JP2005283599A - Defect inspection method, defect inspection device and defect inspection support method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To do away with missing of a defect to improve reliability of defect inspection by allowing easy confirmation of a defect candidate by visual inspection by an image. <P>SOLUTION: This defect inspection method is related to the defect inspection by a magnetic particle inspection method illuminating a test body 1 with ultraviolet rays to make the crack defect 2a generate fluorescence. The surface of the test body 1 is imaged by a color television camera 3 through an ultraviolet cut filter 5. An original image by R, G, B signals outputted from the color video camera 3 is once stored in an image memory 7. The original image is displayed by a color monitor 9, and a computer 8 processes a G image to detect the defect candidate, adds a defect candidate marker to each the defect candidate inside the original image displayed on the color monitor 9, and makes the color monitor 9 display it. An inspector knows the defect candidate inside the original image by the defect candidate marker, and visually confirms whether the defect candidate is the genuine crack defect or a pseudo defect. The original image and an inspection result are stored in a data storage device 11. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to defect inspection such as cracking of the surface of a test body and its support, and more particularly to a defect inspection method, a defect inspection apparatus, and a defect inspection support method based on a nondestructive inspection method such as magnetic particle inspection and penetration inspection.

  In order to manage the quality of products such as tubes and various parts, defect inspection such as cracks on the surface of such products (hereinafter referred to as test specimens) is performed. As the inspection method, a nondestructive inspection method that does not damage the specimen is used, and in particular, there are a magnetic particle inspection method and a penetrant inspection method as defect inspection methods such as cracks on the surface of the specimen.

  The magnetic particle inspection method is a method for inspecting cracks that occur mainly on the surface of a metal having magnetism, such as iron, or a portion close to the inner surface. In this method, first, a solution containing fluorescent magnetic powder is sprayed on the surface of the specimen, and then the surface of the specimen is magnetized. If there is a defect such as a crack on the surface, the magnetic flux concentrates at the defective part, and as a result, the fluorescent magnetic powder collects at the defective part. Therefore, when this surface is irradiated with ultraviolet rays, the phosphor is excited and emits green light, but the light emission becomes stronger at the portion where the magnetic flux is concentrated and the fluorescent magnetic powder is collected, and as a result, the defect is emphasized. .

  By the way, even if the surface of the specimen is magnetized as described above, if there are valleys or other indentations on the surface, the magnetic flux may concentrate there, and these parts and wrinkles are emphasized like defects. Is done. Such an emphasized portion is hereinafter referred to as a pseudo defect with respect to a true defect. Conventionally, the true defect is detected by visually observing the surface of the test body irradiated with ultraviolet rays in this way. I was conducting an examination to search for.

  On the other hand, when the specimen is the same shape as a mass-produced part, the surface of the specimen that has been irradiated with ultraviolet rays is imaged with a monochrome video camera, and the resulting image signal is processed automatically. There is also an example in which defects are inspected (see, for example, Non-Patent Document 1).

  On the other hand, in the penetrant flaw detection method, a red solution called penetrant is applied to the surface of a test specimen, and after a predetermined time, it is removed with a cloth or the like. This leaves only the solution that has penetrated the defects such as cracks. Thereafter, a solvent mixed with white powder called developer is applied to the surface of the test specimen. With this developing solution, the red penetrating solution that has penetrated into the defect is sucked up, and a red indicator pattern by the penetrating solution sucked up on the white developing solution appears. Conventionally, inspection by this penetrant flaw detection method has also been performed visually.

The penetrant flaw detection method can be used regardless of the material of the specimen, and since it is simpler than the magnetic particle flaw detection method, this method is generally used. Has the advantage that the shape of the defect can be obtained relatively accurately, and a defect hidden on the surface, that is, near the surface of the specimen and not opened to the outside can be inspected.
“Summary of the 1998 Spring Conference” sponsored by the Japan Nondestructive Inspection Association, pp.305-308

  By the way, whether it is a magnetic particle flaw detection method or a penetrant flaw detection method, it is common to visually check whether it is a true crack defect or a pseudo defect. There is a problem in inspection reliability that there is a possibility that a true defect may be overlooked or an inspection result may differ depending on individual inspectors, and the inspection result remains only with characters such as “pass”.

  In addition, although there is an example of automatic inspection as described above, in this case, since a monochrome video camera is used, the inspection result image on the display screen is also monochrome, and the real thing and the different inspector are true from the image. It was difficult to check for defects.

  Further, in such automatic inspection, when the specimen is long, the imaging position of the video camera is not known, which part of the specimen is the obtained image, and which part of the specimen is the inspected defect Sometimes I didn't know.

