JPH0470555A - Apparatus for inspecting surface of sphere - Google Patents

Abstract

PURPOSE:To also detect a flat flaw other than a surface flaw and material inferiority with high accuracy by detecting the area and quantity of light of a regular reflection region from the surface image of a sphere. CONSTITUTION:A feed rotary part 2 rotates a sphere 1 under the control of a control part 3 to oppose the entire surface thereof to an inspection part 4. The light from the light source 5 arranged on the lateral side of a coaxial vertical illumination apparatus 6 is turned downwardly at a right angle by the apparatus 6 to be projected on the surface of the sphere 1. The reflected light from the surface of the sphere 1 is two-dimesionally taken imagewise by an imaging camera 7 through the apparatus 6 and the image signal is outputted to an image processing apparatus 8. This signal is subjected to masking processing by a masking processing apparatus 13 and, thereafter, a flat flaw is detected by a flat flaw detection apparatus 17. After masking processing 13, a difference signal due to a definite threshold value is binarized by a binarizing device 15 in a brightness correcting apparatus 14 and the upper and lower parts of an area are removed by an area measuring apparatus 16 to judge the presence of a surface flaw. Further, the sum total of density is compared with a set value in a density measuring apparatus 18 from the apparatus 14 to detect the presence of material inferiority.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is mainly concerned with determining the presence or absence of surface scratches (including flat scratches) or surface roughness (fabric defects) of ceramic spheres used for balls, rings, etc. This invention relates to a device for inspecting the suitability of [Prior Art] Normally, it is essential that a sphere for pole hair ring be perfectly spherical in shape in order to reduce energy loss due to friction as much as possible, and that there be no surface scratches. Therefore, in order to detect surface scratches on spheres, for example, in the case of steel spheres, various devices have been proposed that detect changes in the amount of light reflected from the mirror-finished surface to detect the presence or absence of scratches. ing.

Figure 12 is a schematic diagram of a conventional spherical surface flaw inspection device, Figure 13
The figure is an explanatory diagram showing the locus of the scanning spot light that moves due to the rotation of the sphere 31 during inspection. Reference numeral 32 indicates a spherical object, 32 a transport rotating section, and 34 an inspection section.

The conveyor rotation unit 32 supports the sphere 31 on a plurality of rolls and supports the sphere 31 on a first roll so that the entire surface of the sphere 31 faces the inspection unit 34.
As shown in FIG. 2, the sphere 3j is configured to be able to rotate within a vertical plane and within a horizontal plane, respectively.

On the other hand, the inspection section 34 includes a light source 41, a reflector 42, and a sensor 43.
As shown in FIG. 13, a spot light having a diameter of approximately 1 fl is made incident on the surface of the rotating sphere 31 at right angles or at a predetermined angle as shown in FIG. A spot light having a diameter of about one crane is scanned on the surface as shown by an arrow 31a, and the reflected light from each position on the surface of the sphere 31 is captured by a sensor 43 to detect whether or not there is a change in the light amount with respect to the standard light amount determined in advance. It looks like this.

That is, since the surface of the sphere 31 has a mirror finish, light close to total internal reflection is reflected from the undamaged surface of the sphere 31, but if there is a scratch on the surface, the light is diffusely reflected by the scratch, and is captured by the sensor 43. Since the amount of light remains small, the presence or absence of scratches can be detected by detecting this change in the amount of light.

In such a conventional device, the means for detecting changes in the amount of reflected light can be configured with a relatively simple analog circuit or digital circuit, and if a spot light with a diameter of about 1 crane is used, the rotation of the sphere 31 It is now possible to inspect at a rate of about 3 pieces/second in real time by increasing the speed of the number (3000 revolutions/second).

[Problem to be solved by the invention]

By the way, in recent years, ceramic spheres, which have superior heat resistance and durability compared to copper spheres, have begun to be used, but such ceramic spheres not only have surface scratches but also have partially flat surfaces. The presence or absence of flat scratches, as well as the roughness of the surface due to poor polishing (or poor gloss (referred to as fabric defects))
It is also necessary to detect the presence or absence of.

However, this ceramic sphere has a lower reflectance than a steel sphere, so the change in light intensity due to surface scratches is small, making it impossible to obtain a sufficient S/N ratio. The color of the scratches will be white or black, and in the case of white scratches, the reflectance will increase due to the presence of scratches, and simply capturing the change in light amount between specular reflection and diffuse reflection will not be enough to detect the surface. Difficult to detect scratches.

Therefore, a method was developed to obtain a spherical surface image by capturing the reflected light from the surface of a standard sphere, which usually has no surface scratches, and to detect changes in the brightness of each part of the spherical surface image. However, if the detection ability changes due to changes in the surface condition of the material of the sphere or changes in the amount of light due to aging of the light source that projects light onto the surface of the sphere, stable inspection performance cannot be obtained, and the above-mentioned flatness There has been a problem in that sufficient detection accuracy cannot be obtained for the presence or absence of scratches or fabric defects.

