JPH0154653B2 - - Google Patents

Description

[Detailed description of the invention]

The present invention relates to a sample hardness measuring device, and more particularly to an automatic hardness measuring device that automatically inscribes a predetermined portion of a metal sample and automatically measures the dimensions of the indentation formed on the sample surface to determine the hardness. be. Conventionally, Vickers hardness and Brinell hardness have been widely used to measure the hardness of metal samples. For example, when measuring Bitker's hardness, an indenter made of a diamond square pyramid with a diagonal of 136° is pressed against the surface of a metal sample to form an indentation, and the pressing load is P and the area of the indentation is S. At this time, the Bitkers hardness H _V is determined by H _V =P/S. Therefore, in order to measure Bitker's hardness, the following steps are required: (1) Pressing an indenter against the desired surface of a metal sample with a predetermined load to form an indentation; (2) Measuring the area of the indentation; (3) It is necessary to calculate the hardness from the measured area and load, but in the past, all of these tasks were done manually. Considering the indentation forming process in (1), for example, it is necessary to distinguish between specific parts of the sample, such as the weld metal part of a welded part and the heat-affected zone of the base metal, or the ferrite part and bainite part of ferrite-bainite steel. You may want to measure. In such cases, after the metal structure is revealed, the target area is observed under a microscope, and the indenter of the Vickers hardness meter is manually adjusted so that it is pressed against the desired area. However, this work is very troublesome and requires skill, and has the disadvantage that it requires a lot of time when measurements are to be made at many locations. Also, considering the work in (2), in this case as well, observe the indentation under a microscope with a magnification of 100 to 400 times, move the cursor to the diagonal vertices of the approximately square indentation in the field of view, and measure the interval between them. This was done manually by taking measurements. For this reason, there are individual differences in how the cursor is positioned by the measurer, as well as individual differences in the reading of the length, making it impossible to perform measurements with high reproducibility. In addition, since the person taking the measurements uses a microscope, they are subject to severe fatigue and have the disadvantage of increasing the probability of errors. Additionally, it takes several minutes to measure an indentation in one place, and it takes an extremely long time to measure hardness in multiple areas.
The disadvantage was that the measurer was extremely fatigued and the measurement error became increasingly large. Regarding the calculation work in (3), the calculation itself can be performed using a calculator or a conversion table, but there is a risk of human error occurring when entering or transcribing numbers. Furthermore, when measuring the hardness of multiple points on a sample in conventional manual hardness measurement,
The work of associating the indentation at each point with the area measurement value is extremely troublesome, and if a mistake is made, the measurement becomes unreliable, so the measurer had to be extremely careful in this respect as well. The purpose of the present invention is to eliminate various drawbacks of the prior art,
By automatically performing all of the above-mentioned indentation formation work, indentation area measurement work, and hardness calculation work, the hardness of a desired part of the sample can be measured quickly, with high precision, and with high reproducibility. The present invention aims to provide an automatic hardness measuring device that can significantly reduce the burden. The automatic hardness measurement device of the present invention captures an image of a sample whose hardness is to be measured, and processes the resulting image to obtain a texture feature value and a reference feature value that represent statistically quantified values of texture pattern characteristics. an indentation point determining means that automatically determines the indentation point at which the indenter should be pressed by comparing the indentation points; an indentation dimension measuring means for automatically measuring the dimension of an indentation by finding a point where the gray tone level of an image obtained by imaging the indentation changes rapidly; The present invention is characterized by comprising hardness calculation means for calculating the hardness at the stamping point of the sample from the indentation size and the load. The present invention will be described in detail below with reference to the drawings. FIG. 1 is a diagram showing the configuration of an embodiment of the automatic hardness measuring device of the present invention, in which a sample 1 whose hardness is to be measured is arranged movably in the X-axis and Y-axis directions, and is placed at a predetermined position. Table 2 that can be served
Place it on top. The table 2 is supported by a Bitkers hardness meter body 3, and an indenter 4 for forming an indentation on the sample 1 and an imaging device 5 for taking an image of the sample are attached to the hardness meter body above the table 2. . The indenter 4 and the objective lens 5a of the imaging device 5 are arranged, for example, in alignment in the X-axis direction. The table 2 is configured so that it can be driven by X-axis and Y-axis drive devices 6 and 7, respectively, and the position of the table 2 in the X-axis and Y-axis directions can be detected by position detectors 8 and 9. . These driving devices 6, 7 and position detectors 8, 9 are connected to a table movement control section 10 to move the table 2 as desired. The analog image signal output from the imaging device 5 is supplied to the analog-to-digital converter 11.
The digital image signal is converted into a digital image signal in stages and stored in the gray memory 12. In this example, one screen is vertically ×
Gray memory 1 because it is divided horizontally into 512 x 512 pixels
2 is also configured to be able to store a gray-level digital image signal of 512×512 pixels. Next, the digital image signal stored in the gray memory 12 is read out to the feature value calculating section 13 as appropriate to calculate a predetermined texture feature amount. This texture feature amount is a value that statistically quantifies the characteristics of the texture pattern, such as smoothness and graininess, and the density level of a pixel that is a certain distance δ (γ, θ) from a certain pixel is Find the matrix ij representing the frequency of j, that is, the probability that the density level of a pixel that is a certain distance δ (γ, θ) away from the pixel whose density level is i is j for all combinations of density levels. The matrix Pij (co−occurrence
matrix). This matrix Pij is used to define and calculate various feature amounts corresponding to the texture, that is, data indicating the features of the texture. An example of an arithmetic expression that defines this feature amount is shown in the table below.