  An object of the present invention is to provide a defect inspection method, a defect inspection apparatus, and a defect inspection support method that solve such problems and facilitate the determination of true defects.

  Another object of the present invention is to provide a defect inspection method, a defect inspection apparatus, and a defect inspection support method that make it possible to easily know a defect position even for a long specimen.

  In order to achieve the above object, in the present invention, a specimen is imaged with a color video camera in order to detect a defect by a magnetic particle inspection method or a penetration inspection method. According to this, the defect search is automatically performed by the image processing, and the obtained image is displayed in color, and the image of the specimen is displayed in the same manner as the real object, and the true defect is displayed. It becomes easy to confirm. In particular, in the case of the magnetic particle flaw detection method, the color video camera is configured to image the specimen through the ultraviolet cut filter, so that the ultraviolet rays reflected from the surface of the specimen are blocked from the color video camera. As a result, it is possible to prevent an image due to ultraviolet rays from appearing on the display screen, and it becomes easier to confirm a true defect.

  Further, according to the present invention, a display image of defect candidates (true defects and pseudo defects) obtained by processing an output image signal of a color video camera is indicated by a mark. As a result, the inspector only needs to confirm whether the defect candidate indicated by the mark is a true defect or a pseudo defect, and the work burden on the inspector can be reduced and a highly accurate inspection result can be obtained. become.

  In order to achieve the above-mentioned other object, the present invention attaches a scale to a specimen and images the scale together with the inspection area of the specimen so as to fall within the field of view of the color video camera. Thereby, the scale of this scale is also included and displayed in the image of the specimen, and even if the specimen is long, the position of the defect in the specimen can be easily grasped.

  According to the present invention, since the image input is performed using the color video camera, and further, in the defect inspection by the magnetic particle flaw detection method, the ultraviolet ray reflected from the test body can be cut by the ultraviolet cut filter. The results can be easily confirmed, and defect candidates are automatically instructed and displayed, so there is almost no oversight of defect inspection, and inspection images are saved, improving inspection reliability. .

  Further, according to the present invention, since a color video camera is used, automatic defect inspection for magnetic particle inspection and penetration inspection can be performed with the same sensor probe, and convenience is greatly improved.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.
FIG. 1 is a block diagram showing a first embodiment of a defect inspection apparatus according to the present invention, wherein 1 is a specimen, 2a is a crack defect, 2b is a pseudo defect, 3 is a color video camera, and 4a and 4b are UV illuminations. Means 5 is an ultraviolet cut filter, 6 is an illumination power supply, 7 is an image memory, 8 is a computer, 9 is a color monitor, 10 is an operation unit, and 11 is a data storage device.
Hereinafter, the operation of this embodiment will be described as being based on the magnetic particle flaw detection method.

  In FIG. 1, a test body 1 is placed on an inspection table (not shown). Here, there is a crack defect 2a on the first surface of the test (however, this crack defect 2a may have an opening on the surface, or near the surface which does not have an opening and does not appear on the surface). It is assumed that the fluorescent magnetic powder solution has already been sprayed and magnetized. When power is supplied from the illumination power source 6 to the ultraviolet illuminating means 4a and 4b to illuminate the test body 1 with ultraviolet light, the fluorescent magnetic powder collected on the crack defect 2a on the surface of the test body 1 emits green light. For this reason, when the surface of this test body 1 is imaged with the color television camera 3, the crack defect 2a appears as a dark green line.

  Some parts other than the crack defect 2a are visible as dark green lines. For example, if there is a welded part on the test body 1, the magnetic flux is strong along the valley along the weld bead or when the surface is not completely flat and there is a concave part such as a valley. Fluorescent magnetic powder will collect, and even in this part, it will appear as a dark green line. As will be described later, in the computer 8, the defect is detected by image processing, but the portion along the weld bead is also detected as a defect by this image processing. In this way, defects other than true defects detected by image processing are referred to as pseudo defects, and such a pseudo defect 2b is also present in the specimen 1. In addition, when there is a gentle undulation on the surface, fluorescent magnetic powder may remain in the lower part, and this part also emits green light when irradiated with ultraviolet rays. However, this portion can be removed by image processing to be described later, and therefore this portion is not a pseudo defect.

  The surface of the specimen 1 illuminated with ultraviolet rays is imaged by the color television camera 3. In this case, the color video camera 3 is provided with an ultraviolet cut filter 5, which cuts ultraviolet rays reflected from the surface of the test body 1 and foreign matters on the surface. Although the human eye is not sensitive to ultraviolet rays, the color video camera 3 is sensitive. Therefore, an undesired image is displayed on the ultraviolet reflection portion on the display screen based on the image signal obtained from the color video camera 3. If this overlaps with a true defect, it becomes difficult to confirm the defect. In order to prevent this, an ultraviolet cut filter 5 is provided.

