JP3435224B2 - Inspection equipment for non-metallic inclusions - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a nonmetallic inclusion inspection apparatus for inspecting the presence and type of nonmetallic inclusions in a metal material using an image processing technique. About. [0002] When non-metallic inclusions are present in a metal material, the mechanical properties of the metal material vary depending on the amount and type of the non-metallic inclusions. ,
It is very important to quantitatively analyze these nonmetallic inclusions. [0003] For this reason, ASTM (American) has been proposed as a method of inspecting nonmetallic inclusions present in a metal material.
Society for Testing Mater
Various methods have been developed, such as a method called the ials) method. [0004] In this case, the ASTM method is to observe non-metallic inclusions appearing on the surface of a sample collected from a metal material with a microscope and determine the quality of the metal material according to its shape and distribution. According to the shape and distribution state of the nonmetallic inclusions, the nonmetallic inclusions shown on the surface of the sample are classified into six types of A type, B type, C type, D type, TiB type and TiD type as shown in FIG. Or these A-series, B-series, C-series, D-series,
TiB-based and TiD-based are respectively Thin (thin) and He
The quality of the metal material is determined by classifying the metal material into an avy (thick type) or the like. [0005] By the way, the visual inspection by the inspector with the naked eye has problems in inspection speed and errors.
Inspection apparatuses for automatically inspecting metal materials by using image processing technology have been developed. FIG. 8 is a block diagram showing an example of an inspection apparatus for automatically inspecting the amount and type of nonmetallic inclusions in a sample cut from such a metal material. An inspection apparatus 101 shown in FIG. 1 has a holder 103 on which a sample 102 cut out of a metal material is set.
A microscope 104 for enlarging the sample 102 set in the holder 103, an illumination light source 105 for illuminating the sample 102 set in the holder 103 so as to be coaxial with the optical axis of the microscope 104, The microscope 1
An ITV camera (industrial television camera) 106 for converting an optical image of the sample 102 enlarged by the camera 04 into an electric signal (image signal), and an image processing for processing the image signal of the sample 102 obtained by the ITV camera 106 Device 107, a host computer device 108 for controlling the image processing device 107, and a monitor device 10 for displaying the processing results and the like obtained by the image processing device 107
9 is provided. When a sample 102 cut out from the metal material under a predetermined condition is set in a holder 103 in order to inspect nonmetallic inclusions contained in the metal material to be inspected, While illuminating the sample 102 with the illumination light source 105, the microscope 104 enlarges the optical image of the sample 102, and the ITV camera 106 converts the optical image enlarged by the microscope 104 into an electric signal (image signal). This is processed by the image processing device 107, and the processing result is supplied to the host computer device 108 to analyze the shape and concentration of the non-metallic inclusions, and to analyze the analysis result, the processing result, the image of the sample 102, and the like. The screen is displayed on the monitor device 109. In this case, the image processing device 107 performs A / D conversion (analog / digital conversion) of an image signal output from the ITV camera 106 for each pixel and converts predetermined bits, for example, 8-bit image data. A / D converter circuit 110 to be created and this A / D converter circuit 11
A frame memory circuit 111 for temporarily storing image data output from 0 and image data stored in the frame memory circuit 111, and the image data is stored at a preset threshold value. A CPU circuit 112 for binarizing the image to produce a binarized image in which pixels having non-metallic inclusions are brightened, and converting an image signal output from the ITV camera 106 into an A / D signal for each pixel; After the image data is created by the conversion, the image data is binarized based on a preset threshold value to create a binarized image, and the binarized image is supplied to the host computer device 108. Device 109
To supply. As described above, in the inspection apparatus 101, after the optical image of the sample 102 is converted into an electric signal (image signal), the image signal is A / D converted and converted into image data, which is set in advance. Based on the threshold value, this is binarized to detect non-metallic inclusions, so that the amount of non-metallic inclusions contained in the sample 102 can be detected without visual inspection by an inspector or the like. And its type can be automatically detected. By the way, in such an inspection apparatus 101, each pixel constituting the image data of the sample 102 by the threshold value is defined as a pixel of a metal part and a pixel of a non-metallic inclusion part. The setting of the threshold directly affects the detection accuracy of nonmetallic inclusions. For this reason, when such an inspection apparatus 101 is used, a sample 1 serving as a reference is placed on a monitor 109.
02 while displaying the binarized image of 02, the inspector or the like adjusts the threshold by trial and error to find the threshold that gives the most optimal binarized image.