[Table] On the other hand, a text file in which similar feature quantities are obtained for the reference metallographic structure is stored in the text file section 14, and the feature quantities of the sample obtained by the feature quantity calculation section 13 and stored in the text file section 14 are stored in the text file section 14. The texture discrimination unit 15
A statistical comparison process is performed to classify the tissue of the sample according to the texture information. The maximum likelihood method can be used for this statistical processing. Fig. 2 shows an image obtained by imaging the welded part of sample 1 with the imaging device 5, where part A is the weld metal, part B is the heat affected zone of the base metal, and part A and B are the weld metal.
The boundary between the two parts is the pound part. If it is desired to measure the hardness at several points on a line perpendicular to the welded portion of such a sample, it is first necessary to detect the bonded portion. This bond can be automatically detected as a boundary between two different texture regions. In other words, the tissue state of part A and part B is greatly different, and there is also a difference in the distribution of the feature values obtained from the above matrix Pij, which determines the magnitude of the difference in density level between a certain pixel and neighboring pixels. occurs. Part A and part B can be clearly distinguished by detecting this difference using a maximum likelihood method, which is a statistical method.
FIG. 3 shows an image showing the boundaries determined in this way. As is clear from a comparison between FIG. 2 and FIG. 3, the bond portion can be detected with high accuracy by making a determination based on the above-mentioned feature amounts. The coordinates of the bond portion detected in this way are stored in the coordinate storage section 16. Next, the position where the hardness is to be measured, that is, the position of the stamp point, is determined by the stamp point indicating section 17. For example, when engraving at a pitch of 0.5 mm in the direction perpendicular to the bond, simply input the engraving start point and pitch information, the engraving point will be automatically determined, and its coordinates will be stored in the coordinate storage unit 16. be done. As described above, since the objective lens 5a of the imaging device 5 and the indenter 4 are shifted in the X-axis direction, the coordinates of the bond portion and the coordinates of the embossing point detected based on the image information from the imaging device 5 are It is necessary to shift a predetermined distance in the axial direction. After memorizing the coordinates of the stamping point as described above,
Stamping starts, but in this case, the coordinate storage unit 16
The coordinates of the marking point stored in the table movement control unit 1
0, X-axis and Y-axis drives 6 and 7
is driven so that the table 2, and thus the predetermined position of the sample 1, is directly below the indenter 4. After performing such positioning, a signal is sent to the stamping device 18 to press the indenter 4 against the surface of the sample 1 with a predetermined load to form an impression. After forming an impression at one location, the table 2 is moved so that the next impression point is positioned directly below the indenter 4. In this way, impressions can be automatically formed at successive marking points. After an indentation is automatically formed as described above, the area of the indentation is measured. According to the present invention, this indentation area measurement operation is also performed automatically. This will be explained below. First, the table 2 is moved by the table movement control unit 10, the first impression is positioned with respect to the objective lens 5a of the imaging device 5, and an image of the impression is captured. This analog image signal is converted into a digital image signal of 256 levels of gray tone by the analog-digital converter 11, and the gray memory 12
to be memorized. By converting the image of the indentation into a digital signal at a gray level in this way, it is possible to clearly distinguish the indentation from the matrix tissue, scratches, and corroded areas, and the area of the indentation can be measured with high accuracy and repeatability. Can be done. The digital image signal thus stored in the gray memory 12 is supplied to the indentation area measuring section 19, where the area of the indentation is measured. In order to measure the area of the indentation, it is necessary to detect the contour points of the indentation, and these are determined as points _P1 and _P2 where the image signal suddenly changes, as shown in FIG. In FIG. 4, the gray level of the image signal C of the scanning line B that crosses the indentation A is plotted on the vertical axis, and the pixel position is plotted on the horizontal axis. The outline of the indentation A is found as a point where the gray level value of the image signal changes rapidly, but in order to improve the accuracy of this detection,
You can find the envelope tangent to the curve of the part that includes the inflection point of the image signal curve, pass the image signal through a low-pass filter to cut out high-frequency noise components, or find the average rate of change within a certain section of the image signal curve. A method such as smoothing the noise component can be adopted. After detecting the contour of the indentation in this manner, for example, the length of the diagonal line of the indentation is measured and supplied to the hardness calculating section 20. In this hardness calculation section,
Based on the pressing load P of the indenter 4, the length of the diagonal of the indentation, and the facing angle θ of the indenter 4, the Bitkers hardness H _V is measured using the following formula. H _V = 2Psinθ/2/ ² [Kg/mm ² ] Therefore, for example, when the indenter 4 with a facing angle of 136° is inscribed with a pressing load of 10 Kgf/mm ² , the bitker hardness H _V is H _V = 18.54/ It can be calculated from ² . In this way, the bitker hardness at successive marking points can be determined one after another, and the results can be printed out, for example, by the output section 21. In this case, the table 2 is automatically moved sequentially to the coordinate positions of the embossing points stored in the coordinate storage section 16, so that the hardness at each embossing point can be output accurately without confusing the hardness. The present invention is not limited to the embodiments described above, and can be modified and modified in many ways. For example, in the above-mentioned embodiment, the dimensions of these indentations were measured after making indentations by making indentations at a plurality of positions on the sample one after another. Immediately after the indentation is formed, the dimensions of the indentation may be measured, the table may be moved to the next engraving point, and the same operation may be performed. Also,
In the embodiment described above, the indentation forming work, the indentation dimension measurement work, and the hardness calculation work are performed completely automatically, but for example, the position of the embossing point may be manually input via the embossing point indicating section 17. Semi-automatic operations such as this may also be performed. In addition, in the above example, the hardness was calculated by measuring the length of the diagonal line of the indentation, but in general, the Vickers hardness is given by H _V = P/S, where S is the area of the indentation. In step 19, the area S of the indentation may be determined. Further, in the above example, the Vickers hardness was measured, but the Brinell hardness may also be measured. As described above, according to the automatic hardness measuring device of the present invention, the position of the stamping point to be pressed with the indenter is automatically determined by imaging the sample and detecting its tissue,
This allows you to move the table on which the sample is placed, form an indentation at a desired position, take an image of this indentation, automatically measure the dimensions of the indentation, and automatically measure the hardness. , it is possible to locally measure the hardness of adjacent parts of different metals or parts of metals with different structures. In addition, it automatically determines the marking point, performs the marking work, and measures the impression.
The time required for hardness measurement is significantly shortened, the measurement accuracy is improved, and the labor of the operator is significantly reduced.