  The color television camera 3 obtains R (red), G (green), and B (blue) three primary color signals, and temporarily stores the three primary color signals for one image in the image memory 7. The computer 8 reads an image signal for one screen from the image memory 7 and displays the image signal on the color monitor 9 as an original image. Also, the computer 8 analyzes the image signal read from the image memory 7 by image processing to find defect candidates. Explore. Here, what is searched for as a defect candidate is a true crack defect 2 a and a pseudo defect 2 b on the surface of the specimen 1.

  The above is the outline of one embodiment of the defect inspection method in this defect inspection apparatus. Next, the defect inspection according to the present invention for searching for the true split defect 2a from the defect candidates obtained by this defect inspection method. An embodiment of the support method will be described.

  The computer 8 attaches marks (referred to as defect candidate markers) to the true crack defect 2a and the pseudo defect 2b in the original image displayed on the color monitor 9 according to the result of the defect candidate search. To do. The inspector can easily know the defect candidate in the original image by using the defect candidate marker, and visually discriminate whether the defect candidate is the true crack defect 2a or the pseudo defect 2b. The discrimination result is input from the operation unit 10, and the computer 8 erases the defect candidate marker for the pseudo defect 2b in accordance with the input information. When the inspection for all defect candidates is completed, the computer 8 stores the inspection result and the original image in the data storage device 10.

  The original image and the inspection result stored in the data storage device 10 can be appropriately read out and displayed on the color monitor 9 by the operation of the operation unit 10, or can be printed out. Of course, it is possible to enlarge and display the portion of the true split defect and print it out, and it is also possible to indicate the true split defect with a marker similar to the defect candidate marker based on the inspection result.

  FIG. 2 is a diagram showing the effect of the ultraviolet cut filter 5. FIG. 2A shows a display image based on the output image signal of the color video camera 3 when the ultraviolet cut filter 5 is not used. Indicates a display image based on an output image signal of the color video camera 3 when the ultraviolet cut filter 5 is used.

  In FIG. 2A, when the ultraviolet cut filter 5 is not used, in addition to the image by the emission of fluorescent magnetic particles such as the true crack defect 2a and the pseudo defect 2b, the regular reflection portion of the ultraviolet ray on the surface of the test body 1 2c and the image 2d of foreign matter such as lint that reflects ultraviolet rays on the surface of the test body 1 may be displayed in green, making it indistinguishable from the true split defect 2a. When it overlaps with the true crack defect 2a, the true crack defect 2a may be overlooked.

  On the other hand, when the ultraviolet cut filter 5 is provided in the color video camera 3, as shown in FIG. 2B, the images 2c and 2d caused by the reflection of the ultraviolet rays and entering the color video camera 5 can be removed. Only the image by the emission of the fluorescent magnetic powder is obtained.

  Next, image processing of an embodiment of the defect inspection method according to the present invention used in the defect inspection apparatus shown in FIG. 1 will be described with reference to FIG. FIG. 3 is a flowchart showing an image processing algorithm for analyzing the contents of the image memory 7 by the computer 8 in FIG. 1 in this defect inspection method.

  In the figure, the computer 8 first reads the image memory 7 and takes in a G (green) image (step 100). Then, the G image is differentiated (step 101), and a G differential image in which the portion whose level changes suddenly is enhanced is obtained. As a result, the amount of light emission information from the fluorescent magnetic particles such as crack defects is increased, and the level of the portion where the luminance changes linearly is emphasized. The luminance such as the magnetic particle pool is high, but the level of the portion where the luminance change is small is emphasized. Not. Next, a threshold value for binarization is determined with reference to the average level of the G differential image, and the G differential image is binarized using the threshold value (step 102). Then, image noise such as an isolated point is removed from the binarized G differential image (step 103).

  By the above processing, a binary image consisting of an image of a portion having a large linear luminance change is obtained, and this image represents a temporary defect candidate. Therefore, the position, length, and contrast are calculated for each temporary defect candidate. In FIG. 2, if the true crack defect 2a is detected as a temporary defect candidate, the position of the temporary defect candidate is the center of gravity position of the temporary defect candidate, and the length of the temporary defect candidate is The length of the straight line connecting both ends of the temporary defect candidate is set. The contrast of this temporary defect candidate is the average value of the levels of all the pixels constituting the temporary defect candidate in the G image.