The threshold was set. However, in such a threshold value setting method, not only is it difficult to find an optimum threshold value, but also when each of the non-metallic inclusions in the sample 102 is actually inspected. Of the non-metallic inclusions whose types are determined based on the brightness of the pixels, for example, as shown in FIGS. Inspection device 101
The result of the inspection may be shifted. According to the present invention, in view of the above circumstances, the threshold value can always be set to an optimum value, and the image of the nonmetallic inclusion in the sample image can be accurately distinguished from the image of the metal. ,
It is an object of the present invention to provide a nonmetallic inclusion inspection apparatus capable of accurately inspecting the quantity and type of nonmetallic inclusions. [0015] In order to achieve the above object, the present invention is directed to a photographing apparatus for photographing a sample to be photographed, and an image processing apparatus for acquiring an image signal obtained by the photographing apparatus. In the inspection apparatus for non-metallic inclusions, which detects the amount of non-metallic inclusions and / or the type of the non-metallic inclusions in the image of the sample indicated by the image signal, the width and length directions of the size of non-metallic inclusions was measured in pixels, this
Is the width direction or length direction of the non-metallic inclusion
If the number of pixels in the direction increases, the preset reference
Adjust the threshold value so that it is smaller, and
Pixel in the width or length direction of metal inclusions
The threshold value increases.
And so that adjusting the image signals to optimize the threshold for identifying each non-metallic inclusions to be used for binarization, using this optimized threshold, the image The signal is binarized to check the amount and type of the non-metallic inclusion. In the above arrangement, a sample to be inspected is photographed by a photographing device, and an image signal obtained by the photographing device is binarized by an image processing device, and the image signal is represented by the image signal. The amount of non-metallic inclusions in the image of the sample, in a non-metallic inclusion inspection device that detects at least one of the type, the size of the non-metallic inclusions in the sample is measured in pixel units Based on the measurement result, a preset reference threshold value is adjusted to optimize a threshold value used in binarizing the image signal, and the optimized threshold value is By using the image signal to binarize and inspect the amount and type of the non-metallic inclusion, the set value of the threshold is always set to an optimal value, and the image of the non-metallic inclusion in the sample image is obtained. And the metal image and the positive Identify accurately, and thereby accurately inspect the amount and type of non-metallic inclusions. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS << Method for Setting Threshold Value of Nonmetallic Inclusion Inspection Apparatus According to the Present Invention >> First, prior to detailed description of a nonmetallic inclusion inspection apparatus according to the present invention, a nonmetallic inclusion apparatus according to the present invention will be described. A method of setting a threshold value of the object inspection device will be described. First, as in the inspection apparatus for non-metallic inclusions according to the present invention, a method of A / D converting an image signal obtained by an ITV camera into image data in a digital signal format, and then processing the same. In this method, the surface to be inspected of the sample to be inspected is decomposed into pixels, which are the minimum decomposition units when processing an image, and then the density of each pixel is processed. The density of each pixel is a value obtained by averaging the density of metal or non-metallic inclusions in the area of each pixel for each pixel. Since the pixel density of the non-metallic inclusions is high, even if each pixel includes an image of the non-metallic inclusions, even if each pixel contains an image of the non-metallic inclusions in each pixel, these pixels are determined according to the ratio of the metal image to the non-metallic inclusions. When the density of each pixel changes and a histogram of each pixel is created, the density of this pixel decreases as the ratio of nonmetallic inclusions in the area of one pixel decreases. However, in the conventional threshold value setting method, when setting the threshold value, a histogram of each pixel constituting the reference sample image is created, and based on this histogram, each pixel is set to the pixel of the metal part. , And a threshold value for separation into pixels of non-metallic inclusions is obtained, and the threshold value obtained from the histogram is obtained regardless of the number of pixels occupied by non-metallic inclusions as shown in FIG. Non-metallic inclusions are identified using fixed values, so it is not necessary to detect non-metallic inclusions when the width or length is small. The threshold is higher than the original value.
It will be a small value. In view of the above, in the inspection apparatus for non-metallic inclusions according to the present invention, the threshold value obtained from the histogram as shown in FIG. Each time the width is reduced, and each time the length is reduced, by increasing the threshold value, the metal and the non-metallic inclusions can be accurately identified, and
The type of each nonmetallic inclusion, for example, an A-based inclusion indicated by a white circle and a C-based inclusion indicated by a black circle can be accurately distinguished. Then, based on this threshold setting principle,
When a sample containing non-metallic inclusions, for example, a sample containing A-based inclusions or C-based inclusions was inspected by actually adjusting the threshold value according to the number of pixels occupied by the non-metallic inclusions, FIG. , FIG.