The present invention was made in view of the above circumstances, and its purpose is to provide a spherical surface detection device that can detect not only surface scratches and fabric defects, but also flat scratches with high precision. There is something to do.

[Means to solve the problem]

A spherical surface flaw inspection device according to the present invention: means for capturing an image of the spherical surface while irradiating the rotating spherical surface with light; and means for detecting surface flaws by determining the amount of specularly reflected light.

[Effect]

According to the present invention, the specular reflection area is captured from the spherical surface image, the area light amount of the specular reflection area is detected, and the presence or absence of a flat flaw can be identified.

〔Example〕

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be specifically described below based on drawings showing embodiments thereof.

FIG. 1 is a schematic diagram showing a spherical surface inspection apparatus according to the present invention, in which 1 indicates a spherical body, 2 indicates a conveying rotation section, and 4 indicates an inspection section.

The conveyor rotating section 2 is configured to rotate the sphere 1 under the control of the controller 3 so that its entire surface faces the inspection section 4 side. The controller 3 outputs the control signal to the transport rotation section 2 and also to the image processing device 8 in the inspection section 4.

The inspection unit 4 includes a light source 5 that emits parallel (or non-parallel) light rays.
, a coaxial epi-illumination device 6 composed of a half mirror, etc., an imaging camera 7, and an image processing device 8. FIG. 2 is an explanatory diagram showing the arrangement of the optical system. 5 is disposed on the side of a coaxial epi-illumination device 6 placed directly above the conveyor rotation unit 2, and the light from the light source 5 is turned downward at right angles by the coaxial epi-illumination device 6 and is projected onto the sphere 1. Ru. The reflected light from the sphere 1 enters the imaging camera 7 via the coaxial epi-illumination device 6, where it is imaged two-dimensionally, and the image signal is sent to the image processing device 8.
It is now output to .

From the image processing device 8, the image pickup camera 7 and the system? A control signal is synchronously outputted to the Il device 3, and the sphere 1 is rotated by a dimension corresponding to the previously imaged area in its rotational direction so that the imaging positions with respect to the sphere surface do not overlap. At this time, the next captured image is set to be captured from the imaging camera 7.

FIG. 3 is a block diagram of the image processing device 8, and the image signal input from the imaging camera 7 to the image processing device 8 is first
/D (analog/digital) converter 11,
Here, it is output as a quantized digital value (image data) to the image memory 12 and stored there. The image data input to the image memory 12 is processed by a mask processing device 13.
2 as shown in Figures 4(a) and 5(al).
After being subjected to various types of mask processing, the image data subjected to mask processing as shown in FIG. The image data is output to a brightness correction device 14 for detecting surface flaws and fabric defects.

FIG. 4(a) is an explanatory diagram showing image data that has been masked to detect flat flaws, and is a circular flaw inspection area a (
This is image data from which signals in parts other than the area shown with hunting have been removed. FIG. 5 fa) is an explanatory diagram showing image data subjected to mask processing for detecting surface flaws and fabric defects, and of the image data input via the image memory 12, the image data centered on the center of the spherical surface image is shown in FIG. The non-inspection areas e and f, excluding the annular flaw inspection area (the area shown with hunting) d, have been removed by mask processing.

In FIG. 4, the flaw inspection area a is made circular with the center of the spherical surface image as its center because in the case of a flat flaw, specular reflection occurs only when the flat flaw directly faces the coaxial epi-illumination device 6.1. At this time, the amount of reflected light to be imaged becomes extremely large.

Furthermore, the reason why region b is a non-inspection region is that it is a diffuse reflection region regardless of the presence or absence of a flat flaw, and the presence of a flat flaw hardly causes a change in the amount of light. The size of the specular reflection area on a standard sphere without scratches inside the scratch detection area a is the range shown in Figure 4 (bl), and the specular reflection area when there is a flat scratch is as shown in Figure 4 (C). As shown, since the area is wider, it is sufficient that the area is at least wider than the area shown in FIG. 4(b).

On the other hand, in Figure 5 (al), e and f are non-inspection areas because the non-inspection area e has an extremely large amount of specularly reflected light compared to the amount of diffusely reflected light, and the non-inspection area f has an extremely large amount of specularly reflected light compared to the amount of specularly reflected light. This is because the amount of diffusely reflected light is extremely large, and the S/N ratio for detecting the presence or absence of surface flaws is small and accurate inspection cannot be expected.

