast calculation line 64 is scanned, and the difference between the average luminance and the maximum luminance on this line is obtained. This difference is obtained from P1 to P2, and the average value is used as the defect contrast.
Note that the defect candidate marker is not necessarily a short form. Short side A
The method of making D and BC semicircular may also be used. An important point is that the defect is not hidden by a marker.

FIG. 7 shows an example of a defect candidate display method on the color monitor 29. The inspector is requested to confirm the original image in order from the candidate having the longest defect. At first, all the markers are displayed in white, the markers determined to be true defects are displayed in another color, for example, red, and those determined to be false are deleted.

The automatic inspection of magnetic particle flaw detection images has been described above. However, since the sensor is a color television camera, a common sensor probe capable of both penetrant flaw detection and magnetic particle flaw detection can be realized.

FIG. 8 shows an example of a probe commonly used for magnetic particle flaw detection and penetration flaw detection. The color television camera 23 is the same as in FIG. At the time of magnetic particle inspection, the ultraviolet illumination lamp 24a is turned on, and at the time of penetration inspection, the white illumination 24b is turned on.
The ultraviolet illumination connector 82a is connected to the illumination power supply 26 through an illumination power cable 83. To turn on the white light 24b, the lighting power cable 83 is connected to the white light connector 82b. Hood 81 is attached to prevent the influence of external light. In FIG. 8, a ring-shaped lamp is used, but a rod-shaped lamp may be used.

FIG. 9 shows a method of generating a defect candidate marker in penetrating inspection. A start point P1 and an end point P2 are obtained from the defect candidate 61, and a defect candidate marker 63 is generated in the same manner as in the case of magnetic particle flaw detection shown in FIG.

FIG. 10 shows an example of means for specifying a camera position and synthesizing a plurality of images when the test object 21 is a long object.

First, in order to specify the camera position, a scale 101 with a scale is fixed to the test piece 21 and the scale 101 is fixed.
Is captured so as to enter a part of the camera field of view 102. For the scale of the scale, for example, every 1 centimeter,
There is a way to write numbers. At the time of penetration testing, the scale is, for example, a red number on a white background, and at the time of magnetic particle testing, the scale number is a green fluorescent color. An accurate position is determined by the computer 28 by a pattern matching method or the like.

Next, as for a method of synthesizing a plurality of images, a plurality of images simultaneously picked up by a plurality of color television cameras 23 are set in advance so as to be overlapped 40 and picked up by a pattern matching method or the like. The images are synthesized by the computer 28 and synthesized and recognized as one image.

FIG. 11 shows an example of an imaged screen. The scale 101 is simultaneously imaged at the bottom of the screen, and the camera position on the test object 21 is calculated from the scale 101. That is, the scale numeral 120 is written on the scale 101, and can be recognized by a pattern matching method using the computer. In addition, the scale 101 has, for example, a dividing line 121 every one centimeter.
Is included, and a more detailed camera position can be calculated. A section signal 123 is obtained as an image signal of the search line 122 of C1 and C2 on the image. From now on, the left edge A and right edge of the image
The imaging magnification can be calculated from the positions of E1, E2, E3, E4, and E5 of B and the dividing line 121.
The exact position on the test piece 21 is known.

[0033]

By taking in a plurality of images at once, the inspection time can be shortened, and by storing the inspection images in a storage medium, the reliability of the inspection can be improved.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram of the principle of a penetration test.

FIG. 2 is a diagram showing the configuration of an apparatus according to the present invention.

FIG. 3 is a diagram illustrating an effect of an ultraviolet cut filter.

FIG. 4 is an image processing algorithm for defect inspection.

FIG. 5 is a flowchart showing a defect confirmation process.

FIG. 6 is a diagram showing an example of a defect candidate marker generation method.

FIG. 7 is a diagram showing an example of a defect candidate display method.

FIG. 8 is a diagram showing an example of a magnetic particle flaw detection / penetration flaw detection shared probe.

FIG. 9 is a diagram showing an example of a defect candidate marker generation method in penetrating inspection.

FIG. 10 is a diagram illustrating an example of an inspection position specifying unit for a long object.

FIG. 11 is a diagram illustrating an inspection position specifying method.

[Explanation of symbols]

1 ... test material, 2 ... weld, 3 ... material flaw, 4 ... permeate, 5
… Infiltrate in the flaw, 6… development coating, 7… flaw indication pattern, 21…
Specimen, 22 ... crack defect, 23 ... color TV camera, 24 ...
UV illumination, 25 UV cut filter, 27 image memory, 28 computer, 29 color monitor, 30 data storage device, 101 scale, 102 camera field of view, 103 camera field of view overlap, 120 numerical scale , 121 ... separator line.

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Tetsuo Taguchi 3-1-1 Sachimachi, Hitachi City, Ibaraki Prefecture Within the Nuclear Power Division, Hitachi, Ltd. (72) Inventor Toshiro Asano 3300 Hayano Mobara-shi, Chiba Co., Ltd. Within Hitachi Display Group (72) Inventor Kaoru Sakai 292 Yoshida-cho, Totsuka-ku, Yokohama-shi, Kanagawa Prefecture F-term in Hitachi, Ltd. Production Technology Research Laboratories 2G051 AA37 AB03 BA04 BA05 BA08 CA04 CA07 CA11 CB01 CC07 CD03 EA08 EA11 EA14 EA17 EA19 EA25 EB01 EC03 ED04 ED11 FA01

Claims (2)

[Claims] 1. In a surface inspection such as a crack of a metal, CC is used.
A defect inspection method characterized by using a plurality of D cameras and the like. 2. In the surface inspection, a scale with a numeral and a dividing line is placed on a test object to be inspected, and an image is taken so that the scale is included in a part of a camera field of view, thereby specifying an imaging position, and A defect inspection apparatus characterized in that a wide area image can be taken at the same time by capturing an image, and the image data can be stored in a storage medium.