JP3523764B2 - Foreign object detection device in nonmetallic inclusion measurement - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to non-metallic inclusion measurement for detecting foreign matters such as holes generated in dust and bubbles on the surface of a sample by separating them from non-metallic inclusions when detecting a metallographic structure with a microscope. The present invention relates to a foreign matter detection device.

[0002]

2. Description of the Related Art A television camera is connected to a metallographic microscope to measure non-metallic inclusions in a steel material to be measured from a captured image. These are JIS-G-05
55, ASTM-E45 / 1245, etc., a binary image is created from the captured image, and the size and shape of particles and the number of consecutive particles within a predetermined interval are measured to detect non-metallic inclusions. To do.

The surface of the sample to be measured is one that has been polished into a mirror-finished state by a polishing machine or that has been etched. These polishing machines may cause scratches on the sample surface. Alternatively, after polishing, washing with water, drying with hot air or wiping the surface with a solvent and setting on a microscope, water droplets may remain. In addition, water droplets may be dried and adhere to the sample to cause stains. In addition, dust may adhere to the sample during wiping.

It is necessary to separate such water drops, dust, scratches and the like from non-metallic inclusions, and the applicant of the present invention discloses these identification methods in Japanese Patent Laid-Open No. 8-68765. This method obtains a density profile on a line passing through a foreign substance that appears in a grayscale image of an object, and obtains the number of 0-point passing points of the first derivative to
Dirt, scratches, etc. are identified.

[0005]

However, although relatively large dust can be discriminated by the above method, dust such as abrasive grains having the same color tone (concentration) as non-metallic inclusions, non-metallic inclusions or bubbles can be detected. It was difficult to identify small holes that had come out.

The present invention has been made in view of the above-mentioned problems, and provides a foreign matter detection device for measuring non-metallic inclusions which distinguishes foreign matters such as dust and holes on the sample surface from non-metallic inclusions. With the goal.

[0007]

In order to solve the above-mentioned object, in the invention of claim 1, the surface of the sample is inclined upward or diagonally.
Lighting means to illuminate each separately and at different times
An image pickup means for picking up the reflected light of this illumination, and an upper illumination
The first grayscale image of the image captured by
Grayscale image generating means for obtaining a second grayscale image of the imaged image
And the first image obtained by binarizing the particles appearing in the first gray image
Image and the second image in which the particles appearing in the second gray image are binarized
And binarization means for obtaining the first particle and the first particle appearing in the first image
The second particles corresponding to the first particles appearing in the second image are
And the ratio of the area a1 of the first particles and the area a2 of the second particles.
a2 / a1 is a2 / a1 ≧ T1 with respect to the first determination value T1.
In this case, a particle determination unit that determines the first particles as dust is provided.

[0008]

[0009]

When the surface of the sample is illuminated from above and its image is taken, non-metallic inclusions, dust, holes and other particulate shapes are obtained. When the image is captured by illuminating it from an oblique direction, dust or particles having irregularities such as holes are wholly or partially illuminated. This upper illumination image is used as the first grayscale image, and the binarized image of all the particles including non-metallic inclusions is obtained in the binarized first image. Further, the oblique illumination image is used as the second grayscale image, and the binarized image showing the shape of the illuminated portion is obtained in the binarized second image of the particles. Although the non-metallic inclusions appear in the first image, they do not appear in the second image because they have no unevenness. This makes it possible to distinguish non-metallic inclusions from particles having a shining portion such as dust or holes due to unevenness. Dust has a large area of a shining portion in the whole area of one particle, and a hole occupies a small area of a shining portion. This difference makes it possible to distinguish dust from holes. Therefore, for the same particle, the area a1 of the first particle of the first image
Then, when the ratio a2 / a1 of the area a2 of the second particles of the second image is equal to or larger than the first determination value T1, it is determined as dust. In addition,
The judgment value T1 and the judgment values A and T2 described below are obtained by preparing a measurement sample from a metal material of the same lot as the metal material whose non-metallic inclusions are to be measured or a metal material manufactured by the same manufacturing method, and the magnification of the microscope. Is variously determined, and the determination value is determined from the area and area ratio on the image of the foreign matter determined by a person to be dust or a hole.