Note that the flaw inspection areas a, d, non-inspection areas s, ef vary depending on the type, size, etc. of the sphere 1, so it goes without saying that they are appropriately set according to the sphere 1.

(Flat flaw detection) The image data as shown in Figure 4 (al) is the flat flaw detection device 1.
7, the flat flaw detection device 17 applies a predetermined density ( Bright', i) Binarize with the dark value to obtain a binarized box image, and based on this, find the area of the specular reflection site and find the total amount of light from this specular reflection site.

The area of the specular reflection part and the amount of light reflected from this part are calculated in advance and compared with the area of the specular reflection part and the amount of light reflected from there in the case of a standard sphere, and the same as in the case of the standard sphere,
Or when it is larger than this, it is determined that there is a flat flaw, and a flat flaw detection signal is output.

Meanwhile, the image data as shown in FIG. 5(a) outputted from the mask processing device 13 is as shown in FIG.
).

It has a brightness distribution as shown in (d). Figure 5(b)
(c+, (dl) are both shown with the distance from the optical axis of the imaging camera 7 on the horizontal axis and the brightness on the vertical axis.

As is clear from this graph, the brightness of the part without scratches (5th row)) and the brightness of the part containing white scratches (Figure 5 (C))
Since the surface of the sphere 1 is a three-dimensional curved surface, the brightness of the part containing the black scratch (Fig. 5(d)) has a brightness distribution in which the brightness decreases as the distance from the center increases. There is no need to perform brightness correction in the process of detecting flat flaws unrelated to brightness, but in the process of detecting surface flaws and fabric defects, accurate detection of brightness cannot be achieved even if the data is binarized using a predetermined value. Therefore, the brightness correction device 14 corrects the difference in the amount of reflected light due to the three-dimensional curved surface of the sphere 1.

FIG. 6 is a block diagram showing the brightness correction device 14, in which 21 is a calibration device, 22 is a standard image storage device, 23 is an image memory, and 24 is a subtracter.

The operations of the brightness correction device 14 during calibration performed in the preparation stage and during inspection will be described based on the flowchart shown in FIG. 7.8.

l) At the time of calibration FIG. 7 is a flowchart showing the main process at the time of calibration, in which the brightness correction device 14 collects image signals of the surfaces of spheres of different sizes on the sides of various materials without scratches obtained in advance by the imaging camera 7. After capturing the image without being subjected to mask processing by the mask processing device 13 and storing it in the image memory 23 (step Sl), the size of each sphere 1 is detected based on the image signal stored in the image memory 23. S2) Based on this size, the non-inspection area is
Find the size of f (step S3) and input it to the mask processing device 13 (step S4).
In step 515), a brightness standard image is created for the inspection area d, and this is stored in advance in the standard image storage device 22. The information is stored in memory (step S6).

Of course, it is also possible to select unblemished spheres 1 for each loft at the time of manufacture of each sphere 1, or more frequently prior to the inspection, obtain a standard image of their brightness, and perform the inspection using this. Needless to say.

ii) At the time of inspection When starting the inspection, a brightness standard image corresponding to the material and size of the sphere 1 to be inspected is selected and transferred from the image storage device 22 to the image memory 23. Next, an imaging camera 7 captures the reflected light from the sphere 1 placed on the transport rotating section 2.
The image is taken at a predetermined timing and taken into the image processing device 8 of the inspection section 4, and the image signal is sent to the A/D conversion section 11,
After passing through the image memo January 2 and performing mask processing through the mask processing device 13, the image signal is input to the subtracter 24 of the brightness correction device 14, and the subtracter 24 converts the image signal from the image signal that has been manually subjected to the mask processing. The signal of the brightness standard image stored in the memory 1 is subtracted (step 521), and the difference signal is output to the binarizer 15 for detecting surface flaws and the density measuring device 17 for detecting the presence or absence of fabric defects. (step 522
).

Up to this point, the surface flaw detection process and fabric defect detection process are the same, but from this point onwards, they are processed separately.

[Surface flaw detection]

Surface flaws are detected by passing the output from the brightness correction device 14 through a binarizer 153 and an area measuring device 16.

The binarizer 15 binarizes the image difference signal using a certain threshold value, and extracts the flaw signal within the inspection area a. However, the flaw signal contains false detection signals due to electrical noise.
The area measuring device 1 measures the area where each of the obtained flaw signals occurs.
6, the area on the image is determined, and a flaw signal whose area is less than a certain value is removed as minute noise, and a region occupying a constant area is similarly removed as noise, and the presence or absence of surface flaws is finally determined.