  When the position, length, and contrast of each temporary defect candidate are obtained in this manner, it is next determined for each temporary defect candidate whether the length and contrast are equal to or greater than a preset threshold value. Both temporary defect candidates that are equal to or greater than the threshold are determined as defect candidates (step 105). For example, in FIG. 2B, the true crack defect 2a and the pseudo defects 2b, 2b ′ become temporary defect candidates, and at least one of the length and contrast of the pseudo defect 2b ′ is less than the threshold value. The true crack defect 2a and the pseudo defect 2b are determined as defect candidates.

  Next, an embodiment of the defect inspection support method according to the present invention used in the defect inspection apparatus shown in FIG. 1 will be described in detail with reference to FIG. FIG. 4 is a flowchart showing a process of confirming a true defect from the defect candidates obtained by the process of FIG.

  In this figure, when a defect candidate is determined by the processing shown in FIG. 3, the computer 8 first displays a marker on one defect candidate displayed on the display screen of the color monitor 9. This marker is a defect candidate marker, which will be described with reference to FIG. Here, the defect candidate marker is a frame surrounding the defect candidate.

In FIG. 5, the positions of both ends P ₁ and P ₂ of the defect candidate 12 are obtained when the lengths are obtained. First, a straight line connecting both ends P ₁ and P ₂ is obtained as the center line 13. Then, a point a1 at a distance d is obtained from an end portion P ₁ on a straight line obtained by extending the center line 13 from the end portion P ₁ . Also, determine the point a2, a6 distance d from the end portion P ₁ on the perpendicular straight line as the center line 13 of the end portion P ₁ on both sides. A same applies to the end portion P ₂ side, obtains a point a4 distance d to the center line 13 from the end portion P ₂ on the straight line that extends from an end portion P _2, also passes through the center line of the end portion P ₂ 13 finding the point a3, a5 distance d on either side from the end portion P ₂ on the straight line perpendicular to.

  Next, a straight line connecting points a2 and a3, a straight line connecting points a5 and a6, a straight line passing through points a1 and parallel to a straight line connecting points a2 and a6, and a straight line passing through point a4 parallel to a straight line connecting points a3 and a5 , And the intersection of a straight line connecting points a2 and a3 and a straight line connecting points a2 and a6 and passing through point a1 is a point A, a straight line connecting points a2 and a3, and a straight line connecting points a3 and a5. An intersection with a straight line passing through the point a4 in parallel and the point B, a straight line connecting the points a5 and a6 and a straight line connecting the points a3 and a5 and parallel with the straight line passing through the point a4 are connected to the point C, and points a5 and a6 are connected. An intersection point between a straight line and a straight line passing through the point a1 parallel to the straight line connecting the points a2 and a6 is defined as a point D, and a frame connecting the point A → B → C → D → A is defined as a defect candidate marker 14.

  Here, the true crack defect generally has a substantially straight line shape, and there are almost no curved lines that are largely bent. The distance d is set as small as possible without causing the obtained defect candidate marker 14 to contact or intersect the defect candidate 12.

  In addition, although the short sides DA and BC of the defect candidate marker 14 are linear, other shapes such as an arc, a triangle, and a trapezoid may be used.

  In this way, the defect candidate marker 14 is created for each defect candidate, and in step 200 of FIG. 4, the defect candidate marker 14 is surrounded by the defect candidate in the original image displayed on the color monitor 9 (FIG. 1). For example, it is displayed in white. FIG. 6 shows each defect candidate 2a, 2b, 2b ′ as a defect candidate in which the true split defect 2a and the pseudo defects 2b, 2b ′ shown in FIG. 2B are determined (step 105 in FIG. 3). It is a figure which shows the state which displayed defect candidate marker 14a, 14b, 14c.

In step 104 of FIG. 3, the defect candidate marker 14 may be created for each temporary defect candidate as described in FIG. In this case, the midpoint of the center line 13 between both ends P ₁ and P ₂ is set as the position of this temporary defect candidate, and the length of the center line 13 between both ends P ₁ and P ₂ is set as this temporary defect. The length of the candidate. In this case, the contrast of the temporary defect candidate can be calculated as follows. That is, a straight line between the intersections b and c of the defect candidate marker 14 parallel to the short side AD and the long sides AB and CD is used as the contrast calculation line 15, and the average luminance and the maximum luminance of the pixels on the contrast calculation line 15 are calculated. These brightness differences are obtained, and such brightness differences are obtained for each movement while the contrast calculation line 15 is sequentially moved from one end P ₁ to the other end P ₂ by one pixel or a plurality of pixels. The average of the obtained luminance differences is set as the contrast of the temporary defect candidate.

