Automatic evaluation system of concrete fluidity based on point cloud and machine learning, and method for the same

Abstract

A point cloud and machine learning-based automatic concrete fluidity evaluation system and method are provided, which can accurately and automatically evaluate concrete fluidity by acquiring a point cloud, which is three-dimensional shape information, through a depth measuring means for a concrete specimen for fluidity evaluation and analyzing concrete fluidity characteristics using artificial intelligence-based machine learning. Furthermore, compared to conventional qualitative evaluation methods by operators, concrete fluidity can be evaluated quantitatively, quickly, and accurately by utilizing point clouds and machine learning, thereby ensuring the quality of ready-mix concrete / on-site concrete. Additionally, the diameter, height, and material segregation risk of the concrete specimen are generated as point cloud information, and through a machine learning-based concrete fluidity evaluation module, rheological characteristics of the concrete specimen before hardening, compressive strength after hardening, and material segregation characteristics can be extracted.

Description

Technology Field

[0001] The present invention relates to an automatic concrete fluidity evaluation system for predicting the quality and physical properties of concrete, and more specifically, to a method for an automatic concrete fluidity evaluation system that predicts the quality and physical properties of concrete by applying point cloud and machine learning technologies to concrete slump test results to automatically evaluate concrete fluidity. Background Technology

[0002] Generally, the fluidity of concrete is a key physical property that determines workability at construction sites; using concrete that does not meet standards increases the difficulty of the work and leads to substandard construction.

[0003] In addition, since the addition of additional materials is prohibited after the mixing of concrete is completed at ready-mixed concrete production plants, a large amount of industrial waste is generated due to concrete with unsuitable fluidity after production.

[0004] In addition, due to real-time changes in the physical properties and quality characteristics of raw materials, the fluidity (slump and slump flow) of the final shipped ready-mix concrete can vary significantly even with the same mix ratio.

[0005] Accordingly, cameras are installed inside the mixers of ready-mix concrete batch plants, allowing production personnel to visually check the consistency of the concrete mix through the cameras during production.

[0006] Furthermore, since the actual slump of the final shipped ready-mix concrete can only be verified through quality control testing after production is complete, variations in the slump of the final product still occur depending on the experience and intuitive tendencies of the production personnel; consequently, the initial response at the production stage to sudden changes in flow characteristics caused by fluctuations in raw materials remains inadequate.

[0007] In other words, slump tests are conducted for physical property evaluation and quality control prior to concrete hardening before on-site pouring of ready-mix concrete. At this stage, materials are utilized based on the slump test results, as well as the experience and intuition of on-site workers.

[0008] These qualitative methods lack accuracy and may lead to future quality issues; furthermore, evaluation methods by field workers are unsuitable for material property evaluation and quality control due to inaccurate and subjective inspections.

[0009] Meanwhile, as prior art for solving the aforementioned problems, Korean registered patent No. 10-836464 discloses an invention titled "Automatic measuring device and measuring method for concrete slump flow using a computing device," which will be explained with reference to FIGS. 1a and 1b.

[0010] FIG. 1a is a side view showing the state of an automatic concrete slump flow measuring device using a computing device according to conventional technology during filming, and

[0011] FIG. 1b is a configuration diagram of an automatic concrete slump flow measuring device using a computing device.

[0012] As illustrated in FIGS. 1a and 1b, an automatic concrete slump flow measuring device using a computing device according to the prior art,

[0013] The apparatus comprises: a slump plate (20) having a color different from that of the unhardened concrete being tested; a digital camera (10) installed directly above the slump plate (20) to photograph the concrete flowing on the slump plate (20) at predetermined intervals; and a computing device (50) that measures the slump flow of concrete by dividing each captured image into a preset number of pixels, and by using the color difference between the slump plate (20) and the concrete for each image to distinguish the boundary between the slump plate (20) and the concrete, and specifying the position of the pixel on the boundary line between the slump plate (20) and the concrete.

[0014] Additionally, the lower end (40) of the fixing stand is fixedly installed on one side of the slump plate (20), the upper end (30) of the fixing stand is rotatably connected to the lower end (40), and a digital camera (10) is attached to the end of the upper end (30).

[0015] At this time, the height (H) of the bottom (40) of the digital camera fixing stand is made high enough so that the entire bottom of the slump plate (20) can be seen through the digital camera (10).

[0016] Referring to FIG. 1b, an automatic slump flow measuring device for concrete using a computing device according to the prior art includes a digital camera (10) and a computing device (50), and the computing device (50) includes a digital camera control module (51), a slump flow analysis module (52), and a coarse aggregate distribution ratio evaluation module (53).

[0017] The digital camera control module (51) controls the process of photographing and saving concrete by the digital camera (10).

[0018] The slump flow analysis module (52) performs a slump flow measurement process by reading an image captured by a digital camera (10) and analyzing the image.

[0019] The coarse aggregate distribution ratio evaluation module (53) analyzes the captured image to measure the coarse aggregate distribution ratio in the concrete.

[0020] In the case of an automatic concrete slump flow measuring device using a computing device according to conventional technology, a digital camera (10) is positioned directly above the slump plate (20) at a height sufficient to show the entire bottom of the slump plate (20),

[0021] When the flow of concrete begins, the digital camera (10) photographs the concrete at predetermined intervals (e.g., once every second), and the photography can continue at the same rate until the flow of concrete completely stops.

[0022] Afterwards, the image file of the captured concrete is transmitted to a computing device (50) via a USB interface, and the computing device can analyze the captured image to obtain slump flow data.

[0023] According to an automatic concrete slump flow measuring device using a computing device based on conventional technology, the slump flow of concrete can be measured quantitatively.

[0024] In addition, the distribution ratio of coarse aggregate can be quantitatively measured based on the measured area of ​​the concrete.