According to a second aspect of the present invention, the particle determination means is configured such that when the a2 / a1 is a2 / a1 <T1, the area a2 is a2 ≧ A with respect to the second determination value A, and a
When 2 / a1 is a2 / a1 ≧ T2 with respect to the third determination value T2, the first particles are determined to be holes.

In claim 1 or 2, a2 / a1 <T
Since it was determined that the particle of 1 was not dust, it is determined whether or not it was a hole. First, if the area a2 having a shining portion is equal to or larger than the second determination value A, there is a high possibility that it is a hole. Therefore, if the ratio a2 / a1 of a2 or A or more and the shining area a2 to the total area a1 is the third determination value T2 or more, the first particles are determined to be holes.

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

In the third aspect of the invention, the illumination means uses any one of a halogen lamp, a xenon lamp, a tungsten lamp and a mercury lamp as a light source.

As the illumination light source, a halogen lamp, a xenon lamp, a tungsten lamp, or a mercury lamp containing a wavelength in a wider range than a single wavelength such as a laser beam or a light in a narrow range is used. Images of inclusions, dust, and holes can be obtained at the same time, and dust that has risen on the sample surface, such as dust and holes, glows white, and dented holes have a specific color that is different from the interior. be able to.

According to a fourth aspect of the present invention, the illumination means sets the illumination from an oblique direction to an irradiation angle within 5 degrees from the horizontal.

In the case of illuminating from an oblique direction, in order for dust to effectively shine white, the oblique light should be within 5 degrees, preferably within 2 degrees with respect to the horizontal.

[0023]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the arrangement of an apparatus for realizing this embodiment. An image pickup lens is attached to the eyepiece tube of the microscope 1, and a color image pickup device 16 for picking up an image through the image pickup lens is attached. The stage 17 on which the measurement sample 20 is placed is subjected to plane position adjustment by a plane moving mechanism 18 that is moved back and forth and left and right by a pulse motor provided on the stand according to a signal from the auto stage driver 11, and a vertical moving mechanism 19 is performed by the auto focus driver 12. Actuate stage 17
Move up and down.

The A / D converter 2 converts the input data from the image pickup device 16 from analog to digital, and the input buffer 3 temporarily stores the digital data. The bus 4 transmits signals, the program memory 5 stores a program that defines the operation of the apparatus, and the CPU 6 controls the entire apparatus according to this program.

The image processor 7 performs grayscale processing, binarization processing, image analysis, etc. of the input image data, the grayscale image memory 9 stores the grayscale image data, and the binary image memory 10
Stores binary image data. The IHP processor 8 converts image data input as R, G, and B densities into lightness (I), hue (H), and saturation (P), and performs image processing. In response to an instruction from the CPU 6, the auto stage driver 11 controls the plane moving mechanism 18 to move the stage 17 on which the measurement sample 20 is placed in the X and Y directions to set the measurement position and area of the measurement sample 20. The autofocus driver 12 receives a control command from the CPU 6 to the vertical movement mechanism 19 and controls the vertical movement mechanism 19. The output buffer 13 temporarily stores the data to be output, the D / A converter 14 converts the output data from digital to analog, and the CRT 15 displays the output data on the screen.

2A and 2B are views showing a method of illuminating the measurement sample 20. FIG. 2A is a side view and FIG. 2B is a plan view. The light from the halogen lamp 21 is condensed by the condenser lens 22, bent at a right angle by the half mirror 23, and then the objective lens 2
The sample for measurement 20 is irradiated through the beam path 4. Light from another halogen lamp (not shown) becomes plane light having a width by the line fiber 25 through the optical fiber 26 and is applied to the measurement sample 20. The upper surface of the measurement sample 20 is set horizontally, and the plane light is emitted at an inclination angle of θ with respect to the horizontal plane. θ should be within 5 degrees,
The range of 0 to 2 to 3 degrees is preferable. Although the number of the line fiber 25 is one, the plurality of line fibers 25 may be provided around the measurement sample 20.