  As described above, when the process of step 104 in FIG. 3 is performed, the defect candidate marker 14 is formed for each temporary defect candidate, so that the defect candidate in the process of calculating the position, length, and contrast of the temporary defect candidate. The marker 14 can be created, processing is performed without waste, and when a defect candidate is determined (step 105 in FIG. 3), the defect candidate marker is displayed on the defect candidate immediately determined in step 200 in FIG. Can do.

  As shown in FIG. 6, defect candidate markers may be displayed simultaneously on all defect candidates in the original image, but in FIG. 4, only one defect candidate is displayed as a marker.

  Returning to FIG. 4, when a marker is displayed on one defect candidate of the original image of the color monitor 9 (step 200), the inspector visually checks the defect candidate displayed on the marker (step 201), and this defect candidate. Is a true crack defect or a pseudo defect (step 202). When the inspector determines that this is a true crack defect and inputs information indicating that fact by operating the operation unit 10 (FIG. 1), the computer 8 (FIG. 1) determines this true crack based on this input information. The defect information for the defect (the information obtained in step 104 of FIG. 3 and the defect candidate marker 14 etc.) is stored in the data storage device 11 (FIG. 1) (step 203) and displayed for this true split defect. The display color of the defect candidate marker 14 is changed from white to red and displayed as a defect marker. If there are other defect candidates that have not been visually confirmed (step 206), the marker display is performed for the next defect candidate after a predetermined time has elapsed (step 200). Make sure that you can visually check. When the next defect candidate is displayed as a marker, the marker that is visually confirmed and displayed in red is also deleted.

  In this way, only one marker is displayed, including the defect candidate marker and the defect marker, and the displayed marker does not overlap with the candidate defect to be visually confirmed. Defect candidates can be easily known, and visual confirmation is facilitated.

  Further, when the inspector visually confirms the defect candidate and determines that it is a pseudo defect, and inputs information indicating the defect from the operation unit 10 (step 205), the computer 8 immediately deletes the defect candidate marker 14 for the defect candidate. . If other defect candidates that have not been visually confirmed remain (step 206), the marker display is performed on the next defect candidate after a predetermined time (step 200).

  As described above, defect information of defect candidates determined to be true split defects is sequentially stored in the data storage device 11 (step 203), and visual confirmation of all defect candidates is completed (step 206: this fact) The white defect candidate marker is not displayed, and it can be known that the defect candidate displayed by the marker has disappeared), and the color original image composed of R, G, B is read from the image memory 7 and is stored in the data storage device. (Step 207). Thus, the defect inspection support operation is also completed.

  In the above description of FIG. 4, the white defect candidate markers are displayed one by one every time they are visually confirmed. However, when the defect candidates are not displayed so closely, all defect candidates are displayed. Markers may be displayed at the same time. In this case, in FIG. 4, a white marker is displayed on all defect candidates in step 200, and a red marker is displayed for those that are determined to be true split defects by visual confirmation. Then, it is displayed as it is (step 204), and the marker is deleted for those determined to be pseudo defects (step 205). If there is a defect candidate for which a white marker is displayed, the process returns from step 206 to step 201 so that the next defect candidate can be visually confirmed. In this case, for the defect candidate to be visually confirmed next, the computer 8 causes the defect candidate marker 14 to blink, for example. Thereby, the defect candidate which should be visually confirmed next can be known easily.

  As described above, in the first embodiment, the specimen 1 is imaged by the color video camera 3 so that the original image can be displayed in color. Therefore, the image of the specimen 1 is clearly displayed. Defects can be easily confirmed, and defect candidates to be inspected are automatically narrowed down by image processing, and defect candidates are displayed as markers. Can be easily and clearly known, and there is no omission of confirmation.

  In the first embodiment described above, an automatic inspection is performed by image processing using a magnetic particle flaw detection method, but a color television camera is used as a sensor, and even in a penetrant flaw detection method, as a sensor. Since a color video camera can be used, it is possible to realize a common sensor probe to which both penetration inspection and magnetic particle inspection can be applied.

  FIG. 7 is a block diagram showing the main part of the second embodiment of the defect inspection apparatus according to the present invention, that is, a magnetic particle / penetration flaw detection probe, wherein 16 is a hood, 17a and 17b are connectors, 18 is a power cable, Reference numeral 19 denotes a white illumination means, and parts corresponding to those in FIG.

  In this figure, ultraviolet illumination means 4 and white illumination means 19 are provided here, and in order to avoid the influence of external light, the light incident portions of these illumination means 4 and 8 and the color video camera 3 are the hood 16. Covered by.