[0025] In addition, since the results do not vary depending on the tester or test equipment, objective results can be obtained compared to conventional technology in the slump flow test, which measures the consistency of concrete.

[0026] In the case of an automatic concrete slump flow measuring device using a computing device according to conventional technology, it can be applied to a system that measures the slump or slump flow value of a product after production is completed and reflects it in future production.

[0027] However, considering the time required for testing and the current state of the domestic market with continuous production, there is a problem in that real-time feedback is impossible.

[0028] Meanwhile, in overseas regions such as Japan and some domestic companies, production is carried out by referring to the mixer motor's torque meter to predict slump and slump flow based on the difference in mixer motor load values ​​observed according to the flow characteristics of the ready-mix concrete mixture being produced.

[0029] However, even in such cases, reliability is difficult to guarantee due to limitations in distinguishing changes in viscosity and dough consistency based on raw material characteristics, and the trend is that this data is used merely as reference material for production personnel.

[0030] Meanwhile, as another prior art, Korean registered patent No. 10-1936218 discloses an invention titled "Artificial Intelligence Learning-Based Ready-Mixed Concrete Production System,"

[0031] By analyzing video data output from a camera capturing the interior of a batch plant mixer during the ready-mix concrete production stage using an artificial intelligence algorithm, the slump and slump flow of the shipped ready-mix concrete product can be predicted in real time, and the quality of the aggregates received as raw materials, the concentration of recovered water, and the moisture content of the aggregates can be automatically measured.

[0032] According to a conventional AI learning-based ready-mixed concrete production system, a pre-processed image is produced by analyzing a concrete mixing video within a mixer, and the flow characteristics of the concrete shown in the image are quantified through slump and slump flow tests. By applying a learning algorithm to the AI, slump and slump flow can be predicted during the ready-mixed concrete production process.

[0033] However, it was necessary to develop a technology that diversifies slump prediction technology by calculating slump prediction values ​​based on physical data obtainable during the concrete mixing process, and to derive more accurate prediction results through cross-validation with slump prediction values ​​obtained through artificial intelligence image analysis.

[0034] Meanwhile, as prior art for solving the aforementioned problems, Korean registered patent number 10-2235712 discloses an invention titled "Concrete Slump Prediction System," which will be explained with reference to FIGS. 2 and FIGS. 3.

[0035] FIG. 2 is a diagram showing the concept of a concrete slump prediction system according to conventional technology, and FIG. 3 is a detailed configuration diagram of the flow characteristic monitoring unit shown in FIG. 2.

[0036] The reduction gear equipped in the concrete mixer shown in Fig. 2 performs the role of maintaining a constant rotational speed (rpm) of the mixer, and is subjected to a load due to the mixing state of the concrete raw materials inside the mixer, the consistency of the mixture, etc., and consumes power in proportion to the amount of load.

[0037] Accordingly, by measuring the power required by the reduction gear, the consistency of the mixer, that is, the slump of the fresh concrete, can be determined.

[0038] In addition, by combining artificial intelligence that deep learns big data of videos inside and outside the mixer and slump test values, the accuracy of concrete slump prediction values ​​can be improved.

[0039] A concrete slump prediction system according to conventional technology may be configured to include: a power meter equipped to measure the power required by a concrete mixer reducer, as shown in FIG. 2; a data logger that records the power measured by the power meter according to the elapsed mixing time; a calculation unit that extracts power data for a certain period after the completion of concrete raw material input from the data recorded in the data logger and derives an average power value; and a memory unit that matches the average power value with the slump measurement value of fresh concrete discharged from the mixer and stores it as coordinate data on an xy plane.

[0040] At this time, the computation unit can derive a regression equation by performing regression analysis on multiple coordinate data stored in the memory unit, and derive a slump prediction value by substituting the newly derived average power value into the regression equation.

[0041] In addition, by utilizing the conventional mixer camera inside the batch plant as is or using a separate dedicated camera, and by including an image analysis unit equipped with artificial intelligence, a slump prediction value can be derived by artificial intelligence.

[0042] Specifically, as illustrated in FIG. 3, the video analysis unit may include a splitter, a video recording device (such as a DVR or NVR), artificial intelligence, and a monitor. In this case, the splitter simultaneously distributes and transmits the video captured by the camera to the video recording device and the monitor, and the video recording device stores the video so that the artificial intelligence analyzes and learns from the video stored in the video recording device.

[0043] At this time, artificial intelligence analyzes the video captured by the camera and predicts the slump of the final shipment of ready-mix concrete in real time by deep learning big data on the correlation between video analysis data inside the mixer for a certain period of time (e.g., 5 to 6 seconds before discharge) before the fresh concrete is discharged and the slump test value of the fresh concrete discharged from the mixer, and can be divided into a learning module and a prediction module according to its function.

[0044] According to a concrete slump prediction system based on conventional technology, a slump prediction value can be easily calculated based on power values, which are physical data acquired during the concrete mixing process. Additionally, by cross-verifying the calculated slump prediction value with the slump prediction value obtained through artificial intelligence image analysis, the slump can be predicted more accurately during the ready-mix concrete production process.

[0045] However, in the case of a concrete slump prediction system based on conventional technology, there are limitations in accurately predicting concrete fluidity because it utilizes 2D RGB images (photos) captured by a digital camera.

[0046] Meanwhile, as a series of accidents have recently occurred due to concrete quality and performance issues, research is actively underway to rapidly and accurately evaluate concrete quality and performance on-site.

[0047] Here, the slump test is a simple field test method for evaluating concrete fluidity, but according to the U.S. ASTM C1611, it has been reported to be capable of indirectly predicting quality and physical properties. However, because the subjectivity of field workers is involved, there is a need for technology to evaluate concrete quality and physical properties more objectively.