Next, the first embodiment will be described. In the present embodiment, a method of distinguishing dust and holes from an image of the measurement sample 20 having non-metallic inclusions, dust, holes, etc. will be described. An image of the three primary colors of red (R), green (G), and blue (B) is obtained by the color image pickup device 16, and one of these, for example, a grayscale image of G is used. If a monochrome imaging device is used, that image may be used. FIG. 3 shows a grayscale image obtained by simultaneously illuminating the measurement sample 20 from above and obliquely with the illumination device shown in FIG. 2 and this binarized image.
(A) is a grayscale image, (B) is a binarized image of (A),
(C) is a binarized image of a white glowing portion of the particles of (A). The binarized image of (B) is obtained by binarizing the grayscale image of (A) with a threshold value that reveals the contours of particles, and the binarized image of (C) reveals a white glowing part. It is obtained by binarizing with a threshold value. For this threshold value, the method disclosed in JP-A-7-234942 is used. This is determined by trial adjustment by a person for the first image, and then automatically corrects the device-determined threshold for images similar to the first image.

FIG. 3 (A) is a grayscale image of the measurement sample 20, and a, b and c show particles. “A” represents dust adhering to the surface and is shining white. Reference numeral b indicates a small hole generated by bubbles, which has a partially white glowing range. c is a non-metallic inclusion, and unlike a and b, has no shining range.
(B) is a binarized image obtained by binarizing the shape of each particle in (A), a1 corresponds to a, b1 corresponds to b, and c1 corresponds to c, and a1, b1, and c1 correspond to particles. The area is shown together with the name. (C) is a binarized image obtained by binarizing the shape of the white shining portion of each particle in (A), where a2 corresponds to particles a and b2 correspond to particles b, and a2 and b2 are white. The area of the shining part is also shown. Such binarization, calculation of the area of particles of the image, and discrimination between dust and holes described below are performed by the apparatus shown in FIG.

FIG. 4 shows two grayscale images obtained by illuminating the measurement sample 20 from above and obliquely at different times by the illumination device shown in FIG. 2 and the binarized image. (A) is a grayscale image illuminated by illumination from above, and (B) is a grayscale image illuminated by oblique illumination, which is a black image except for the white portion. (C) is the binarized image of (A),
(D) is the binarized image of (B). The binarized image of (C) is obtained by binarizing the grayscale image of (A) with a threshold value that reveals the contours of particles, and the binarized image of (D) is the grayscale image of (B). It is obtained by binarizing the image with a threshold that reveals the contours of the particles.

FIG. 4A is a grayscale image when the measurement sample 20 is illuminated only from above, and a10, b10, and c10 represent particles. a10 represents dust attached to the surface. b10 indicates a small hole generated by bubbles. c10 is a non-metallic inclusion. (C) is a binarized image in which the shape of each particle in (A) is binarized, and a11 is a10,
b11 corresponds to the particles of b10 and c11 correspond to the particles of c10, and a11, b11 and c11 represent the area of the particles as well as their names. (B) is a grayscale image when the measurement sample 20 is illuminated only obliquely, and a20 and b20 represent particles or a part thereof. The a20 is a dust adhering to the surface and appears as white. Reference numeral b20 denotes a portion of the small hole generated by the bubbles that is illuminated by oblique light and shines white.
(D) is a binarized image obtained by binarizing each particle in (B), where a21 corresponds to the particles of a20, b21 corresponds to the particles of b20, and a21 and b21 also represent the area of the white glowing portion. . Such binarization, calculation of the area of particles of the image, and discrimination between dust and holes described below are performed by the apparatus shown in FIG.

