Cement mixture quality prediction method and cement mixture quality prediction system

Abstract

To provide a technique that enables non-contact prediction of cement mixture quality regardless of skills of measurers.SOLUTION: A cement mixture quality prediction method is provided, comprising acquiring a captured image of an uncured cement mixture, and predicting the quality of the cement mixture on the basis of a brightness distribution in the image.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a technique for predicting the quality of a cement mixture such as concrete. [Background technology]

[0002] Strength, slump or slump flow, air content, and chloride content are known as quality evaluation values ​​for cement mixtures such as concrete. Concrete is made up of materials including water, cement, sand, and gravel, but also contains air. The air content of concrete affects the workability during concrete pouring, freeze-thaw resistance, and strength, so it is important that the air content of concrete is appropriate. It is also important to manage the indicators used to evaluate consistency, such as concrete slump and slump flow, in order to ensure the desired workability.

[0003] Known methods for measuring air content in concrete include the air chamber pressure method (JIS A 1128: Testing Method for Air Content of Fresh Concrete by Pressure), the mass method (JIS A 1116: Testing Method for Unit Volume Mass of Fresh Concrete and Testing Method for Air Content by Mass), and the volumetric method (JIS A 1118: Testing Method for Air Content of Fresh Concrete by Volume). However, these measurement methods require the collection and measurement of concrete samples after mixing, which requires a certain level of skill. Therefore, it is desirable to be able to predict quality, such as air content, without contacting the concrete and without relying on the skill of the measurer. This issue is not limited to air content, but also applies to other qualities. For example, the conventional slump test (JIS A 1101: Slump Testing Method for Concrete) similarly suffers from the problem of being dependent on the skill of the measurer.

[0004] Patent Document 1 discloses a technique for measuring the air content of fresh concrete based on the pressure inside a concrete pumping pipe, the measured unit volume mass of fresh concrete, and the unit volume mass of fresh concrete that does not contain air. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2020-125662 Summary of the Invention [Problem to be solved by the invention]

[0006] The present disclosure provides a technology that can predict the quality of a cement mixture in a non-contact manner, without relying on the skill of the person performing the measurement. [Means for solving the problem]

[0007] The quality prediction method of the present disclosure for a cement mix includes the steps of acquiring an image of an unhardened cement mix, and predicting the quality of the cement mix based on the brightness distribution in the image.

[0008] The quality prediction system for cement mixes disclosed herein includes an acquisition unit that acquires an image of an unhardened cement mix, and a prediction unit that predicts the quality of the cement mixes based on the brightness distribution in the image. [Effects of the Invention]

[0009] According to the present disclosure, since an image of the cement kneaded body is used, it is possible to predict the quality of the unhardened cement kneaded body in a non-contact manner. Furthermore, it is possible to make a prediction that is not dependent on the skill of the measurer. [Brief explanation of the drawings]

[0010] [Figure 1]1 is a flowchart showing a method for predicting the quality of an unhardened cement mix according to the present disclosure. [Figure 2A] FIG. 2 is a block diagram showing the configuration of a quality prediction system for a cement mix that executes the method described in FIG. 1. [Figure 2B] FIG. 2 is a block diagram showing the configuration of a quality prediction system for a cement mix that executes the method described in FIG. 1. [Figure 2C] FIG. 2 is a block diagram showing the configuration of a quality prediction system for a cement mix that executes the method described in FIG. 1. [Figure 3] FIG. 1 is a diagram showing brightness histograms of images taken of seven types of concrete (A to G). [Figure 4] FIG. 10 is a diagram showing a luminance histogram of only concrete A. [Figure 5] FIG. 10 is a diagram showing the relationship between the estimated air amount obtained from correlation data and the actually measured air amount, with the brightness parameter being "number of occurrences of maximum brightness / half-value width." [Figure 6] FIG. 10 is a diagram showing brightness histograms of three types of concrete with different strengths. [Figure 7] FIG. 10 is a diagram showing brightness histograms of three types of concrete with different slumps. [Figure 8] FIG. 10 is a diagram showing brightness histograms of three types of concrete with different slump flows. [Figure 9] FIG. 10 is an explanatory diagram regarding the number of luminance occurrences at the boundary near the peak. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0012] The prediction method of this embodiment includes the steps of acquiring an image of an unhardened cement mix and predicting the quality of the cement mix based on the brightness distribution in the image. It has been found that the brightness distribution (particularly the shape of the brightness histogram) of the image of the cement mix correlates with the quality of the cement mix, such as the air content, of the concrete. Therefore, the quality of the concrete mix is ​​estimated based on the brightness distribution of the image of the cement mix.

