JPH02298837A - Measuring method of foam structure - Google Patents

Abstract

PURPOSE:To measure a foam structure precisely in a short time by applying a light onto the surface of a sample having foam in the surface from a plurality of directions, from one direction after another, by picking up an image each direction and by measuring the foam from a difference in brightness in the same position. CONSTITUTION:A light is applied onto the surface of a sample, e.g. set concrete, having foam in the surface from a plurality of directions, from one direction after another, and an image in each direction of application is picked up. The brightness of reflection thereof is different since a foam part has a shape of depression from the surface of the sample. In the foam part, the level of the brightness is different even in the same place according to the direction of application of the light, while the brightness in a part other than the foam part has the same level. Based on this difference in the brightness, e.g. a brightness difference, accordingly, a foam structure can be measured precisely in a short time.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a method for measuring the cell structure of a sample having air bubbles on its surface, and particularly to a method for measuring the cell structure of hardened concrete in a short time.

[Conventional technology]

In order to ensure the frost damage resistance of concrete structures, it is necessary that appropriate entrained air be entrained in the concrete. Therefore, quality control is generally performed by measuring the amount of air in fresh concrete. However, in order to obtain frost damage resistance, the spacing between air bubbles is more important than the amount of air itself, so in order to accurately judge frost damage resistance, it is necessary to determine the air bubble structure in hardened concrete. .

The commonly used methods for investigating the cellular structure in hardened concrete are the A37MC457 point count method or the linear traverse method. These methods involve observing and measuring the cross section of concrete under a stereomicroscope, but in order to eliminate the drawbacks of this method, such as the long time required for measurement and the considerable patience of the person taking the measurements, Recently, methods using image analysis have been attempted.

(S, Chatterji and H, Gudmu
Characterization
of Entrained Air Bub
ble System in Concrete
by MeanSof anImage Anal
ysing Microscope+ Cemen
tand Concrete Re5earch+ V
ol, L No. 4, PP-423~428 (
1,977), Katsumi Genga et al. 2 each; Method for measuring air bubble structure in hardened concrete using image analysis equipment/equipment, Cement/Concrete, No. 471, pp. 22-28 (19
86), Toshitaka Ota and three others; Method for measuring air bubble parameters in hardened concrete using an automatic image analysis system, Proceedings of the 8th Annual Conference on Concrete Engineering, pp.
389-392 (1986)) This is a method in which images input from a television camera are processed by a computer, and the burden on the measurer is significantly reduced compared to the ASTM method.

[Problem to be solved by the invention]

However, in the case of the above method, it is difficult to accurately determine air bubbles in concrete directly from images obtained with a television camera, and in order to distinguish between air bubbles and other parts, the sample is binarized in advance. It is necessary to apply A common method of treatment is to fill the air bubbles on the concrete surface with a white solvent, and then paint the area other than the air bubbles with black paint. For this reason, even if the time involved in the measurement itself is shortened compared to the ASTM method, it still requires time for preprocessing for binarization, and has the disadvantage that it does not substantially shorten the test time. In addition, since the bubbles to be measured are extremely small, the field of view of the image to be captured needs to be as narrow as several Ilv, and in order to accurately determine the air bubbles in concrete, the number of fields of view to be measured must be several hundred. Even using an image analysis device did not necessarily lead to a drastic reduction in measurement time.

An object of the present invention is to provide a method of irradiating the surface of a sample with light from a plurality of directions one by one, capturing images, and measuring the bubble structure in a short time based on the difference in brightness of each image. .

[Means for solving the problem] In order to achieve the above object, it is sufficient to irradiate the cell structure with light from multiple directions one by one, measure each reflected light, and compare the results to detect the difference. The bubble structure measuring method of the present invention irradiates light from a plurality of directions onto the surface of a sample having bubbles, captures images for each irradiation direction, and calculates the difference in brightness at the same position of each image. It is characterized by measuring air bubbles. Further, it is preferable that the plurality of directions are three directions. Further, concrete may be used as the sample, in which case the surface of the concrete sample is first divided into an aggregate region and a cement paste region,
Next, it is recommended to measure air bubbles only in the cement paste area.

[For production]

When there is a bubble structure on the surface of a sample, if light is irradiated from multiple directions, one at a time, the brightness of the reflected light will differ from each other because the bubbles are shaped like depressions from the sample surface. Become something. Bubbles can be measured based on this difference in brightness, that is, the difference in brightness. This will be explained using FIG. 1 for the case where light is irradiated from three directions.