[0048] According to conventional technology, digital cameras are used for concrete slump tests and applied to concrete mixers to evaluate fluidity. However, these digital cameras have limitations, such as the necessity of calibration due to lens distortion and the difficulty of obtaining consistent results caused by various factors like lighting and shadows. Additionally, there is a problem in that it is impossible to address changes in physical properties and quality caused by temperature, humidity, etc., prior to ready-mix concrete pouring. Prior art literature

[0049] Republic of Korea Registered Patent No. 10-2235712 (Registration Date: March 29, 2021), Title of Invention: "Concrete Slump Prediction System" Republic of Korea Registered Patent No. 10-1936218 (Registration Date: January 2, 2019), Title of Invention: "AI Learning-Based Ready-Mixed Concrete Production System" Republic of Korea Registered Patent No. 10-1115235 (Registration Date: February 3, 2012), Title of Invention: "Concrete Slump Measuring Device" Republic of Korea Registered Patent No. 10-953942 (Registration Date: April 13, 2010), Title of Invention: "Concrete Flow Tester and Method for Measuring Concrete Consistency Using the Same" Republic of Korea Registered Patent No. 10-836464 (Registration Date: June 2, 2008), the invention Title: "Automatic measuring device and measuring method for slump flow of concrete using a computing device" Japanese Registered Patent No. 6,963,258 (Registration Date: October 19, 2021), Title of Invention: "Method for measuring rheology constant of concrete" The problem to be solved

[0050] The technical objective of the present invention to solve the aforementioned problems is to provide a point cloud and machine learning-based system and method for automatically evaluating concrete fluidity, which can automatically evaluate concrete fluidity by acquiring a point cloud, which is three-dimensional shape information, through a depth measuring means for a concrete test specimen for fluidity evaluation, and analyzing concrete fluidity characteristics using artificial intelligence-based machine learning.

[0051] Another technical objective of the present invention is to provide a point cloud and machine learning-based automatic concrete fluidity evaluation system and method that can quantitatively, quickly, and accurately evaluate concrete fluidity by utilizing point clouds and machine learning, compared to conventional qualitative evaluation methods performed by operators.

[0052] Another technical objective of the present invention is to provide a point cloud and machine learning-based automatic concrete fluidity evaluation system and method capable of generating point cloud information for the diameter, height, and material segregation risk of a concrete test specimen, extracting pre-hardening rheological characteristics of the concrete test specimen through a machine learning-based concrete fluidity evaluation module, predicting post-hardening compressive strength, and predicting material segregation characteristics. means of solving the problem

[0053] As a means for achieving the aforementioned technical challenge, the point cloud and machine learning-based automatic concrete fluidity evaluation system according to the present invention comprises: a data analysis terminal that generates point cloud information by collecting and analyzing a point cloud including the diameter, height, and material segregation risk of a concrete test specimen obtained through a depth measuring means, and automatically evaluates concrete fluidity to derive quantitative concrete properties for the concrete test specimen; wherein the data analysis terminal includes a concrete mix ratio for the concrete test specimen and a machine learning-based concrete fluidity evaluation module that performs machine learning using point cloud information generated by a point cloud information generation module that collects and analyzes a point cloud obtained through a depth measuring means to generate point cloud information, automatically evaluates concrete fluidity, and derives quantitative concrete properties for the concrete test specimen; and wherein the machine learning-based concrete fluidity evaluation module includes a machine learning module that performs machine learning using a concrete mix ratio for the concrete test specimen and point cloud information including the diameter, height, and material segregation risk of the concrete test specimen as input. The method includes a concrete fluidity evaluation module that automatically evaluates concrete fluidity according to the learning results of the machine learning module and derives quantitative concrete properties for a concrete specimen, wherein the concrete fluidity evaluation module includes a rheology characteristic extraction unit that extracts rheology characteristics of the concrete specimen before hardening; a compressive strength prediction unit that predicts the compressive strength of the concrete specimen after hardening; and a material separation characteristic prediction unit that predicts the material separation characteristics of the aggregate of the concrete specimen.

[0054] Meanwhile, as another means for achieving the aforementioned technical task, the point cloud and machine learning-based automatic concrete fluidity evaluation method according to the present invention comprises, in the point cloud and machine learning-based automatic concrete fluidity evaluation method using the point cloud and machine learning-based automatic concrete fluidity evaluation system, a) a step of fabricating and arranging a concrete test specimen for concrete fluidity evaluation; b) a step of generating a point cloud, which is three-dimensional shape information, by photographing the concrete test specimen before hardening through a depth measuring means; c) a step in which a point cloud information generation module collects the point cloud from the depth measuring means; d) a step of deriving point cloud information including the diameter, height, and material segregation risk of the concrete test specimen by analyzing the point cloud; e) a step in which a machine learning-based concrete fluidity evaluation module performs machine learning using the concrete mix ratio and point cloud information as input; and f) a step in which the machine learning-based concrete fluidity evaluation module automatically evaluates quantitative concrete physical properties for the concrete test specimen. Effects of the invention

[0055] According to the present invention, for a concrete test specimen for fluidity evaluation, a point cloud, which is three-dimensional shape information, is obtained through a depth measuring means, and concrete fluidity characteristics are analyzed using artificial intelligence-based machine learning, thereby enabling accurate automatic evaluation of concrete fluidity.

[0056] According to the present invention, the quality of ready-mix concrete / on-site concrete can be secured by quantitatively evaluating concrete fluidity quickly and accurately using point clouds and machine learning, compared to conventional qualitative evaluation methods by workers.