Below, the binarized images of FIGS. 3B and 3C or the binarized images of FIGS. 4C and 4D are used to detect dust and holes and distinguish between them. However, (B) of FIG. 3 and (D) of FIG. 4 or (C) of FIG. 3 and (C) of FIG. 4 may be used.

FIG. 5 is an operation flow chart of the first embodiment. In the description of this operation, the image of the particles shown in FIG. 3 will be described as an example. A color image (R, G, B) of the measurement sample 20 is input by the color imaging device 16, and 1 of these is input.
One, for example, a grayscale image of a G image is obtained (ST1). As a result, the grayscale image of FIG. 3A is obtained. Next, a binarized image of all the particles appearing in the grayscale image is generated and stored in the memory S1 (ST2). This binarized image is shown in FIG.
It is an image of (B). In addition, a binarized image of white-lighted particles and white-lighted portions of the particles appearing in the grayscale image is generated and stored in the memory S4 (ST3). This 2
The binarized image is the image of FIG.

Areas a1 and a of each particle of the binarized image stored in the S1 memory and the S4 memory, and a white glowing portion of each particle.
2, b1, b2, c1 are measured (ST4). Set the total number of particles to N and save the measured data in the data memory (ST
5). The nth particle is taken out (ST6), and the measurement data of the nth particle is read (ST7). Next, the area ratios a2 / a1 and b2 / b1 of the shining portion of each particle are obtained,
It is compared with the judgment value T1 (ST8), and if it is the judgment value T1 or more, it is judged as dust (ST9). Although such a determination is automatically performed by the apparatus shown in FIG. 1, the determination value T1 and the determination values A and T2 described below are determined by the operator or the like by the microscope 1 for some of the measurement samples 20. , It is set by checking the dust and holes by changing the magnification.

In ST8, the area ratio of the shining portion is a2 /
Although a close to 1 such as a1 is determined to be dust, if it is smaller than T1, for example, b, it is checked whether the area b2 of the further shining portion is the determination value A or more (ST
10) If A or more, it is checked whether the area ratio b2 / b1 of the shining portion is T2 or more (ST11), and if T2 or more, it is determined to be a hole (ST12). If less than A in ST10,
If it is less than T2 in ST11, it is determined that it is not a hole, and the routine proceeds to a check for non-metallic inclusions (ST13).
The inclusion determination routine is omitted because it is outside the scope of the present invention.
Japanese Patent No. 34190 discloses a routine for determining titanium. In this way, dust and holes are discriminated for one particle, and inclusions are discriminated. If n does not reach the total number N of particles (ST14), n is incremented by 1 (ST15), n + 1. About the 7th particle ST7 ~ 13
The process of the step is performed, the determination is performed up to the number N of all particles (ST14), and the process ends. As described above, the apparatus can discriminate dust and small holes that were conventionally difficult to discriminate from non-metallic inclusions.

When detecting and distinguishing dust and holes using the binarized images (C) and (D) of FIG. 4, the same procedure can be performed as in the flow chart of FIG. 5, but the ratio of the shining portion is a21 / a1
1, b21 / b11 may be used. Further, FIG. 3B
And (D) of FIG. 4, the ratio of the shining part is a2
When 1 / a1 and b21 / b1 are used and (C) of FIG. 3 and (C) of FIG. 4 are used, the ratio of the shining portion is a2 / a1.
1, b2 / b11 may be used.

Next, a second embodiment will be described. In the present embodiment, holes and dust are identified by hue. First, the color will be described. When an image is displayed in color, it is possible to separate and input the densities of the three primary colors of red (R), green (G), and blue (B), and synthesize and reproduce them. On the other hand, as a method of expressing colors, there is a method of expressing by lightness (I), hue (H), saturation or saturation (P), and conversion from display in three primary colors to display using lightness, hue and saturation. You can This is because when a color solid is considered, the designation of a specific color is to determine one of the points, so designating I means designating one position in the vertical direction of the color solid. The target color exists on the plane when the color solid is sliced horizontally in this place. Designating H here designates the angle from the reference axis when this plane is considered as rectangular coordinates with the achromatic color axis of the color solid as the origin. Therefore, the target color exists on the line segment representing this angle. This is because the designation of P further designates the diameter of a concentric circle centered on the achromatic color axis of the color solid, and one point on the line segment representing the angle is determined.