  When such a shared probe is used in a defect inspection apparatus using a magnetic particle flaw detection method, the color video camera 3 is connected to the image analysis apparatus so as to have the configuration shown in FIG. 1, the power cable 18 is connected to the connector 17a, and the ultraviolet illumination means is used. 4 may be driven. In this case, defect inspection and support thereof are the same as those in the embodiment shown in FIG. When used in a defect inspection apparatus using the penetrant flaw detection method, the color video camera 3 is connected to an image analysis apparatus having the same configuration, the power cable 18 is connected to the connector 17b, and the white illumination means 19 is driven. What should I do?

  In addition, as a lamp | ramp of the ultraviolet illumination means 4 or the white illumination means 19, although it is a ring-shaped thing here, a rod-shaped thing may be sufficient.

  Next, a specific example of the defect candidate extraction method based on the penetrant flaw detection method will be described with reference to FIG. 8. The basic configuration of the image analysis apparatus in this case is the same as that shown in FIG. Since it is composed of a color monitor, an operation unit and a data storage device, and only the processing by the computer is different, it will be described with reference to FIG.

  In the case of defect inspection by the penetrant flaw detection method, in FIG. 7, the crack defect portion on the surface of the specimen 1 is represented in red. The surface of the test body 1 is illuminated with white light by the white illumination means 19, and this surface is imaged by the color video camera 3. One image of the R, G, and B primary color signals output from the color video camera 3 is written and stored as an original image in the image memory 7 in FIG.

  When the writing is completed, the computer 8 takes in the R, G, B images from the image memory 7 (step 300), and the chromaticity values x, y from the R, G, B pixel values (levels) for each pixel. Is obtained (step 301).

Now, assuming that the pixel values for R, G, and B of the pixels are R, G, and B, respectively, the stimulus values X, Y, and Z given to the color video camera 3 from the surface of the specimen 1 are expressed by the following equation (1). It is.

However, a _{11 to} a ₃₃ are coefficients determined by the color video camera 3. Therefore, the chromaticity x and y in this pixel are expressed by the following formula 2.

Next, the obtained chromaticity x and y are converted into data of hue θ and saturation r (step 302). In FIG. 9, assuming that the white chromaticity is (x _C , y _C ), the hue θ and the saturation r of the arbitrary chromaticity (x, y) can be obtained by the following equations (3) and (4).

Next, it is determined whether or not the obtained hue θ and saturation r are within a red range defined as a defect, and an image within the range is set as a temporary defect candidate (step 303). That is, as shown in FIG. 9, the ranges of hues θ _{1 to} θ ₂ and saturations r _{1 to} r ₂ are set as the red range, and the hue θ and saturation r obtained as described above are set. But
θ ₁ ≦ θ ≦ θ ₂ and r ₁ ≦ r ≦ r ₂
When the above condition is satisfied, the pixels having the data of hue θ and saturation r are in the red range, and a pixel group formed by adjoining such pixels is a temporary defect candidate.

Next, for each temporary defect candidate obtained as described above, the position, area A, length L, and maximum saturation r _MAX are obtained (step 304). Then, the area A, the length L, and the maximum saturation r _MAX are respectively set with respect to the preset area threshold A ₀ , the length threshold L ₀ , and the maximum saturation threshold r ₀ .
A ₀ ≦ A, L ₀ ≦ L, r ₀ ≦ r _MAX
Are satisfied at the same time, a temporary defect candidate that satisfies this is determined as a defect candidate (step 305).

  The defect inspection support method for the determined defect candidate is the same as that in the embodiment shown in FIG. 4. For example, a white defect candidate marker is displayed on each defect candidate, and a true crack is confirmed by visual confirmation. For the defect, the defect candidate marker is changed from white to, for example, blue, and for the pseudo defect, the defect candidate marker is erased.

FIG. 10 shows a state where a defect candidate marker is attached to a defect candidate. As described above, the defect candidate 20 in the penetration flaw detection method is a region having an area. Both ends P ₁ and P ₂ of the defect candidate 20 are detected, a straight line connecting these both ends P ₁ and P ₂ is set as a center line 21, and the length of the center line 21 is the length L of the defect candidate 20. The number of pixels in the defect candidate 20 is the area A of the defect candidate 20, and the center of gravity of the defect candidate 20 (or the center position of the center line 21) is the position of the defect candidate 20. Further, the defect candidate marker 22 for the defect candidate 20 is obtained as shown in FIG. 5 from both ends P ₁ and P ₂ and the center line 21, and the distance d from both ends P ₁ and P ₂ is determined as magnetic powder. The defect candidate marker 22 is not overlapped with the defect candidate 20 by making it larger than in the case of the flaw detection method. Of course, as in the case of the magnetic particle flaw detection described above, the defect candidate marker 22 may be created when the area and length of the temporary defect candidate are obtained in step 304 of FIG.