[0057] According to the present invention, the diameter, height, and material segregation risk of a concrete test specimen are generated as point cloud information, and the rheological characteristics of the concrete test specimen before hardening are extracted through a machine learning-based concrete fluidity evaluation module, the compressive strength after hardening is predicted, and the material segregation characteristics are predicted. Brief explanation of the drawing

[0058] FIG. 1a is a side view showing the state of an automatic concrete slump flow measuring device using a computing device according to conventional technology during filming, and FIG. 1b is a configuration diagram of an automatic concrete slump flow measuring device using a computing device. FIG. 2 is a diagram showing the concept of a concrete slump prediction system according to conventional technology. FIG. 3 is a detailed configuration diagram of the flow characteristic monitoring unit illustrated in FIG. 2. FIG. 4 is a diagram for comparing and explaining conventional 2D visual analysis and 3D point cloud analysis according to an embodiment of the present invention. FIG. 5 is a diagram showing the determination of ROI from data acquired by a depth measuring means in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 6 is a schematic configuration diagram of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 7 is a detailed configuration diagram of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 8 is a detailed configuration diagram of a concrete fluidity evaluation module in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 9 is an embodiment of the present invention This is a diagram schematically illustrating the implementation of a point cloud and machine learning-based automatic concrete fluidity evaluation system. Fig. 10 is a diagram showing the overall view of a field experiment of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. Fig. 11 is a diagram illustrating a point cloud obtained from a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. Fig. 12 is a diagram showing a concrete test specimen for fluidity evaluation in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. Fig. 13 is a diagram illustrating a depth measurement means in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.FIG. 14a is a diagram showing axis alignment and noise removal of a point cloud in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and FIG. 14b is a diagram showing a point cloud of a region of interest (ROI). FIG. 15 is a diagram schematically showing a machine learning algorithm in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 16a to 16c are diagrams specifically showing the output of a machine learning algorithm in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 17 is a diagram showing the point cloud analysis process in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 18 is a diagram showing the diameter and height analysis of a concrete test specimen in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 19 is a diagram showing the material segregation risk evaluation of a concrete test specimen in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. FIG. 20a is a point cloud and machine learning-based Figure 20 is a diagram showing material separation prediction data evaluated by an automatic concrete fluidity evaluation system, and Figure 20b is a diagram specifically showing area A shown in Figure 20a. Figure 21 is a diagram illustrating material separation characteristics evaluated by a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention. Figure 22 is a flowchart illustrating a point cloud and machine learning-based automatic concrete fluidity evaluation method according to an embodiment of the present invention. Specific details for implementing the invention

[0059] Embodiments of the present invention are described below with reference to the attached drawings so that those skilled in the art can easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0060] Throughout the specification, when a part is described as "comprising" a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components. Furthermore, terms such as "…part" as used in the specification refer to a unit that processes at least one function or operation, and this may be implemented in hardware, software, or a combination of hardware and software.

[0061] First, FIG. 4 is a diagram for comparing and explaining conventional two-dimensional visual analysis and three-dimensional point cloud analysis according to an embodiment of the present invention.

[0062] Referring to Fig. 4, in the case of conventional two-dimensional visual analysis, a digital camera is used to evaluate fluidity by applying it to a concrete slump test and a concrete mixer, and dynamic separation and static separation can be predicted.

[0063] In contrast, in the case of three-dimensional point cloud analysis according to an embodiment of the present invention, data (e.g., point cloud and RGB colors) is obtained by utilizing a depth measuring means to test the slump flow of a concrete specimen for rapid evaluation of the separation degree of a fresh mixed concrete specimen.

[0064] That is, post-processing of data obtained through a depth measurement means is performed to determine the region of interest (ROI) and perform rotational alignment; subsequently, the diameter, depth, and curvature are determined through slump flow analysis, and the degree of separation can be predicted for the suspected separation area.

[0065] Hereinafter, with reference to FIGS. 5 to 21, a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention will be described, and with reference to FIG. 22, a point cloud and machine learning-based automatic concrete fluidity evaluation method according to an embodiment of the present invention will be described.

[0066] [Point Cloud and Machine Learning-Based Automated Concrete Fluidity Evaluation System]

[0067] First, in the case of the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, first, a concrete slump point cloud is acquired using a depth measuring means. Second, using the acquired point cloud, a Region of Interest (ROI) is established, such as axis alignment and noise removal, and quality is analyzed, such as fluidity analysis and material segregation. Third, concrete material properties are predicted through machine learning using information such as concrete mix ratios and rheological characteristics, along with the point cloud.

[0068] FIG. 5 is a diagram showing the determination of ROI from data obtained by a depth measuring means in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.

[0069] As illustrated in FIG. 5, in the case of the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, more than 50,000 point clouds can be collected by acquiring point cloud data of the concrete surface after a slump test using Realsense Viewer software provided by Intel.

[0070] In addition, in the case of the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, as a point cloud analysis technique, noise such as floor slabs is removed and a point cloud algorithm for camera-concrete horizontal alignment is applied to acquire more than 30,000 concrete point clouds corresponding to the ROI.

[0071] By utilizing this, information such as slump diameter and curvature can be analyzed, and the risk of material separation based on point cloud non-linear fitting can be evaluated.

[0072] In addition, in the case of the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, to predict concrete physical properties and quality, a machine learning model is trained using the results of rheological characteristics and changes in compressive strength according to the mix ratio of the concrete used in the experiment and the point cloud analysis results, and subsequently, concrete physical properties and quality can be predicted using the trained machine learning model.

[0073] Meanwhile, FIG. 6 is a schematic diagram of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.

[0074] Referring to FIG. 6, a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention includes a data analysis terminal (100), a depth measuring means (200), and a concrete test specimen (300). The data analysis terminal (100) may include a point cloud information generation module (110), a machine learning-based concrete fluidity evaluation module (120), a fluidity evaluation DB (130), and a display (140), but is not limited thereto.

[0075] The point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention is based on a database of concrete pre-hardening physical properties (fluidity, rheology, material separation, etc.) based on various experiments, and

[0076] In addition, by photographing the surface of concrete before it hardens after a fluidity (slump) test using a depth measuring means (200) including a depth camera and a lidar, the results are analyzed using a point cloud, which is three-dimensional shape information, and accordingly, the physical properties before hardening, such as concrete fluidity, risk of material segregation, and rheological properties, can be automatically evaluated and derived.