As shown in the figure, the result of the visual inspection by the inspector and the result of the inspection by the inspection device could be matched. << Description of the Preferred Embodiment >> A non-metallic inclusion inspection apparatus according to the present invention using the above-described principle of setting the threshold will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing an embodiment of the inspection apparatus for nonmetallic inclusions according to the present invention. The inspection apparatus 1 for non-metallic inclusions shown in FIG. 1 includes an image input mechanism 2 and a processing mechanism 3. The image input mechanism 2 takes an image of a sample 4 to generate an image signal, and performs processing. After processing the image signal by the mechanism 3 and determining a reference threshold value necessary for detecting the non-metallic inclusions, the width and length of each non-metallic inclusion and the density thereof are measured in pixel units. According to the width and length of the non-metallic inclusion, the reference threshold is set to an optimum threshold, and the amount and type of the non-metallic inclusion in the sample 4 are inspected. The inspection result is displayed on a screen or printed out. The image input mechanism 2 includes a holder 5 on which a sample 4 cut out of a metal material is set, and a holder 5
A microscope 6 for enlarging the sample 4 set in the holder 6, an illumination light source 7 for illuminating the sample 4 set in the holder 5 so as to be coaxial with the optical axis of the microscope 6, and holding the holder 5 Meanwhile, the XY stage 8 is moved in the X direction and the Y direction to bring the inspection target area of the sample 4 into the field of view of the microscope 6 and a stage control command output from the processing mechanism 3. XY stage 8
XY stage controller device 9 for controlling the operation of
An autofocus controller device 10 that adjusts the focus, magnification, and the like of the microscope 6 based on a focus control command output from the processing mechanism 3 to focus on the surface of the sample 4; And an ITV camera 11 for converting the optical image of the sample 4 into an electric signal (image signal). When a sample 4 cut out from the metal material under predetermined conditions is set in a holder 5 in order to perform inspection of non-metallic inclusions contained in the metal material to be inspected, While illuminating the sample 4 with the illumination light source 7, the XY stage 8 is operated based on the stage control command and the focus control command output from the processing mechanism 3, and the inspection target of the sample 4 is within the visual field of the microscope 6. The region is arranged, and in this state, the microscope 6 focuses on the surface of the inspection target region and enlarges the surface. The ITV camera 11 converts the optical image enlarged by the microscope 6 into an electric signal (image signal). And supplies it to the processing mechanism 3. The processing mechanism 3 takes in the information processing device 12 for controlling the entire inspection apparatus 1 and processing the measurement data, and the measurement data processed by the information processing device 12, and the like.
A printer device 23 for printing this, an image processing device 13 for taking in an image signal output from the ITV camera 11 based on an instruction from the information processing device 12 and performing necessary image processing, An image monitor device 14 for taking in an output image signal (raw image signal obtained by the ITV camera 11) and displaying it on a screen, and an image-processed image signal output from the image processing device 13 And a video printer device 16 for capturing a screen signal output from the image processing device 13 and hardcopying it. Then, the XY stage controller 9 and the autofocus controller 10 of the image input mechanism 2 are controlled so that the sample 4
After taking an image of the inspection area of interest and taking in the image signal obtained thereby, and processing this image signal to determine a reference threshold necessary for detecting non-metallic inclusions, each non-metallic The width and length of the inclusion and the concentration thereof are measured in pixel units, and the reference threshold is set to an optimum threshold in accordance with the width and length of the non-metallic inclusion. The amount and type of the non-metallic inclusions are inspected, and the inspection result is displayed on a screen or printed out. In this case, as shown in FIG. 2, the image processing device 13 performs A / D conversion of an image signal output from the ITV camera 11 and converts the image signal into image data. A frame memory circuit 18 for taking in and storing image data output from the D converter circuit 17, an image processing dedicated processor circuit 19 for performing image processing on the image data stored in the frame memory circuit 18, and an image processing dedicated processor A control CPU circuit 20 for controlling the operation of the circuit 19 and the A / D converter circuit 17 and the like, and a work RAM circuit 21 used as a work area for the control CPU circuit 20 and the like
And an interface circuit 22 that supports data transfer between the control CPU circuit 20 and the information processing device 12 or the like. Then, based on an instruction from the information processing device 12, an image signal output from the ITV camera 11 is fetched and supplied to the monitor device 14 to be displayed on a screen. A / D conversion and conversion into image data, image processing, generation of a histogram, and supply of the histogram to the information processing device 12, based on the image data, the shape (width and length) of the non-metallic inclusion ) Is measured and supplied to the information processing device 12. Based on the threshold value output from the information processing device 12, the image data is binarized to determine a pixel having non-metallic inclusions. A process of creating a bright binary image,