[0013] This method uses images captured by a camera 12 (see FIGS. 2A to 2C), making it possible to predict the quality of the unhardened cement mixture, such as the air content, without contacting the object. Furthermore, this method makes it possible to predict the quality without relying on the skill of the person performing the measurement.

[0014] Examples of unhardened cement mixtures include fresh concrete and fresh mortar. Camera 12 (see Figures 2A-C) can capture images of the cement mixture at various times during or after mixing, such as when the materials are being mixed in a mixer 10 (see Figure 2A), when the mixture is stored in a hopper or other container after mixing, when the mixture is being loaded into an agitator car after mixing, when the mixture is removed from the chute of the agitator car, or after pouring but before hardening. In particular, as shown in Figure 2B, when the cement mixture is in a stagnant state, such as when it is removed from a hopper or agitator car and flows down the chute, or when pouring, it is subject to the effects of sunlight outdoors. To mitigate this, an enclosure can be installed and images captured under certain environmental conditions. Note that the same conditions can also be applied to images of the mixture flowing down the chute, as shown in Figure 2C.

[0015] In the following embodiment, as shown in Fig. 2A, an example will be described in which the prediction target is fresh concrete and fresh mortar at the stage of mixing the materials in a mixer 10. Fig. 1 is a flowchart showing the quality prediction method for an unhardened cement mix according to the present disclosure. Figs. 2A to 2C are block diagrams showing the configuration of a quality prediction system for a cement mix that executes the method shown in Fig. 1. The prediction method and prediction system will be described below with reference to Fig. 1 and Figs. 2A to 2C.

[0016] (Measurement step) In step ST1 shown in FIG. 1, the surface of the cement mixture being mixed in the mixing chamber 10a of the mixer 10 is imaged by the camera 12, and the image is acquired by the acquisition unit 22 of the computer (see FIGS. 2A to 2C). The image is captured continuously at multiple points during mixing. In this embodiment, an example is given in which 450 images of 512 pixels x 512 pixels are acquired, but this is just an example and is not limiting. The number of pixels can be changed as appropriate. Furthermore, the number of images acquired can be changed as appropriate, provided that it is one or more.

[0017] (Prediction step) In step ST2 following step ST1 shown in FIG. 1, the luminance data generating unit 23 (see FIGS. 2A to 2C) generates luminance data based on the acquired image. The generated image is a grayscale image in which one pixel is represented by a numerical value of 256 gradations from 0 to 255. Note that R (red), G (green), B (blue) may be converted into a color image expressed by a numerical value of 256 gradations from 0 to 255. The luminance data includes a luminance value of the number of images x the number of pixels per image. The luminance data of this embodiment includes 450 x 512 x 512 luminance values, and each luminance value is a value from 0 to 255. A known method can be used to generate the luminance data.