[0077] In addition, the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention utilizes three-dimensional shape information and uses one depth measuring means (200).

[0078] At this time, after the concrete fluidity test, the depth measuring means (200) is used to take a picture, and the physical properties such as fluidity, quality, and material separation of the concrete before hardening can be evaluated.

[0079] In addition, in the case of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, concrete fluidity (slump) test results are obtained using a point cloud obtained using a depth measuring means (200) and software, and the diameter, height, and material segregation risk of a concrete test specimen (300) are analyzed using a predetermined algorithm.

[0080] At this time, the material segregation risk is algorithmized to be automatically analyzed immediately after acquiring the point cloud, and after acquiring the point cloud, point cloud information including the diameter, height, and material segregation risk of the concrete test specimen can be derived all at once.

[0081] Therefore, based on machine learning results, artificial intelligence can be utilized to automatically derive quantitative concrete properties, such as rheological characteristics before hardening, compressive strength after hardening, and material segregation characteristics, by using concrete mix ratios and point cloud information.

[0082] When the point cloud and machine learning-based automatic concrete fluidity evaluation system according to the embodiment of the present invention is applied, the concrete characteristic evaluation technology that relied on field workers can be replaced with a technology for rapidly and accurately predicting concrete physical properties and quality based on objective data, which can be utilized for future construction automation.

[0083] Meanwhile, FIG. 7 is a detailed configuration diagram of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and

[0084] FIG. 8 is a detailed configuration diagram of a concrete fluidity evaluation module in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and FIG. 9 is a diagram schematically illustrating the implementation of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.

[0085] Referring to FIGS. 7 to 9, a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention comprises a data analysis terminal (100), a depth measuring means (200), and a concrete test specimen (300).

[0086] At this time, the data analysis terminal (100) may include a point cloud information generation module (110), a machine learning-based concrete fluidity evaluation module (120), a fluidity evaluation DB (130), and a display (140).

[0087] A concrete test specimen (300) is fabricated and placed for evaluating concrete fluidity, as shown in FIG. 9.

[0088] As illustrated in FIG. 9, the depth measuring means (200) captures the concrete test specimen (300) to generate a point cloud, which is three-dimensional shape information. Specifically, the depth measuring means (200) can acquire a point cloud and RGB images of the concrete fluidity test before hardening of the concrete test specimen using commercial RealSense viewer software.

[0089] As illustrated in FIG. 9, the data analysis terminal (100) collects and analyzes the point cloud obtained through the depth measuring means (200) to generate point cloud information, and automatically evaluates the fluidity of the concrete through machine learning to derive quantitative concrete properties for the concrete test specimen (300).

[0090] For example, the data analysis terminal (100) may be implemented as a desktop or laptop, but is not limited thereto.

[0091] Here, the point cloud information may include the diameter, height, and material segregation risk of the concrete test specimen (300). Additionally, the quantitative concrete properties for the concrete test specimen (300) may include rheological properties before concrete hardening, compressive strength after hardening, and material segregation properties.

[0092] Specifically, the point cloud information generation module (110) of the data analysis terminal (100) generates point cloud information by collecting and analyzing the point cloud obtained through the depth measurement means (200).

[0093] The machine learning-based concrete fluidity evaluation module (120) of the data analysis terminal (100) performs machine learning using the concrete mix ratio for the concrete test specimen (300) and the point cloud information generated by the point cloud information generation module (110) as input, and automatically evaluates the concrete fluidity to derive quantitative concrete properties for the concrete test specimen (300).

[0094] The fluidity evaluation DB (130) stores concrete fluidity evaluation results obtained through experiments on various concrete test specimens and quantitative concrete physical property evaluation results derived from the machine learning-based concrete fluidity evaluation module (120).

[0095] The display (140) outputs the quantitative concrete physical property evaluation results derived from the machine learning-based concrete fluidity evaluation module (120).

[0096] Here, the point cloud information generation module (110) may include: a data collection unit (111) that collects point clouds obtained through the depth measuring means (200); a data processing unit (112) that performs axis alignment and noise removal on the point clouds collected by the data collection unit (111); and a data analysis unit (113) that analyzes the point clouds to generate point cloud information including the diameter, height, and material separation risk of the concrete test specimen (300). Here, the axis alignment of the data processing unit (112) is performed to rotate and flatten the angle that is inevitably tilted when photographing the concrete test specimen (300) through the depth measuring means (200).

[0097] Additionally, the machine learning-based concrete fluidity evaluation module (120) may include: a machine learning module (121) that performs machine learning by inputting point cloud information including the concrete mix ratio for the concrete test specimen (300) and the diameter, height, and material segregation risk of the concrete test specimen (300); and a concrete fluidity evaluation module (122) that automatically evaluates concrete fluidity according to the learning result of the machine learning module (121) to derive quantitative concrete physical properties for the concrete test specimen (300).

[0098] Here, the concrete fluidity evaluation module (122) may include a rheology characteristic extraction unit (122a) for extracting rheology characteristics before hardening of the concrete test specimen (300), as shown in FIG. 6; a compressive strength prediction unit (122b) for predicting the compressive strength after hardening of the concrete test specimen (300); and a material separation characteristic prediction unit (122c) for predicting the material separation characteristics of the aggregate of the concrete test specimen (300).

[0099] Accordingly, according to the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, point cloud data of the surface can be acquired using a depth measuring means (depth camera) after a concrete slump test.

[0100] Furthermore, it is possible to analyze fluidity information and assess the risk of material segregation by utilizing point cloud analysis algorithms. Additionally, various concrete material properties can be predicted by integrating point cloud analysis technology with a database constructed through machine learning and experiments.