FIG. 1 is a diagram in which a bubble on the surface of a sample is assumed to be a hemisphere, and light of the same illuminance is irradiated onto this cross section from directly above, diagonally to the right, and diagonally to the left. (a) The figure shows the shape of the bubble and the direction of light. (b) The figure shows the distribution of brightness (brightness) and the formation of shadows when light is irradiated from directly above.Brightness indicates its size in 256 levels from 0 to 255, and UP indicates bubbles. UP is the brightness at a certain position outside, and up is the brightness at a position where the bubble is present. Also, in this case there is no shadow. Figure (C) shows the brightness distribution and shadow formation when light is irradiated diagonally from the right. RP and rp are the brightness at the same position as in the (b
This is the brightness at the same position as in Figure J. Here (b), (C1
, (d) When comparing the figures, the brightness of areas other than bubbles, for example,
U.P.

RP and LP are at the same level, but in the bubble portion, the brightness level differs as up+rp+Ip+ depending on the direction of light irradiation even at the same location. This allows bubbles to be measured.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 1 to 9.

Figure 2 shows a block diagram of an image analysis processing device that implements the present invention. It is in a state where it does not. 2 is sample 1
3 is an XY stage that performs two-dimensional movement with the sample 1 mounted on it, 4 is an xY stage control device that controls the movement of the XY stage 3, and 5 is the vertical position of the sample 1. 6 is an automatic focus control device that controls the automatic focus mechanism 5, 7 is a microscopic camera that measures bubbles in the sample 1,
Reference numeral 8 indicates a macro camera for identifying the aggregate region and cement paste region of sample 1; 9.10 indicates camera control units for the micro camera 7 and macro camera 8; and 11 indicates the micro camera 7 and the macro camera 8. 12 is an illumination device for the sample 1, which is equipped with three lights to illuminate the upper surface of the sample 1 with oblique light from the left and right and light from directly above. 13 processes images from the micro camera 7 and the macro camera 8, and also includes an XY stage control unit 4. Automatic focus control device6. An image processing device that controls the video switching unit 11; 14 a personal computer 1 that controls the image processing device 13; 15 a microdisk or floppy that records the program of the personal computer 14 and measurement data of the image processing device 13; A secondary storage device 16 is a printer that outputs measurement data etc. of the image processing device 13.

Next, a method for measuring the cellular structure of hardened concrete using this device will be explained.

Measuring the cellular structure of hardened concrete using an image analysis system involves macroscopic measurements that separate the cement paste area and aggregate area using an image of the entire cut surface of the concrete specimen.
It is divided into micro-measurement, which measures the cell structure using an enlarged image of the cut surface.

(1) Macro measurement Macro measurement is performed based on a binarized grayscale image taken from a television camera.
The TV signal input from the TV camera is digitized for each element (image component) according to its density.
For example, white is set to ON, black is set to "225", and is divided into 256 levels). Binarization is this “0”
This is done by setting an arbitrary value between ``255'' and ``1'' above the set value and ``0'' below the set value.

In the binarization process, the captured gray scale image of the cut surface of the hardened concrete is binarized into an aggregate region and a cement paste region. In other words, from among 256 levels of density, values are adjusted so that areas that are judged to be aggregate (aggregate area) and areas that are not (cement paste area) are appropriately determined, and the aggregate area is black ( 1), Divide the cement paste area into white (0).

Here, the light and shade on a certain arbitrary straight line that crosses the sample is assumed as follows. The length of a straight line is measured in pixels.

In the binarization process, the values above the set binarization level are set to 1, and those below it are set to 0, thereby simplifying the grayscale image.

To explain this using FIG. 3, the area above the binarization level is the aggregate area, and the area below the binarization level is the cement paste area.

However, this method makes all judgments based on color shading, so whitish-colored aggregates or aggregates mixed with crystals are difficult to recognize as aggregates, so manual correction is required.

(2) Micro measurement Micro measurement is performed by capturing three images by shining light of the same intensity on the object from three different directions, and measuring three images of the same point corresponding to the three images. The position of that point is measured from the difference in brightness (brightness).

Here, it is assumed that the reflectance of light at a point on the object corresponding to a point on the image is constant. In order for the reflected light from three different directions to have equal brightness when the target part is a completely diffuse reflective surface, the projection angle C05INH
A light source whose intensity is proportional to the reflection and inversely proportional to the square of the distance is desirable.