[0018] In step ST3 following step ST2 shown in FIG. 1, the brightness parameter generating unit 24 (see FIGS. 2A to 2C) generates brightness parameters (see FIG. 4) related to peak P2 or peak portion P1 including the vicinity of peak P2, which has the highest frequency in a brightness histogram based on one or more images, based on the brightness data. FIG. 3 shows brightness histograms of images captured of seven types of concrete (A to G). The brightness values ​​(0 to 255) are divided into 10-point intervals, and the vertical axis shows the number of brightness occurrences (frequency) in each interval. Concrete A is shown as a black bar, and concretes B to G are arranged in order from left to right of concrete A. FIG. 3 shows that the number of brightness occurrences (frequency) with a brightness value of 110 to 110 corresponds to peak P2. The number of brightness occurrences (frequency) shown in FIG. 3 is not the frequency for a single image, but the total frequency of brightness across multiple captured images (450 images). FIG. 4 shows the brightness histogram for only concrete A. As shown in FIG. 4, the luminance histogram has a peak P2 where the frequency is maximum and a peak portion P1 that includes the vicinity of peak P2. Peak portion P1 is a part of the luminance distribution. Peak portion P1 includes not only peak P2 where the frequency is maximum, but also the luminance in the vicinity of peak P2. Peak portion P1 can be said to be the top of a mountain in the luminance histogram. In this embodiment, the vicinity of peak P2 is defined as a portion other than peak P2 that has a luminance equal to or greater than the maximum frequency (number of occurrences of maximum luminance) multiplied by a coefficient (0.5). In the example of FIG. 4, the number of occurrences of maximum luminance is approximately 18×10 6 Therefore, the number of luminance occurrences is about 9 × 10 6 The above brightness corresponds to the peak portion P1. The brightness value is 80 to 139. In this embodiment, the peak portion P1 includes both the peak P2 and the vicinity of the peak P2, but is not limited to this. The peak portion P1 may be only the peak P2, or may be only the vicinity of the peak P2 without including the peak P2. The vicinity of peak P2 is a part other than peak P2 that has a brightness equal to or greater than the maximum frequency (number of occurrences of maximum brightness) multiplied by a coefficient (0.5) when the peak position is clear, and is a part that has a brightness equal to or greater than the maximum frequency multiplied by a coefficient (0.6) when the peak position is unclear.

[0019] For example, the brightness parameters are listed as follows: As shown in FIG. Half-width: This indicates the range W1 of luminance values ​​equal to or greater than the maximum frequency H1 multiplied by a specified coefficient. The specified coefficient is 0.5 when the peak position is clear; if the peak position is unclear, the coefficient can be shifted to 0.3, 0.4, 0.5, or 0.6 to clearly correct the ambiguity. The half-width is an index representing the luminance width of the peak portion P1 in the luminance histogram waveform; the wider the luminance width, the larger the half-width, quantitatively representing the characteristics of the luminance histogram waveform. In Figure 4, this corresponds to the range W1 of luminance values ​​from 80 to 139. Peak position: This refers to the luminance value where the number of luminance occurrences (frequency) in an image is at its maximum. By extracting the peak position, it is possible to quantitatively express the characteristics of waveform analysis and the number of luminance occurrences. In Figure 4, this corresponds to a luminance value of 110. - Number of maximum luminance occurrences: This refers to the number of occurrences (maximum frequency) of luminance at the peak position. The number of maximum luminance occurrences quantitatively expresses the characteristics such as the brightness of the image and the kurtosis of the waveform. In Figure 4, the maximum frequency H1 (number of maximum luminance occurrences: approximately 18 x 10 6 ) applies. Number of luminance occurrences at the boundary near the peak: The number of luminance occurrences at the boundary near the peak. In the example of FIG. 9, this corresponds to the number of luminance occurrences at the boundary position (luminance value 130) of the range W1 near the peak.

[0020] In step ST4 following step ST3 shown in FIG. 1, quality (e.g., air content) is predicted based on the brightness parameters. Specifically, correlation data 20 correlating the brightness parameters for peak portion P1 with the quality of the cement mix is ​​used to identify the quality corresponding to the brightness parameter as a predicted value. The quality may be any of the air content, slump, slump flow, strength, surface water content of the concrete raw material, and granularity of the cement mix. Note that the air content, slump, slump flow, strength, surface water content of the concrete raw material, and granularity of the cement mix as qualities are all correlated with brightness parameters. However, among these qualities, the air content, strength, surface water content of the concrete raw material, and granularity, which are parameters that capture the mass scale of the volume (volume factor quality), have a significantly higher correlation with brightness parameters than slump or slump flow, thereby achieving higher prediction accuracy. The correlation data 20 may be any data correlating brightness parameters with quality. For example, it may be an approximation equation or correlation equation generated based on multiple measured data including brightness parameters and measured quality, or it may be a prediction model generated by machine learning to output quality using one or more brightness parameters as input data.