[0101] Meanwhile, FIG. 10 is a diagram showing the overall view of a field experiment of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and

[0102] FIG. 11 is a diagram illustrating a point cloud according to an embodiment of the present invention and a point cloud obtained from a machine learning-based automatic evaluation system for concrete fluidity.

[0103] The point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention can perform field experiments for evaluating concrete fluidity by installing a concrete test specimen (300), a depth measuring means (200), and a data analysis terminal (100), as shown in FIG. 10.

[0104] At this time, a point cloud as illustrated in FIG. 11 can be obtained by the depth measuring means (200). That is, in the point cloud and machine learning-based automatic concrete fluidity evaluation system according to the embodiment of the present invention, three-dimensional shape information for a concrete test specimen can be obtained by the point cloud.

[0105] Specifically, a point cloud refers to data collected by a Lidar sensor, an RGB-D sensor, etc., and is obtained by photographing a concrete test specimen through a depth measuring means (200) in an embodiment of the present invention.

[0106] For example, a Lidar sensor or an RGB-D sensor sends light / signals to an object and records the time it takes for them to return to calculate distance information for each light / signal and generates a single point, and a point cloud refers to a set cloud of multiple points spread out in three-dimensional space.

[0107] At this time, unlike images, point clouds have depth (z-axis) information, so they are basically represented as N, X, and 3 NumPy arrays, where each N line is mapped to a single point and has 3(x, y, z) information.

[0108] In the case of such image data, the position information of the pixel is always positive, the reference point is from the top left, and the coordinate values ​​are expressed as integers; however, in the case of point cloud data, the position of the point is positive or negative, the reference point is the position of the sensor, and the coordinate values ​​are expressed as real numbers. For example, regarding the x, y, and z values ​​in the coordinates of the reference point, x represents the front and back, y represents the left and right, and z represents the top and bottom, respectively.

[0109] Meanwhile, FIG. 12 is a drawing showing a concrete test specimen for fluidity evaluation in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and

[0110] FIG. 13 is a drawing illustrating a depth measuring means in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.

[0111] In the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, the concrete test specimen (300) can be manufactured and placed as shown in FIG. 12.

[0112] As illustrated in FIG. 13, the depth measuring means (200) in the point cloud and machine learning-based concrete fluidity automatic evaluation system according to an embodiment of the present invention is of a type that is easy to attach and detach and carry, and may be, for example, an Intel RealSense L515, but is not limited thereto.

[0113] At this time, the depth measuring means (200) can acquire point cloud and RGB images of the concrete fluidity test of the concrete test specimen before hardening using Realsense viewer software developed by Intel.

[0114] Accordingly, concrete slump test results can be obtained using the point cloud obtained using the above software, and the diameter, height, and material segregation risk can be analyzed using an algorithm.

[0115] In particular, the above material separation risk is acquired through an algorithm that automatically performs immediate analysis after point cloud acquisition. Furthermore, as the algorithm predicts the material separation risk in advance before solidification, it differs from the experimental result material separation prediction data that will be output later.

[0116] Ultimately, after acquiring the above point cloud, the diameter, height, and material segregation risk of the concrete test specimen can be derived all at once.

[0117] Meanwhile, FIG. 14a is a diagram showing the axis alignment and noise removal of a point cloud in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and FIG. 14b is a diagram showing a point cloud of a region of interest (ROI).

[0118] In the case of a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, as shown in FIG. 14a, axis alignment and noise removal are performed on the point cloud collected by the data collection unit (111).

[0119] Here, the axis alignment is performed to flatten the angle that is inevitably tilted when photographing the concrete test specimen (300) through the depth measuring means (200) by rotating it. At this time, FIG. 14b shows the point cloud / RGB or the point cloud alone of the region of interest (ROI), respectively.

[0120] Meanwhile, FIG. 15 is a diagram schematically illustrating a machine learning algorithm in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and

[0121] FIGS. 16a to 16c are drawings specifically illustrating the output of a machine learning algorithm in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.

[0122] In the case of the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, the machine learning module may be implemented as a Deep Neural Network (DNN), but is not limited thereto. For example, as shown in FIG. 15, the machine learning module utilizes a Deep Neural Network (DNN) analysis tool, and the construction of such a DNN model is described in detail as follows.

[0123] First, a Deep Neural Network (DNN) is an Artificial Neural Network (ANN) that includes multiple hidden layers between an input layer and an output layer, as illustrated in FIG. 15.

[0124] Such deep neural networks (DNNs) are a method of machine learning and can learn various non-linear relationships by including multiple hidden layers, and such deep neural network (DNN) analysis can derive a function equation that predicts certain characteristics based on training data.

[0125] The point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention is configured to take concrete mix ratio, diameter and height, and material segregation risk as inputs, and to output rheological characteristics, compressive strength, and material segregation characteristics, and is excellent in terms of material property evaluation and quality control by providing quantitative results based on such machine learning.

[0126] As a result of performing machine learning with the diameter, height, and curvature of the concrete specimen as input, yield stress and plastic viscosity shown in FIG. 16a, the adsorption ratio of the high-performance superplasticizer shown in FIG. 16b, and bleeding shown in FIG. 16c can be output to evaluate the fluidity of the concrete specimen.

[0127] Meanwhile, FIG. 17 is a diagram illustrating the point cloud analysis process in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and

[0128] FIG. 18 is a diagram showing the diameter and height analysis of a concrete test specimen in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, and

[0129] FIG. 19 is a diagram showing the assessment of material segregation risk for a concrete specimen in a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.

[0130] In the case of the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, a point cloud analysis process as shown in FIG. 17 is performed, and specifically, as shown in FIG. 18, the diameter and height of a concrete test specimen can be analyzed, and additionally, based on the results of the diameter and height analysis, the risk of material segregation can be evaluated as shown in FIG. 19.