A process of supplying the binarized image to the monitor device 15 to display it on a screen, a process of supplying the binarized image to the video printer device 16 to make a hard copy, and the like are performed. At this time, the information processing device 12 calculates a reference threshold value for determining the optimum threshold value based on the histogram of the reference sample 4 output from the image processing device 13, Each time the shape information of the non-metallic inclusion obtained by performing image processing on the actual sample 4 to be inspected is supplied from the image processing device 13, the non-metallic According to the number of pixels in the width direction and the number of pixels in the length direction occupied by the inclusion, as shown in FIG. The reference threshold value is adjusted so as to be smaller, an optimum threshold value is calculated, and the calculated optimum threshold value is supplied to the image processing device 13 to binarize image data of an actual sample to be inspected. Based on the binarized image data, The amount of non-metallic inclusions are on the 4 on the surface and to identify its type. << Effects of Embodiment >> As described above, in this embodiment, the sample 4 is photographed by the image input mechanism 2 to generate an image signal, and the image signal is processed by the processing mechanism 3 to produce a non-metallic image. After determining the reference threshold necessary for detecting inclusions, the width and length of each nonmetallic inclusion and its concentration are measured in pixel units, and the width and length of this nonmetallic inclusion are determined. Correspondingly, the reference threshold value is set to an optimum threshold value, and the amount and type of the nonmetallic inclusions in the sample 4 are inspected. With the optimum value, the image of the non-metallic inclusions in the sample image can be accurately distinguished from the image of the metal, so that the amount and type of the non-metallic inclusions can be accurately inspected. . As described above, according to the present invention, the threshold value is always set to an optimum value, and the image of the non-metallic inclusion and the image of the metal in the sample image are compared. Accurate identification can be made, and thereby the amount and type of non-metallic inclusions can be accurately inspected.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram showing one embodiment of a nonmetallic inclusion inspection apparatus according to the present invention. FIG. 2 is a block diagram illustrating a detailed configuration example of the image processing apparatus illustrated in FIG. 1; FIG. 3 is a diagram showing an example of a threshold used in a conventional inspection device. FIG. 4 is a diagram showing an example of a threshold used in the inspection apparatus for nonmetallic inclusions according to the present invention. FIG. 5 is a diagram showing an example of the relationship between the evaluation of the A-based inclusion and the visual evaluation when the threshold value is adjusted. FIG. 6 is a diagram showing an example of a relationship between evaluation of C-based inclusions and visual evaluation when a threshold value is adjusted. FIG. 7 shows the types of nonmetallic inclusions contained in the metal,
It is a figure for explaining the judgment method. FIG. 8 is a configuration diagram showing an example of a conventional inspection apparatus for automatically inspecting the amount and type of nonmetallic inclusions in a sample cut out of a metal material. 9 is a diagram showing an example of the relationship between the evaluation of A-based inclusions by the inspection device shown in FIG. 8 and the visual evaluation. 10 is a diagram showing an example of the relationship between the evaluation of C-based inclusions by the inspection device shown in FIG. 8 and the visual evaluation. [Description of Signs] 1 Inspection device for nonmetallic inclusions 2 Image input mechanism 3 Processing mechanism 4 Sample 5 Holder 6 Microscope 7 Illumination light source 8 XY stage 9 XY stage controller device 10 Autofocus controller device 11 ITV camera ( 12) Information processing device 13 Image processing device 14 Image monitor device 15 Image monitor device 16 Video printer device 17 A / D converter circuit 18 Frame memory circuit 19 Image processing dedicated processor circuit 20 Control CPU circuit 21 Work RAM circuit 22 Interface circuit 23 Printer

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Tsukasa Hamada 2nd Nadahama-Higashicho, Nada-ku, Kobe City, Hyogo Prefecture Inside Kobe Steel, Ltd. (72) Inventor Satoshi Matsushita 66-2 Horikawa-cho, Saiwai-ku, Kawasaki City, Kanagawa Prefecture (56) References JP-A-55-94147 (JP, A) JP-A-2-52251 (JP, A) (58) Fields studied (Int. Cl. ⁷ , DB name) G01N 21 / 84-21/958 G01N 33/00-33/46

Claims (1)

(57) [Claim 1] A sample to be imaged is photographed by a photographing device, and an image signal obtained by the photographing device is binarized by an image processing device. In a non-metallic inclusion inspection apparatus for detecting the amount and / or type of the non-metallic inclusions in the image of the sample shown, the width direction and the length of the non-metallic inclusions in the sample the direction of <br/> size measured in pixels, the measurement result is the non-metallic
Many pixels occupy the width or length direction of the inclusion
Then, the preset reference threshold value will be reduced.
Adjustment, the measurement result is the width direction of the non-metallic inclusions or
Is preset if less pixels occupy the length direction.
The threshold value for identifying each non-metallic inclusion used in binarizing the image signal is optimized by adjusting the reference threshold value to be large. A non-metallic inclusion inspection apparatus, comprising: binarizing the image signal using a threshold to inspect the amount and type of the non-metallic inclusion.