JPH0511563B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an automatic surface roughness measuring device that quantitatively and automatically evaluates the surface finish, etc. of parts of machined structures at high speed. It is something.

(Conventional technology) Conventionally, in order to check the surface finish of machined parts, the roughness of the surface was determined qualitatively by an operator observing it with the naked eye, or by using a surface roughness meter. Judgments were made based on surface roughness.

(Problem to be solved by the invention) However, qualitative judgment by the naked eye of the worker requires a great deal of effort and time, and the judgment varies depending on the skill level of the worker, resulting in problems in terms of accuracy. It was hot.

Furthermore, when a surface roughness meter is used, since the surface roughness meter measures values on a certain line, the information becomes two-dimensional and has the disadvantage of inefficiency. Furthermore, since the measurement was performed while in contact with the sample surface, there were problems such as damage to the sample.

On the other hand, as an example of image processing, texture analysis, which is a method of quantifying patterns, is known, and the quantitative values include texture features. However, from an engineering perspective, surface roughness is regarded as a pattern,
No attempt has been made to quantify this as a texture feature.

The purpose of the present invention is to solve the above-mentioned problems, and to quantitatively and automatically perform surface processing at high speed by using texture analysis that distinguishes image homogeneity and contrast differences among image processing at gradation level. An object of the present invention is to provide an automatic surface roughness measurement device that can measure the roughness of surfaces.

(Means for Solving the Problems) The automatic surface roughness measuring device of the present invention includes: an optical system that forms an image of the surface condition of a sample; an imaging device that captures this image and outputs an analog image signal; An A/D converter converts this analog image signal into a digital image signal with a gradation level of four or more gradations, and from the digital image signal from this A/D converter, the feature amount F for texture analysis is calculated. _LVL 〓 ⁱ⁼¹ _LVL 〓 ^j=1 {p(i, j)} ^2Here , p(i, j): Co-occurrence matrix LVL: Determine the gradation level, and based on this feature amount F, roughly The present invention is comprised of a texture feature discriminator that determines the surface roughness from a previously determined and stored calibration curve that indicates the relationship between the texture feature F and a value required for the surface roughness, and a display section that displays the discrimination result. This is a characteristic feature.

(Function) The configuration of the present invention will be specifically explained below.
FIG. 1 is a block diagram showing a specific configuration of an embodiment of the apparatus of the present invention. In Fig. 1, optical system 1
An ordinary optical microscope is used to obtain images preferably from 2 times to 100 times the size of the sample. The optical system 1 needs to be able to change the magnification, preferably between 2x and 100x. The reason why the magnification is limited to the range mentioned above is that if it is 2 times or less, there will be no difference in the texture feature amount corresponding to roughness.
Furthermore, if the magnification is increased by 100 times or more, the intervals between the concave and convex portions become too large, which may make it difficult to extract the texture feature amount properly.

The image magnified by the optical system 1 described above is captured by the imaging device 2 connected to the optical system, read out as an analog image signal, and supplied to the A/D converter 3. The A/D converter 3 performs A/D conversion on the analog image signal at a gradation level of 4 or more gradations, preferably 256 or more gradations, to obtain a digital image signal. As for the gradation level, four or more gradation levels are required in order to quantify the difference in roughness using the texture feature amount. In addition, the size of the pixel must be such that the surface roughness of the object to be measured can be determined. It is preferable to divide the image into digital images suitable for texture analysis. The digital image signal from the A/D converter 3 is temporarily stored in a frame memory in the texture feature determining unit 4, and texture analysis is performed on the stored digital image signal to quantify roughness.

Texture analysis will be explained below. First, the inventors found that the surface roughness of a sample can be quantified by the index F = _LVL 〓 ⁱ⁼¹ _LVL 〓 ^j=1 {p(i, j)} ² , which expresses the fineness of the "texture." This has been revealed through research. Here, F is a value representing the texture feature amount, LVL represents a division level of light and shade at the time of texture analysis, and p
(i,j) is a Co-occurrence matrix. here
The co-occurrence matrix is a matrix obtained by determining the probability that the gray level of an adjacent pixel in a micro region consisting of, for example, 32×32 pixels changes from i to j for all gray level changes. The value of the texture feature F described above uniquely corresponds to the roughness of the sample as described later, so the relationship between the necessary values for roughness (Ra, PC...) and the texture feature F is calculated in advance using a calibration curve. If it is determined as follows and stored in the texture feature discriminator, the necessary value for roughness can be determined by determining F of the target image. Here, Ra: center line average roughness (JISB0601), PC: number of ridges above a certain height or depth (ANSI B-46-1 (1985) appendix (c)) are defined.

Therefore, if the above-mentioned texture feature F is calculated and obtained for the digital image signal stored in the frame memory in the texture feature discriminator 4, the required value (Ra, PC
) can be found. The values determined by the method described above are displayed on the display section 5 together with a calibration curve determined in advance.

(Example) Ordinary steel material was coated with #240 dry paper.
Samples were prepared that were uniformly polished to a roughness of #320, #600, and #1000. With the device of the present invention,
These samples were enlarged 20 times and texture analysis was performed to determine the texture feature amount F. FIG. 2 is a graph showing the relationship between the roughness and texture feature amount F of each sample. As is clear from FIG. 2, since the texture feature amount F can be uniquely determined with respect to roughness, it can be seen that the roughness can be quantified using this texture feature amount F.

In addition, in order to investigate the effect of the magnification of the optical system on the texture feature F, texture analysis was performed on a sample polished with #600 at different magnifications to obtain the texture feature F. FIG. 3 is a graph showing the relationship between the magnification of the optical system and the texture feature amount F. From Figure 3, the magnification of the optical system is from 2x to
It can be seen that a magnification of 100x is appropriate.

The present invention is not limited to the embodiments described above, and many modifications and changes are possible. For example, in the above-mentioned embodiments, an example was shown in which a microscope was used as the optical system, but any optical system can be used as long as it can magnify the sample in the range of 2 times to 100 times, and for example, a macrostad etc. can be suitably used.

(Effects of the Invention) As is clear from the detailed explanation above, according to the automatic surface roughness measuring device of the present invention, the roughness of the sample surface can be measured by performing texture analysis on the surface image of the sample. Can be measured quantitatively and automatically at high speed. Furthermore, since the surface image of the sample is processed, the information becomes three-dimensional and the surface roughness can be measured without contact.

[Brief explanation of the drawing]

Fig. 1 is a block diagram showing the specific configuration of one embodiment of the present invention device, Fig. 2 is a graph showing the relationship between the roughness of each sample and the texture feature amount, and Fig. 3 is the magnification of the optical system. FIG. DESCRIPTION OF SYMBOLS 1... Optical system, 2... Imaging device, 3... A/D conversion part, 4... Texture feature quantity determination part, 5... Display part.

Claims (1)

[Scope of Claims] 1. An optical system that forms an image of the surface state of a sample, an imaging device that captures this image and outputs an analog image signal, and an optical system that captures this image and outputs an analog image signal. From the A/D converter that converts into a digital image signal and the digital ^image signal from this _A /D converter, the feature amount F for _texture analysis ^is calculated. j)} ^2Here , p(i, j): Co-occurrence matrix LVL: Find the gradation level, and based on this feature F, calculate the relationship between the value required for roughness and the texture feature F. 1. An automatic surface roughness measurement device comprising: a texture feature discriminator that determines surface roughness from a previously determined and stored calibration curve; and a display that displays the determination results.