[0131] Meanwhile, FIG. 20a is a diagram showing material separation prediction data evaluated by a point cloud and machine learning-based concrete fluidity automatic evaluation system according to an embodiment of the present invention, and FIG. 20b is a diagram specifically showing region A illustrated in FIG. 20a, and

[0132] FIG. 21 is a diagram illustrating material separation characteristics evaluated by a point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention.

[0133] In the case of the point cloud and machine learning-based automatic concrete fluidity evaluation system according to an embodiment of the present invention, quantitative concrete properties (e.g., rheological properties before hardening, compressive strength after hardening, and aggregate segregation properties) can be automatically derived by utilizing machine learning to utilize the concrete mix ratio and the point cloud information (diameter, height, and risk of segregation).

[0134] For example, as shown in FIGS. 20a and 20b, material separation prediction data can be derived. Additionally, as shown in FIG. 21, material separation characteristics for aggregates can be derived.

[0135] Ultimately, according to the point cloud and machine learning-based automatic concrete fluidity evaluation system according to the embodiment of the present invention, the diameter, height, and material segregation risk of a concrete test specimen are generated as point cloud information, and through a machine learning-based concrete fluidity evaluation module, the rheological characteristics of the concrete test specimen before hardening are extracted, the compressive strength after hardening is predicted, and the material segregation characteristics are predicted.

[0136] [Point Cloud and Machine Learning-Based Automated Concrete Fluidity Assessment Method]

[0137] FIG. 22 is a flowchart illustrating a point cloud and machine learning-based automatic concrete fluidity evaluation method according to an embodiment of the present invention.

[0138] Referring to FIG. 22, the point cloud and machine learning-based automatic concrete fluidity evaluation method according to an embodiment of the present invention is,

[0139] First, a concrete test specimen (300) for evaluating concrete fluidity is prepared and placed (S110).

[0140] Next, a point cloud of three-dimensional shape information is generated by photographing a concrete test specimen (300) before hardening using a depth measuring means (200) (S120).

[0141] For example, the depth measuring means (200) acquires point cloud and RGB images of the concrete fluidity test of the concrete test specimen before hardening using RealSense viewer software.

[0142] Next, the point cloud information generation module (110) collects a point cloud from the depth measuring means (200) (S130).

[0143] Next, the point cloud information generation module (110) performs axis alignment and noise removal of the collected point cloud (S140).

[0144] Specifically, the point cloud information generation module (110) collects a point cloud obtained through the depth measuring means (200), performs axis alignment and noise removal on the collected point cloud, and analyzes the point cloud to generate point cloud information including the diameter, height, and material separation risk of the concrete test specimen (300).

[0145] Here, the axis alignment is characterized by rotating the angle that is inevitably tilted when photographing the concrete test specimen (300) through the depth measuring means (200) to flatten it.

[0146] Next, the point cloud information generation module (110) analyzes the point cloud to derive point cloud information including the diameter, height, and material separation risk of the concrete test specimen (300) (S150).

[0147] Next, a machine learning-based concrete fluidity evaluation module (120) performs machine learning using concrete mix ratio and point cloud information as input (S160).

[0148] At this time, as a result of performing the above machine learning, yield stress, plastic viscosity, adsorption ratio, and bleeding are output to evaluate the fluidity of the concrete test specimen.

[0149] Next, the machine learning-based concrete fluidity evaluation module (120) automatically evaluates the quantitative concrete properties of the concrete test specimen (300) (S170).

[0150] Specifically, the machine learning-based concrete fluidity evaluation module (120) performs machine learning using the concrete mix ratio for the concrete test specimen (300) and the point cloud information generated by the point cloud information generation module (110) as input, and automatically evaluates the concrete fluidity to derive quantitative concrete properties for the concrete test specimen (300).

[0151] Here, the quantitative concrete properties for the concrete test specimen (300) may include rheological properties before concrete hardening, compressive strength after hardening, and material segregation properties.

[0152] Ultimately, according to an embodiment of the present invention, for a concrete test specimen for fluidity evaluation, a point cloud, which is three-dimensional shape information, is obtained through a depth measuring means, and concrete fluidity characteristics are analyzed using artificial intelligence-based machine learning, thereby enabling accurate automatic evaluation of concrete fluidity.

[0153] In addition, compared to conventional qualitative evaluation methods by operators, the quality of ready-mix concrete and on-site concrete can be ensured by utilizing point clouds and machine learning to quantitatively, quickly, and accurately evaluate concrete fluidity.

[0154] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical spirit or essential features of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0155] The scope of the present invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof should be interpreted as being included within the scope of the present invention. Explanation of the symbols

[0156] 100: Data analysis terminal 200; Depth measuring means (Depth Camera) 300: Concrete test specimen 110: Point Cloud Information Generation Module 111: Data Collection Unit 112: Data Processing Unit 113: Data Analysis Department 120: Machine Learning-Based Concrete Fluidity Evaluation Module 121: Machine Learning Module 122: Concrete Flowability Evaluation Module 122a: Rheology Property Extraction Unit 122b: Compressive strength prediction section 122c: Material segregation characteristic prediction section 130: Liquidity Assessment DB 140: Display