This principle states that if a point on a slope with a certain area is illuminated with light of the same illuminance from three directions at different angles, each point will appear to have different brightness from a fixed viewpoint. It is applied.

Here, we assume the following about air bubbles in hardened concrete. Assuming that the reflectance of light at all points on the concrete surface is constant, as explained in Figure 1, the viewpoint is It is set directly above, and its brightness is expressed in 256 levels from 0 to 225. When you look at these three figures together, the brightness level is the same for the parts other than the bubbles, but the brightness level of the shadowed parts of the bubbles is lower, which is different from the data for the parts other than the bubbles. I understand that there is.・Next, a binarized image for identifying bubbles is obtained as follows. Now, the three-dimensional coordinates shown in FIG. 4 are used to express the brightness when light is applied from above, diagonally to the right, and diagonally to the left. One direction of light corresponds to one coordinate axis, and the size of the coordinate axis is a value of the brightness level, ranging from 0 to 255. Three-dimensional coordinates (1,0°0), (0,1,0), (0,0,1)
Consider a triangular (111) plane with vertices. Also, here, a point where all three coordinate values are the same (for example, coordinates (1,
Consider a vector passing through point 1, 1, ) and use this as the reference vector.

In Figure 1, the brightness level at any point in the bubble when light is applied from above, diagonally to the right, and diagonally to the left, Uρ + rl' + lp
is placed on this coordinate system, and a vector starting from (0, 0, 0) is created.

Here, (111) plane and vector (up, rP, 1f
Since a point (referred to as the brightness point) is created where the reference vector intersects with the (111) plane (referred to as the reference point), we can create a vector (referred to as the brightness vector) to the brightness point starting from the point where the reference vector intersects with the (111) plane (referred to as the reference point). can.

Bubbles are identified based on the size of this brightness vector.

The larger this brightness vector is, the greater the possibility that it is a bubble. On the other hand, if this brightness vector is small, it can be said that the possibility that it is a bubble is small. Brightness level in parts other than bubbles in Figure 1, U
Since P, RP, and LP are the same, they will pass through the reference point,
It can be said that there is almost no possibility that it is a bubble. Therefore, from the above, it is sufficient to consider a point having a brightness vector of a certain value or more as a bubble.

Here, as shown in Fig. 5, 3 of the triangular (111) plane
Divide a straight line drawn from the reference point to each vertex into 256 levels from O to 255 starting from the reference point, and connect the points at the same level on the three straight lines to draw the reference point within this triangular plane. It can be leveled into 256 levels at the center.

This is the binarization level, and the brightness level tlp+ rP+
The brightness vector of a point with 'p is expressed at this level by its magnitude. Then, if the brightness vector is binarized with levels above a certain level as 1 and levels below as 0, points on the sample having that brightness vector are determined to be bubbles. This is binarization processing.

Therefore, when dividing all points in the measurement range on the sample into bubbles and non-bubbles, if you set a binarization level that allows the bubbles to be clearly distinguished, the brightness above this binarization level Points with vector levels will be recognized as bubbles.

The setting of the level for this binarization process can be adjusted while referring to the image on the display.

The above-described processing of grayscale images is an application of color analysis for processing color images.

Color images have R (red) and G for each pixel that makes up the image.
It consists of a combination of the three primary colors of light (green) and B (blue). Each color of R, G, and B is 256 from 0 to 255.
It has data on the density of stages, and the color is expressed by the degree of mixing of the three colors. Furthermore, when the data of the density of all three colors is the same, the image becomes monochrome.

Therefore, for example, if the density data of R and G are the same and the data of B is larger than the other data of R and G, then
The same data as R and G of B and B are combined and become monochrome, so the protruding part of B remains and the color becomes bluish. Conversely, if B is smaller than the other R and G data, the displayed color will be a mixture of R and G.

Color analysis is to perform IHP processing on R, G, and B density data to convert it into luminance, five hue, and purity data. I HP processing places R, G, and B data on three-dimensional coordinates as shown in Figure 4, and considers the (111) plane (Figure 6).
. The point where a straight line passing through the coordinates (1, 1, 1) from the coordinate origin intersects this plane is defined as a black and white point.

Purity refers to the degree of mixing of the three primary colors, and as shown in the brightness vector in Figure 5, the vector created from the R, G, and B data starting from the black and white point intersects the (111) plane. Represented by a vector. The level of purity is determined by dividing a straight line drawn from the black and white point to the apex of the triangle into 256 points as black and white point O, and connecting the same level points on the three straight lines to level the triangular plane in 256 levels around the black and white point. It is expressed as

Hue is the angle between a straight line drawn from a black and white point to R (red) of a triangle and a purity vector in the R-G-B-R...- direction. This indicates which colors are biased towards the colors represented by the three primary colors. The hue level is expressed by dividing 360 degrees into 256.