  In this embodiment, the defect inspection apparatus including the image analysis apparatus including the image memory, the computer, the color monitor, the operation unit, and the data storage device as shown in FIG. 1 is based on the magnetic particle inspection method and the penetration inspection method. There are two types, one that can be used for both, and one defect inspection apparatus in which the above-described image analysis apparatus is connected to the magnetic particle / penetration flaw detection shared probe of this embodiment By enabling the computer 8 (FIG. 1) to perform the image processing of the defect inspection of the magnetic particle flaw detection method shown in FIG. 3 and the image processing of the defect inspection of the penetrating flaw detection method shown in FIG. The method can be shared by the law and the penetrant flaw detection method. In this case, when it is desired to perform a defect inspection by the magnetic particle inspection method, the power cable 18 is connected to the connector 17a in FIG. 7, and in FIG. In addition, when it is desired to perform a defect inspection by the penetrating flaw detection method, the power cable 18 is connected to the connector 17b in FIG. 7, and in FIG. It may be instructed that the inspection is a defect inspection.

FIG. 11 is a diagram showing still another embodiment of the defect inspection method according to the present invention, in which 23 is a scale, 24 is a camera field of view, and the other portions are the same as in the previous embodiment. In addition, the same reference numerals are given to the portions corresponding to the above-mentioned drawings, and the overlapping description is omitted.
This embodiment is the same as the above-described embodiment as a basic configuration, but is particularly effective for a long specimen.

  In FIG. 11, the test body 1 is long so as not to fit within the camera field of view 24 of the color video camera (FIG. 1). For such a test body 1, a scale 23 is arranged along the longitudinal direction of the test body 1. The arrangement of the scale 23 in this case is such that the surface to be inspected of the specimen 1 is not hidden by the scale 23 when viewed from the color video camera, and the scale 23 falls within the camera field of view 24.

  As this scale 23, the dividing line 23a is provided in equal intervals (for example, every 1 cm), and the scale number 23b which shows an order is attached | subjected for every area divided by this dividing line 23a. Also, the scale 23 in the case of defect inspection by the magnetic particle flaw detection method and the scale 23 in the case of defect inspection by the penetrant flaw detection method can be different in color. For example, in the scale 23 used for defect inspection by the magnetic particle flaw detection method, the separation line 23a and the number 23 are colored green in white on the white background, and in the scale 23 used for defect inspection by the penetrant flaw detection method, the separation line 23a is white on the background. And the color of the number 23 is red.

  FIG. 12A is a diagram showing a specific example of the display screen by imaging shown in FIG. 11, and reference numeral 25 denotes the display screen. The parts corresponding to those in FIG. Omitted. FIG. 12B is a waveform diagram showing an image signal on this display screen.

  In FIG. 9A, a display screen 25 (display screen of the color monitor 9 in FIG. 1) includes a partial image of the surface of the specimen 1 (FIG. 11) and a partial image of the scale 23 (FIG. 11). The original image consisting of When only the display image of the test body 1 is viewed, the position of the color video camera 3 with respect to the test body 1 (the position in this case is the end position of the test body 1 as a reference position, and the test body from this reference position is The distance in the longitudinal direction of 1) is unknown, but the position becomes clear when the image of the scale 23 is viewed. In the example shown in the drawing, it can be seen that the color video camera 3 is located at a position approximately 17 cm in the longitudinal direction from the end of the specimen 1 and the defect candidate 2 is located at this position of the specimen 1.

  Further, the position can be obtained automatically and accurately using the image of the scale 23. That is, the computer 8 (FIG. 1) can recognize the scale numbers 23b of the scale 23 by pattern matching, and can detect the division line 23a of the scale 23. Therefore, the scale recognized as the detected division line 23a. It is possible to recognize how far each position on the display screen 25 is from the end of the specimen 1 by the positional relationship with the numeral 23 b.

  In this case, the image signal of the search line 26 parallel to the image of the scale 23 on the display screen 25 has peaks E1, E2,... With respect to the dividing line 23a of the scale 23, as shown in FIG. By detecting this, the position of the dividing line 23a on the display screen 25 can be detected. That is, for each separation line 23 a displayed on the display screen 25, the position from the end of the specimen 1 and the position on the display screen 25 are associated with each other.