[0021] The brightness parameters to be used may be one of the following, or a combination of multiple parameters. Case 1: Half width W1 Case 2: Half-width W1, maximum brightness occurrence count Case 3: Half-width W1, peak position, specified coefficient (0.5, 0.3, etc.)

[0022] As an example, 100 sets of data were prepared for mixing mortar or concrete with a twin-shaft mixer. The data items were as follows: Water-cement ratio (W / C): 35-65% Cement types: 5 types Admixture types: 10 types ·Admixture amount: 2~10kg / m 3 Unit water volume: 150-180 kg / m 3

[0023] The quality items obtained from the 100 cases are as follows: Slump: 5~25.5cm Slump flow: 22.5~74.5cm Air content: 1.5-7.0% Compression strength: 20~50N / mm 2 Surface water content of aggregate: 0-5%

[0024] For some of the above 100 cases, the brightness parameter was set to "number of maximum brightness occurrences / half-width" and correlated with the air volume to create correlation data 20 as a correlation equation. Figure 5 is a diagram showing the relationship between the estimated air volume obtained from correlation data 20 using the brightness parameter as "number of maximum brightness occurrences / half-width" and the actually measured air volume. The data used in Figure 5 is the remaining data out of the 100 cases used for the prediction in Figure 5. As shown in Figure 5, it can be confirmed that there is a correlation between the brightness parameter and quality (air volume).

[0025] As another example, the correlation data 20 may be expressed by the following conversion formula: Air volume = Number of maximum luminance occurrences / Peak position × Half-value rate (0.5) × Normalization coefficient

[0026] Next, if the quality is air volume, <1> ~ <4> The accuracy rate is shown when the brightness parameter is changed as shown in Figure 1. <1> The brightness parameter was "half-width," and the accuracy rate for tolerances within ±1.0% was 50%, the accuracy rate for tolerances within ±1.5% was 86%, and the accuracy rate for tolerances within ±2.0% was 94%. <2> The brightness parameter was "half-width / number of maximum brightness occurrences," and the accuracy rate for tolerances within ±1.0% was 73%, the accuracy rate for tolerances within ±1.5% was 89%, and the accuracy rate for tolerances within ±2.0% was 97%. <3> The brightness parameter was "number of maximum brightness occurrences / peak position," and the accuracy rate was 85% within a tolerance of ±1.0%, 92% within a tolerance of ±1.5%, and 98% within a tolerance of ±2.0%. <4> The brightness parameters were "half-width / peak position / coefficient (0.3)", and the accuracy rate for tolerances within ±1.0% was 89%, the accuracy rate for tolerances within ±1.5% was 96%, and the accuracy rate for tolerances within ±2.0% was 100%. As described above, it can be seen that accuracy tends to improve when the brightness parameters include both W1 (the brightness range where the frequency is equal to or greater than the maximum frequency multiplied by a predetermined coefficient) and the maximum frequency H1 (the maximum number of brightness occurrences) shown in Figure 4. W1 and H1 can be considered parameters related to the shape of the peak portion (parameters that represent the shape). In other words, it was found that accuracy improves when the brightness parameters are changed based on skewness and kurtosis, which are parameters based on the waveform shape, and when the brightness parameters are changed based on the waveform appearance position.

[0027] (Prediction System) The prediction method may be performed by a person or, as shown in FIGS. 2A to 2C, by a system. As shown in FIGS. 2A to 2C, the quality system for a cement mix includes a camera 12 that captures an image of the cement mix, an acquisition unit 22 that acquires an image of the cement mix captured by the camera 12, and a prediction unit 21 that predicts the quality of the cement mix based on the brightness distribution in the image. The prediction system is configured to acquire correlation data 20 from memory or externally via a network. The acquisition unit 22 and the prediction unit 21 are implemented by one or more processors of a computer loading a predetermined program. The prediction unit 21 includes a brightness data generation unit 23, a brightness parameter generation unit 24, and a quality identification unit 25. The brightness data generation unit 23 generates brightness data based on the acquired image. The brightness parameter generation unit 24 generates brightness parameters related to peak portions with the highest frequency in a brightness histogram based on one or more images, based on the brightness data. The quality identification unit 25 uses the correlation data 20 to identify a quality (such as air content) corresponding to the brightness parameter as a predicted value. This enables the system to automatically predict the quality of the cement mix.