Luminance refers to R, G on three-dimensional coordinates starting from the coordinate origin.
, B data is expressed by the magnitude of the vector drawn to the coordinate point (2).
55, 255, 255) is divided into 256 parts with the coordinate origin set at 0.

In this image analysis system for bubble structure measurement, instead of the R, G, and B data for this color analysis, we process the data of the grayscale image when light is applied in one direction from above, diagonally right, and diagonally left. ing. Therefore, hue and brightness of color analysis are not useful for bubble discrimination, only purity is effective.

Micro measurements must be performed over a large number of fields of view, but the images are processed at the initially set binarization level and automatically repeated on the autostage until the end of the measurement. Further, in a field of view of a certain micro-measurement, if the measurement range is determined to be 100% aggregate area based on the macroni value image, the field of view is automatically moved to the next field of view without performing measurement.

Next, a measurement example will be explained.

The sample concrete is 10 cm square, the measurement range of the macro image is 9 x 9 cm, and the measurement range of the micro image is 6 cm.
It is x6mm. The air content of the fresh concrete used was 4.6%.

The following three methods were used to determine the cell structure from the binarized micro image of Yago, which shows a binarized image of the cell structure in FIG.

(1) This method is based on the ASTM linear traverse method, in which a traverse line is set in the image, and the cell structure is determined from the number and length of bubbles crossed by the line.

(2) This is based on the modified point counting method of ASTM, in which points are set in the image and the bubble structure is determined from the number of bubbles that overlap with the points.

(3) Determine the bubble structure from the number and area of bubbles obtained directly from the image.

FIG. 8 shows the measurement results obtained using the three methods described above, and also shows the results obtained using the ASTM modified point counting method for comparison. (1) in Figure 8,
(2), (3) are the above three methods (1), (
2), 4) representing (3) is the ASTM modified point counting method (Modified Point-Cou
nt Method of ASTM).

Note that the cement paste capacity in method (4) is determined from the formulation.

FIG. 9 shows the distribution of the measured bubble sizes.

The horizontal axis shows the bubble diameter, and the vertical axis shows the frequency in %.

From FIG. 8 and FIG. 9, it can be seen that the cell structure of concrete obtained by this image analysis device is quite reliable.

The time required for image analysis is approximately one hour for all three methods, and the time required for capturing and processing micro images, which requires the most time, is automatically measured, so the time required for the measurement operator is reduced. The measurement takes only about 20 minutes.

〔Effect of the invention〕

As is clear from the above description, according to the present invention, the bubble structure can be measured with high precision in a short time by shining light onto the sample from multiple directions and analyzing the reflected light. It has this excellent effect.

[Brief explanation of the drawing]

Fig. 1 is an explanatory diagram showing the principle of the present invention, Fig. 2 is a diagram showing an example of a block diagram of an image analysis device implementing the present invention, and Fig. 3 is a binary representation of the concrete aggregate area and cement paste area. Figure 4 is an illustration of the illuminance vector, Figure 5 is an illustration of the binarization of the illuminance vector, Figure 6 is an illustration of color analysis, and Figure 7 is an illustration of the binary representation of the bubble structure. FIG. 8 is a diagram showing the measurement results of the example, and FIG. 9 is a diagram showing the bubble distribution of the example. 1--1 sample 2--1 vibration isolation table 3--XY stage 4--XY stage control device 5--1 automatic focus mechanism 6--1 automatic focus control device 7-
--Micro camera 8-m-Macro camera 9.1
0-11 Camera control unit 11---Video switching unit 12-m-Lighting device 13-
--Image processing device

Claims (4)

[Claims] (1) From multiple directions on the surface of a sample with air bubbles on the surface.
Irradiates light in each direction and captures images for each irradiation direction.
A method for measuring bubble structure, characterized in that bubbles are measured from the difference in brightness at the same position in each of the images. (2) The cell structure measuring method according to claim 1, wherein the plurality of directions are three directions. (3) The cell structure measuring method according to claim 1 or 2, characterized in that concrete is used as the sample. (4) The cell structure measuring method according to claim 3, characterized in that the surface of the concrete sample is divided into an aggregate region and a cement paste region, and the bubbles are measured only in the cement paste region.