[Fabric defect detection]

The presence or absence of fabric defects on the surface of the sphere is determined by the concentration measuring device 18. FIG. 8 is a block diagram showing a specific configuration of the density measuring device 18, in which image data after brightness correction from the brightness correction device 14 is input to the subtracter 25. Subtractor 25
subtracts the density of each corresponding part of the brightness standard image of the standard sphere stored in the storage device 22 from the density of each part (for example, for each pixel) of the brightness correction image, and adds the density difference image for each part. Output to 27. The adder 27 calculates the sum of the densities for each portion and outputs this to the fabric defect detector 28. The fabric defect detector 28 compares the sum of the densities with a predetermined set value, and outputs a signal indicating that there is a fabric defect if it exceeds this value, and outputs a signal indicating that there is no fabric defect if it does not exceed the reference value. It has become.

If there is no flaw signal or fabric defect signal, the next image is captured in accordance with the control signal sent from the image processing device 8 to the controller 3 of the transport rotation unit 2 so that the imaging positions do not overlap, and the process is carried out in the same manner as described above. Detects flaws and fabric defects. If no flaws are detected, continue inspecting the entire circumference of the sphere 1,
If no flaws are detected even then, the product is determined to be non-defective.

When a flaw signal or fabric defect signal is output, the inspection is stopped even during the inspection, replaced with the next sphere 1, and the inspection is continued.

The image is taken into the image processing device 8 in the process shown in FIG. FIG. 10 is a flowchart showing the timing of manual image operation, in which the conveyance rotation unit W2 is controlled through the controller so that the rotation speed of the sphere 1 becomes a predetermined value, and the rotation speed at that time is detected. Step 532: It is determined whether the surface of the sphere 1 located within the field of view of the camera 7 does not overlap with the position of the previously input image.
), if they overlap, the process returns to step S31 and detects the number of revolutions since the image was captured by light again, and if it is determined that the positions do not overlap, a trigger signal is output from the image input trigger output device ( Step 533).

As a result, the image processing device 8 captures the image from the imaging camera 7 and performs image processing (step 534).

It is determined whether or not the entire surface of the sphere 1 has been inspected (step 535); if not, step S
Returning to step 31, the above-described process is repeated, and when the process is completed, this routine is ended.

FIG. 11 is an explanatory diagram showing the locus of the scanning light on the surface of a sphere when the device of the present invention is used, and it is clear from the comparison with the conventional case shown in FIG. In contrast to scanning with a spot light having a radius of about 100 yen, the device of the present invention scans with a spot light having a radius of the radius of the sphere 1 or close to it, so the inspection area is wide, and the inspection area is wider than that of the conventional device. While it is necessary to rotate the sphere several hundred times to complete the process, the device of the present invention only requires a few rotations, which is efficient.

〔effect〕

As described above, the present invention has excellent effects such as being able to efficiently inspect the presence or absence of flat scratches on a sphere used for ball hair rings and the like.

[Brief explanation of the drawing]

Fig. 1 is a schematic diagram of the device of the present invention, Fig. 2 is an explanatory diagram of the optical system in the device of the present invention, Fig. 3 is a block diagram of the image processing device, and Fig. 4 is a diagram of flat flaw detection by the mask processing device. An explanatory diagram showing the mask processing area for
An explanatory diagram showing the mask processing area for detecting fabric defects, Fig. 6 is a block diagram of the brightness correction device shown in Fig. 3, Fig. 7 is a flowchart showing the main process during calibration, and Fig. 8 is during inspection. Figure 9 is a block diagram of a density measuring device for detecting fabric defects; Figure 10 is a flowchart showing the main processes of
The figure is a flowchart showing the process of capturing images from the imaging camera to the image processing device, FIG. 11 is an explanatory diagram showing the inspection trajectory of the spherical surface in the device of the present invention, FIG. 12 is a schematic diagram of the conventional device, and FIG. 13 is the inspection FIG. 3 is an explanatory diagram showing the locus of a spot light on the surface of a sphere at a time. 1... Sphere 2... Transport rotation unit 3... Controller 4... Inspection unit 5... Light source 6... Coaxial epi-illumination device 7... Imaging camera 8... Image processing device I
I...-A/[) converter I2... Image memory 13
...Mask processing device 14...Brightness correction device 1
5... Binarizer 16... Area measuring device 17...
Flat flaw detection device 18... Density measurement device 21... Calibration device 22... Image recording device 23... Image memory 24... Subtractor 25... Subtractor 26...
・Storage device 27-... Adder 28... Fabric defect detector patent Applicant Sumitomo Metal Industries Co., Ltd. Agent Patent attorney Noboru Kono] figure

Claims (1)

[Claims] 1. A means for capturing an image of the rotating spherical surface while irradiating light onto the spherical surface, and based on the captured spherical surface image, the area of the specularly reflected area from the spherical surface and the amount of specularly reflected light from this are determined to eliminate surface scratches. 1. A spherical surface inspection device comprising: detecting means.