[0028] [Another embodiment] (A) In the above embodiment, the concrete (cement mixture) is imaged during mixing in the mixer 10, but this is not limiting. For example, images may be captured when mixing is completed and the concrete is stored in a hopper, or when mixing is completed and the concrete is being loaded into an agitator truck. The camera 12 is not limited to being mounted on the mixer 10, but may be mounted on a separate device, a person-held camera, or an information and communication terminal such as a smartphone with a camera function. In this case, for example, the prediction unit 21 and the acquisition unit 22 may be connected to the camera or the information and communication terminal via a network so as to be able to communicate with each other. The prediction unit 21 and the acquisition unit 22 may be mounted on a terminal that displays the prediction results, or on an external server other than the terminal that displays the prediction results.

[0029] (B) In the above embodiment, only the amount of air is used as an example of quality, but the quality is not limited to this. As an example, Figure 6 shows brightness histograms of three types of concrete with different strengths. It can be seen from Figure 6 that brightness parameters such as the number of maximum brightness occurrences and half-width are correlated with strength. Therefore, correlation data 20 that associates brightness parameters with the strength of concrete (cement mix) can be used. As an example, Figure 7 shows brightness histograms of three types of concrete with different slumps. It can be seen from Figure 7 that brightness parameters such as the number of maximum brightness occurrences and half-width are correlated with slump. Therefore, correlation data 20 that associates brightness parameters with the slump of concrete (cement mix) can be used. As an example, Figure 8 shows brightness histograms of three types of concrete with different slump flows. From Figure 8, it can be seen that brightness parameters such as the number of maximum brightness occurrences and half-width are correlated with slump flow. Therefore, correlation data 20 that associates brightness parameters with the slump flow of concrete (cement mix) can be used. The same is also thought to be true for the surface water content and particle size of the concrete raw materials.

[0030] [1] As described above in the above embodiment, the method for predicting the quality of a cement mix may include the steps of acquiring an image of an unhardened cement mix, and predicting the quality of the cement mix based on the brightness distribution in the image. In this way, by using an image of the cement mixture, it is possible to predict the quality of the unhardened cement mixture without contacting it, and it is possible to make a prediction that is not dependent on the skill of the person making the measurement.

[0031] [2] The method for predicting the quality of a cement mix described in [1] above may also include predicting the quality of the cement mix based on brightness parameters related to a peak portion P1 including a peak P2 having the highest frequency in a brightness histogram based on an image. Since the peak portion P1 in the luminance histogram correlates with quality, more accurate prediction may be possible.

[0032] [3] The method for predicting the quality of a cement mix described in [2] above may also include using correlation data 20 that associates brightness parameters with the quality of the cement mix to identify the quality corresponding to the brightness parameters based on the acquired image as a predicted value. Since the correlation data 20 that associates the quality of the unhardened cement mixture with the brightness parameters is used, more accurate prediction may be possible.

[0033] [4] In the method for predicting the quality of a cement mixture described in [2] or [3] above, the brightness parameter may include at least one of a brightness value at which the frequency is maximum, a maximum frequency, and a brightness range having a frequency equal to or greater than a value obtained by multiplying the maximum frequency by a predetermined coefficient. These luminance parameters have a high correlation with quality, which makes it possible to further improve prediction accuracy.

[0034] [5] In the method for predicting the quality of a cement kneaded product according to any one of the above [1] to [4], the quality may be the air content of the cement kneaded product. As a quality factor, it becomes possible to predict the amount of air in the cement mix.