A technique for converting R, G, B data into I, H, P data has been disclosed, and is disclosed in, for example, Japanese Patent Publication No. 5-14944 by the applicant. The conversion to I and H used in this embodiment will be briefly described.
The conversion from R, G, B data to I is performed for each R, G,
It can be obtained by multiplying the data of B by an appropriate coefficient and then taking the sum. The appropriate coefficient is determined based on the human visibility and the difference in the reproduction level for each R, G, B of the apparatus.

Next, a method of converting R, G, B data into H will be described. As is well known, when a hue is expressed on a plane, it forms one ring. So if you put this ring on Cartesian coordinates, all the hues on Cartesian coordinates will be
It can be expressed as an angle from a reference axis, for example, the X axis. This state is shown in FIG. As can be seen from the example of the hue circle, a specific hue represented by three-dimensional data of R, G, B can be represented by one-dimensional data of an angle. Therefore, considering that the minimum value of R, G, B data is zero and the maximum value is normalized, the hue that can be expressed using R, G, B is the red color shown in FIG. 7B. (R), yellow (Y),
Hexagon with green (G), cyan (C), blue (B) and magenta (M) as vertices (this is the hexagonal shape drawn by the composite vector when considering the density data of R, G and B as vectors). Example) and its position is given in degrees from the reference axis. That is, when the R, G, and B density data are considered as vectors and placed on a plane at arbitrary angles (for example, at equal intervals), the hue is expressed as an angle from an arbitrary reference position indicated by the combined vector. .

The reference axis is the horizontal axis (A in FIGS. 7A and 7B).
X), the R vector is made to coincide with AX, and the G vector and the B vector are arranged at equal intervals, the converter from R, G, B to H may be configured as shown in FIG. it can. In FIG. 8, GX of the coefficient unit group is a horizontal component of G vector, GY is the same vertical component, BX is a horizontal component of B vector, BY is the same vertical component, RX is a horizontal component of R vector, and R component is a vertical component. There is no. These respective components are combined by adders 30 and 31 to create a horizontal component KX and a vertical component KY of the combined vector. KX and KY are converted into an angle of a composite vector through an arctangent calculator 32, and this angle represents a hue. This angle indicates Θ in FIG. 7 (A).

In this way, the R, G, and B images input from the color image pickup device 16 are converted into lightness I, hue H, and saturation P by the IHP processor. Illumination is performed only from above by the halogen lamp 21 shown in FIG. Figure 6
Indicates an image converted from RGB to hue H. “A” indicates dust particles, which are displayed as white dots. Symbol b represents a hole particle, and a yellow or yellowish green border b1 appears around the particle. This interior is often black. c is a nonmetallic inclusion. When the image is displayed in the hue H in this way, yellow or yellowish green shine appears around the hole. When the holes are small, the entire holes shine yellow or yellowish green, so that even small holes can be reliably identified. Note that this edging may not be continuous. Rather, intermittent yellow or yellow-green color often appears around. In such a case, it is also determined as a hole. In addition, when it is not yellow or yellowish green but pink, it is judged that it is not a hole but a titanium compound.

In FIG. 6, b represents the particles of the holes and the area of the particles, and b1 represents the area of the trimming. The edge area ratio b1 / b is calculated, and k is set as the judgment value.
When ≧ k, the accuracy of determination is improved by determining that the hole is a hole. The judgment value k is set in advance by the operator or the like as in the case described with reference to FIG.