In FIG. 12A, for example, the dividing line 23a between the scale numbers 17 and 18 on the scale 23 is at a position 17 cm from the end of the test body 1, but as a result of calculation by the computer 8, the display screen 25 is displayed. in there from the left end at a distance h ₁ at pixel, separators 23a between the scale numbers 18 and 19, but from an end portion of the specimen 1 to the position of the 18cm, the left end portion on the display screen 25 at the position of distance h ₂ in a pixel unit and a and, at a distance h in pixels from the left end portion at the location of the defect candidate 2 (center of gravity position M) of the display screen 25 on (however, h ₁ <h <h ₂ And the actual distance between the adjacent separation lines 23a is H (here, H = 1 cm as described above), the position of the defect candidate 2 on the specimen 1 Is expressed by the following equation (5).

As described above, the position of the defect candidate 2 on the specimen 1 can be detected with high accuracy, and further, the distance between the dividing lines 23a on the display screen 25, the number of pixels between them, the pixel width, and the like are parameters. Since the imaging magnification is obtained from the calculated distance and the actual distance H, the distances h ₁ and h in FIG.
17+ (h−h ₁ ) × α (cm)
(Where α is the imaging magnification)
From the above, the position M of the defect candidate 2 can also be obtained.

It is a block diagram which shows 1st Embodiment of the defect inspection apparatus by this invention. It is a figure which shows the effect of the ultraviolet cut filter in embodiment shown in FIG. It is a flowchart which shows one Embodiment of the defect inspection method by this invention used for the defect inspection apparatus shown in FIG. It is a flowchart which shows 1st Embodiment of the defect inspection assistance method by this invention used for the defect inspection apparatus shown in FIG. It is a figure which shows one specific example of the production | generation method of the defect candidate marker displayed with the defect inspection assistance method shown in FIG. It is a figure which shows the defect candidate displayed with the marker by the defect inspection assistance method shown in FIG. It is a block diagram which shows the principal part of 2nd Embodiment of the defect inspection apparatus by this invention. It is a flowchart which shows one Embodiment of the defect inspection method by this invention used for the defect inspection apparatus shown in FIG. It is explanatory drawing of step 302,303 of FIG. It is a figure which shows a specific example of the defect candidate marker used by the penetration flaw detection in embodiment shown in FIG. It is principal part explanatory drawing of 3rd Embodiment of the defect inspection apparatus by this invention. It is a figure which shows the display screen in embodiment shown in FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Specimen 2a True crack defect 2b, 2b 'Pseudo defect 3 Color television camera 4 Ultraviolet illumination means 5 Ultraviolet cut filter 7 Image memory 8 Computer 9 Color monitor 10 Operation part 11 Data storage device 12 Defect candidate 13 Center line 14, 14a -14c Defect candidate marker 16 Hood 17a, 17b Connector 18 Power cable 19 White illumination means 20 Defect candidate 22 Defect candidate marker 23 Scale 23a Separator line 23b Scale number 24 Camera field of view

Claims (8)

In defect inspection method by magnetic particle inspection method,
A defect inspection method comprising imaging a test object to be inspected using a color television camera. In claim 1,
The defect inspection method, wherein the video camera images the specimen irradiated with ultraviolet rays through an ultraviolet cut filter. In the defect inspection method by the penetrant flaw detection method,
A defect inspection method comprising imaging a test object to be inspected using a color television camera. In claim 1, 2 or 3,
A defect inspection method, wherein a scale is arranged in a field of view of the color video camera that images the test body. In defect inspection equipment by flaw detection method,
A color video camera for imaging the specimen to be inspected;
An ultraviolet illuminating means and a white illuminating means are provided as illuminating means for the test body, and defect inspection by a magnetic particle flaw detection method using the ultraviolet illuminating means and defect inspection by a penetrating flaw detection method using the white illumination means can be performed. A defect inspection apparatus characterized by being configured as described above. In claim 5,
A defect inspection apparatus, wherein a scale is arranged in a field of view of the color video camera that images the specimen.   An image signal output from the color video camera according to any one of claims 1 to 6 is arithmetically processed to determine and detect a defect candidate of the specimen, and color based on an output color image signal of the color video camera A defect inspection support method, wherein an original image is displayed on a screen, and an image determined to be a defect candidate in the color original image is indicated by a marker. In claim 7,
When the defect candidate displayed on the screen is a true defect, the display color of the marker for the true defect is changed, and when the candidate defect is a pseudo defect, the marker for the pseudo defect is erased. A defect inspection support method characterized by the above.

Images