[Brief explanation of drawings]

Fig. 1 is a diagram showing the configuration of an embodiment of the automatic hardness measuring device of the present invention, Fig. 2 is a diagram showing the texture of the metal structure in the welded part, and Fig. 3 is an image of the welded part shown in Fig. 2. FIG. 4 is a diagram showing an image processed to reveal a bond portion, and is a diagram for explaining the operation of measuring the dimensions of an indentation. 1... Sample, 2... Table, 3... Hardness meter body, 4... Indenter, 5... Imaging device, 6, 7...
X-axis and Y-axis direction drive device, 8, 9... X-axis and Y-axis direction position detector, 10... Table movement control unit,
11...Analog-digital converter, 12...
Gray memory, 13... Feature amount calculation unit, 14...
Text file section, 15... Texture discrimination section, 16... Coordinate storage section, 17... Embossing point instruction section, 18... Embossing device, 19... Indentation dimension measuring section, 20... Hardness calculation section, 21 ...Output section.

Claims (1)

[Claims] 1 The indenter is pressed by comparing the texture feature amount, which represents the statistically quantified value of the texture pattern feature obtained by image processing the image obtained by imaging the sample whose hardness is to be measured, with the reference feature amount. an indentation point determining means that automatically determines the indentation point to be indented; an indentation means that automatically presses an indenter against the sample with a predetermined load to form an indentation at the determined indentation point position; an indentation dimension measuring means that automatically measures the dimension of the indentation by finding a point where the gray tone level of the image obtained by capturing the image of the indentation changes rapidly; An automatic hardness measuring device comprising: hardness calculation means for calculating the hardness at a stamping point of a sample.