Next, a third embodiment will be described. In the present embodiment, the second embodiment only illuminates from above, whereas the illuminating from oblique direction, the line fiber 25 shown in FIG.
Is the same as the others. By adding the oblique illumination, the dust particles a shown in FIG. 6 appear white and clear in the image, and the dust discrimination accuracy is improved. It should be noted that with illumination from only the oblique direction, only dust and relatively large holes will not appear in the image, and non-metallic inclusions and small holes will not be detected. Simultaneous irradiation of one is preferable.

[0043]

As is apparent from the above description, the present invention illuminates the measurement sample from above and obliquely at the same time or separately at different times, and collects dust from non-metallic inclusions depending on the area ratio of the bright portion. And the hole can be identified. Further, by illuminating from above, a color image can be obtained, and particles having a color different from the inside can be identified as holes. At this time, dust also appears in white and can be identified. By adding oblique illumination to the illumination from above, dust can be accurately identified.

[Brief description of drawings]

FIG. 1 is a block diagram of an apparatus implementing an embodiment of the invention.

FIG. 2 is a diagram illustrating an illumination method.

FIG. 3 is a diagram showing a grayscale image obtained by simultaneously illuminating from above and obliquely and two binarized images with different threshold values.

FIG. 4 is a diagram showing two grayscale images obtained by illuminating from above and diagonally at different times and each binarized image.

FIG. 5 is an operation flow chart of the first embodiment.

FIG. 6 is an image of the second embodiment in which R, G, and B images are hues H;
It is an image represented by.

FIG. 7A is a diagram illustrating a state in which a color represented in a hue circle is represented by an angle, and FIG. 7B is a hexagon drawn by a composite vector when R, G, and B are vectors representing the densities of respective colors. It is a figure explaining.

FIG. 8 is a circuit diagram of an R, G, B to H converter.

[Explanation of symbols]

6 CPU 7 Image processor 8 IHP processor 9 Gray image memory 10 binary image memory 16 color imaging device 21 Halogen lamp 25 line fiber

─────────────────────────────────────────────────── --- Continuation of the front page (56) References JP-A-3-158708 (JP, A) JP-A-8-68765 (JP, A) JP-A-3-221849 (JP, A) JP-A-7- 234942 (JP, A) JP 63-284455 (JP, A) (58) Fields investigated (Int.Cl. ⁷ , DB name) G01N 21/00-21/61 G01N 21/84-21/958 PATOLIS

Claims (4)

(57) [Claims] 1. An illuminating means for illuminating the surface of a sample separately from above and obliquely and at different times, an imaging means for capturing reflected light of this illumination, and a first image captured by the upward illumination. A grayscale image generating means for obtaining a grayscale image and a second grayscale image of an image picked up by oblique illumination;
Binarizing means for obtaining a first image in which the particles appearing in the grayscale image are binarized and a second image in which the particles appearing in the second grayscale image are binarized; the first particles appearing in the first image and the second image When the ratio a2 / a1 of the area a1 of the first particles and the area a2 of the second particles is a2 / a1 ≧ T1 with respect to the first determination value T1, for the second particles corresponding to the first particles appearing in the two images , First
A particle detection device for determining particles as dust, the foreign matter detection device for measuring non-metallic inclusions. 2. The particle determining means determines, when the a2 / a1 is a2 / a1 <T1, the area a2 is the second determination value A.
If case a and a2 / a1 of a2 ≧ A is the third determination value T2 relative to a2 / a1 ≧ T2 contrast, non according to claim 1, wherein the determining the first particles hole and Foreign object detection device for measuring metal inclusions. Wherein said illuminating means is a halogen lamp as a light source, a xenon lamp, a tungsten lamp, foreign object detection apparatus in the non-metallic inclusions measurement according to claim 1 or 2, characterized by using any of a mercury lamp. Wherein said illuminating means according to claim 1, characterized in that the irradiation angle within 5 degrees from the horizontal illumination from oblique
4. A foreign matter detection device for measuring non-metallic inclusions according to any one of 3 to 3 .