[0035] [6] In the method for predicting the quality of a cement mix according to any one of [1] to [4] above, the quality may be any one of the slump, slump flow, strength, surface water ratio of concrete raw materials, and particle size of the cement mix. As quality, it is possible to predict the slump, slump flow, strength (e.g., early strength, compressive strength, flexural strength, splitting strength), surface water content of concrete raw materials, granularity, etc. of the cement mix. Note that, although the air content, slump, slump flow, strength, surface water content of concrete raw materials, and granularity of the cement mix as quality factors all correlate with brightness parameters, the air content, strength, surface water content of concrete raw materials, and granularity, which are parameters that can capture the mass scale of volume (volume factor quality), have a significantly higher correlation with brightness parameters than slump or slump flow, and can achieve higher prediction accuracy.

[0036] [7] As in the above embodiment, the quality prediction system for cement mix may include an acquisition unit 22 that acquires an image of an unhardened cement mix, and a prediction unit 21 that predicts the quality of the cement mix based on the brightness distribution in the image. This allows the system to automatically predict the amount of air in the cement mix.

[0037] [8] The present invention can also be specified as a computer program invention. That is, the program may cause one or more processors to acquire an image of an unhardened cement mixture and predict the quality of the cement mixture based on the brightness distribution in the image. By executing a program having such functions, it becomes possible to automatically predict the quality of the cement mixture. The program may predict the quality of the cement mixture based on brightness parameters relating to a peak portion including the peak with the highest frequency or the vicinity of the peak in a brightness histogram based on the image. The program may use correlation data that associates the brightness parameters with the quality of the cement mixture to identify the quality corresponding to the brightness parameters based on the acquired image as a predicted value. The brightness parameter may include at least one of a brightness value having a maximum frequency, a maximum frequency, and a brightness range having a frequency equal to or greater than the maximum frequency multiplied by a predetermined coefficient. The quality may be the amount of air in the cement mixture. The quality may be any one of the slump, slump flow, strength, surface water content, and particle size of the cement mixture.

[0038] Although the embodiments of the present disclosure have been described above with reference to the drawings, the specific configurations should not be considered to be limited to these embodiments. The scope of the present disclosure is defined not only by the description of the above embodiments but also by the claims, and further includes all modifications within the meaning and scope of the claims.

[0039] The structures employed in the above-described embodiments can be employed in any other embodiment. The specific configurations of the components are not limited to the above-described embodiments, and various modifications are possible within the scope of the present disclosure. [Explanation of symbols]

[0040] 20: Correlation data 21: Prediction section 22: Acquisition part H1: Maximum frequency P1: Peak part P2: Peak

Claims

1. acquiring an image of an unhardened cement mixture; and predicting the quality of the cement mix based on the brightness distribution in the image.

2. The quality prediction method for a cement mix described in claim 1, wherein the quality of the cement mix is ​​predicted based on brightness parameters relating to a peak portion including a peak with the highest frequency or a peak portion near the peak in a brightness histogram based on the image.

3. The method for predicting the quality of a cement kneaded body according to claim 2, further comprising: using correlation data relating the brightness parameter to the quality of the cement kneaded body, identifying the quality corresponding to the brightness parameter based on the acquired image as a predicted value.

4. 4. The method for predicting the quality of a cement mixture according to claim 2 or 3, wherein the brightness parameter includes at least one of a brightness value at which the frequency is maximum, a maximum frequency, and a brightness range in which the frequency is equal to or greater than a value obtained by multiplying the maximum frequency by a predetermined coefficient.

5. The method for predicting quality of a cement kneaded product according to claim 1 , wherein the quality is an air content of the cement kneaded product.

6. 4. The method for predicting the quality of a cement mix according to claim 1, wherein the quality is any one of the slump, slump flow, strength, surface water content of concrete raw materials, and particle size of the cement mix.

7. an acquisition unit that acquires an image of an unhardened cement kneaded body; a prediction unit that predicts the quality of the cement mix based on the brightness distribution in the image.

8. Acquiring an image of an unhardened cement mixture; predicting the quality of the cement mixture based on the brightness distribution in the image; A program that causes one or more processors to execute the above.

Images