Claims

Claim 1 The system includes a data analysis terminal (100) that generates point cloud information by collecting and analyzing a point cloud including the diameter, height, and material segregation risk of a concrete test specimen (300) obtained through a depth measuring means (200), and automatically evaluates concrete fluidity to derive quantitative concrete properties for the concrete test specimen (300); wherein the data analysis terminal (100) performs machine learning using the point cloud information generated by a point cloud information generation module (110) that collects and analyzes the point cloud obtained through the depth measuring means (200) and generates point cloud information for the concrete test specimen (300), automatically evaluates concrete fluidity, and includes a machine learning-based concrete fluidity evaluation module (120) that derives quantitative concrete properties for the concrete test specimen (300); wherein the machine learning-based concrete fluidity evaluation module (120) includes the concrete mix ratio for the concrete test specimen (300), the diameter, height, and A point cloud and machine learning-based automatic concrete fluidity evaluation system comprising: a machine learning module (121) that performs machine learning using point cloud information including material separation risk as input; and a concrete fluidity evaluation module (122) that automatically evaluates concrete fluidity according to the learning result of the machine learning module (121) to derive quantitative concrete properties for a concrete test specimen (300), wherein the concrete fluidity evaluation module (122) comprises: a rheology characteristic extraction unit (122a) that extracts rheology characteristics before hardening of the concrete test specimen (300); a compressive strength prediction unit (122b) that predicts the compressive strength after hardening of the concrete test specimen (300); and a material separation characteristic prediction unit (122c) that predicts material separation characteristics for aggregates of the concrete test specimen (300). Claim 2 In claim 1, the quantitative concrete properties for the concrete test specimen (300) include a point cloud including rheological properties before concrete hardening, compressive strength after hardening, and material segregation properties, and a machine learning-based automatic evaluation system for concrete fluidity. Claim 3 A point cloud and machine learning-based automatic concrete fluidity evaluation system according to claim 1, wherein the depth measuring means (200) acquires point cloud and RGB images of a concrete fluidity test before hardening of a concrete test specimen using Realsense viewer software. Claim 4 In claim 1, the data analysis terminal (100) further comprises: a point cloud information generation module (110) that generates point cloud information by collecting and analyzing a point cloud obtained through a depth measuring means (200); a fluidity evaluation DB (130) that stores concrete fluidity evaluation results obtained through experiments on concrete test specimens and quantitative concrete physical property evaluation results derived from the machine learning-based concrete fluidity evaluation module (120); and a display (140) that outputs the quantitative concrete physical property evaluation results derived from the machine learning-based concrete fluidity evaluation module (120), thereby forming a point cloud and machine learning-based concrete fluidity automatic evaluation system. Claim 5 In claim 1, the point cloud information generation module (110) comprises: a data collection unit (111) that collects a point cloud obtained through the depth measuring means (200); a data processing unit (112) that performs axis alignment and noise removal on the point cloud collected by the data collection unit (111); and a data analysis unit (113) that analyzes the point cloud to generate point cloud information including the diameter, height, and material segregation risk of the concrete test specimen (300), thereby forming a point cloud and machine learning-based automatic concrete fluidity evaluation system. Claim 6 In claim 5, the axis alignment of the data processing unit (112) is characterized by rotating the angle of inevitably tilted when photographing the concrete test specimen (300) through the depth measuring means (200) to flatten it, in a point cloud and machine learning-based automatic concrete fluidity evaluation system. Claim 7 delete Claim 8 A point cloud and machine learning-based automatic concrete fluidity evaluation system, wherein, in claim 1, the machine learning module (121) outputs yield stress, plastic viscosity, adsorption ratio, and bleeding as a result of performing machine learning to evaluate the fluidity of a concrete test specimen. Claim 9 delete Claim 10 A method for automatically evaluating concrete fluidity based on a point cloud and machine learning using a point cloud and machine learning-based concrete fluidity evaluation system according to claim 1, comprising: a) a step of manufacturing and arranging a concrete test specimen (300) for concrete fluidity evaluation; b) a step of generating a point cloud, which is three-dimensional shape information, by photographing the concrete test specimen (300) before hardening through a depth measuring means (200); c) a step in which a point cloud information generation module (110) collects the point cloud from the depth measuring means (200); d) a step of analyzing the point cloud to derive point cloud information including the diameter, height, and material segregation risk of the concrete test specimen (300); e) a step in which a machine learning-based concrete fluidity evaluation module (120) performs machine learning using the concrete mix ratio and point cloud information as input; and f) a step in which the machine learning-based concrete fluidity evaluation module (120) automatically evaluates the quantitative concrete physical properties of the concrete test specimen (300). Claim 11 A point cloud and machine learning-based automatic concrete fluidity evaluation method according to claim 10, characterized in that, before performing step d), the point cloud information generation module (110) performs axis alignment and noise removal of the collected point cloud. Claim 12 In claim 10, the quantitative concrete properties for the above concrete test specimen (300) include a point cloud including rheological properties before concrete hardening, compressive strength after hardening, and material segregation properties, and a machine learning-based automatic evaluation method for concrete fluidity. Claim 13 In claim 10, the depth measuring means (200) is characterized by acquiring a point cloud and RGB image of the concrete fluidity test before hardening of the concrete test specimen using RealSense Viewer software, in a point cloud and machine learning-based automatic evaluation method for concrete fluidity. Claim 14 In claim 10, the point cloud information generation module (110) collects a point cloud obtained through the depth measuring means (200), performs axis alignment and noise removal on the collected point cloud, and analyzes the point cloud to generate point cloud information including the diameter, height, and material segregation risk of the concrete test specimen (300), thereby forming a point cloud and machine learning-based automatic concrete fluidity evaluation method. Claim 15 In claim 14, the above-mentioned axis alignment is characterized by rotating the angle of inevitably tilted when photographing a concrete test specimen (300) through the depth measuring means (200) to flatten it, in a point cloud and machine learning-based automatic evaluation method for concrete fluidity. Claim 16 In claim 10, the machine learning-based concrete fluidity evaluation module (120) performs machine learning using the concrete mix ratio for the concrete test specimen (300) and the point cloud information generated by the point cloud information generation module (110) as input, and automatically evaluates the concrete fluidity to derive quantitative concrete properties for the concrete test specimen (300), thereby forming a point cloud and machine learning-based method for automatically evaluating concrete fluidity. Claim 17 A point cloud and machine learning-based automatic evaluation method for concrete fluidity, characterized in that, in claim 16, yield stress, plastic viscosity, water absorption rate, and bleeding are output as a result of performing the machine learning above to evaluate the fluidity of a concrete test specimen.

Images