A real-time control method and device for concrete formula based on vision

Abstract

The present invention provides a visual-based real-time control method for concrete formula and a device thereof, wherein the method includes a material property detection method, a mixing quality detection method, and a production quality adjustment formula method. Through the above structure, the present invention can effectively predict the working performance of the concrete mixing process and detect aggregate gradation and particle shape parameters in real time through the cooperation of the material property detection method, the mixing quality detection method, and the production quality adjustment formula method, thereby reducing the necessary performance detection time in the production process and improving production efficiency. When the performance of the current batch of concrete does not meet the requirements, the amount of aggregate to be supplemented is calculated through the current aggregate gradation and particle shape parameters, and adjusted in real time to ensure that the performance of the concrete can meet the requirements when it is discharged, reduce resource waste, and improve the efficiency of adjusting the formula.

Description

Technical Field

[0001] The present invention relates to the technical field of concrete formula, and in particular to a vision-based real-time control method and device for concrete formula. Background Art

[0002] The traditional concrete production model has problems such as energy waste, large carbon emissions and excessive consumption of raw materials. The main source of carbon emissions in commercial concrete is cement, which accounts for more than 90% of the carbon emissions in the entire concrete production process. Under the special domestic supply and demand market, the shortage of high-quality raw materials has led to fluctuations in their parameters. The compatibility problem between water reducers and concrete raw materials has become prominent. In addition, the grading, moisture content and fluctuations of sand and gravel aggregates and stone powder in aggregates have increased the instability of indicators such as concrete strength. To ensure the quality of concrete, the mixing station chooses to increase the amount of cement. Cement fills the gaps between aggregates in concrete, especially when the aggregate grading and particle shape are not ideal, reduces the friction between aggregates, and provides excellent concrete fluidity and stability. However, this leads to a negative cycle: in order to solve the quality problems that may be caused by raw material fluctuations, the mixing station must increase the amount of cement, further exacerbating the carbon emission problem, thereby increasing the environmental burden while improving the quality of concrete.

[0003] In order to minimize cement consumption and carbon emissions, the entire concrete production process must be monitored and intelligently controlled in real time to better feedback and adjust the formula parameters. There are two key stages: the raw material entry stage; the production mixing and unloading stage.

[0004] Among the raw materials, fluctuations in the properties of sand and coarse aggregate have a significant impact on concrete. Before the raw materials are put into the mixing tank, it is necessary to detect and determine their actual parameters such as particle grading, porosity, moisture content, etc. Traditional material performance measurement methods often rely on offline laboratory testing and sample collection, which are cumbersome and inefficient. Now some digital image processing technologies such as filtering and thresholding methods are effective in discrete aggregate conditions, but cannot cope with complex stacking conditions. For moisture content detection, in addition to the traditional drying method, researchers have successively proposed resistance and capacitance methods and neutron detection methods, but these contact methods have large errors, poor repeatability, and are easily affected by adherent materials. Non-contact methods include infrared sensor measurement methods, but they are sensitive to particle size and stacking, and can only measure surface moisture, which is affected by the surface color of sand and gravel, causing deviations.

[0005] There is no perfect monitoring system and mature implementation plan for the concrete production and mixing stage. Currently, it mainly relies on product sampling of manual slump cone test, which has errors and lags. The traditional method is affected by operator factors and differences in test conditions, resulting in inaccurate measurement results, making it difficult to monitor and adjust concrete fluidity in real time, affecting the immediate control and adjustment of the construction process. A small number of studies have used grayscale co-occurrence matrices and other methods to extract concrete surface texture features, and then used machine learning to regress and fit the slump, but these solutions are all in ideal working conditions in the laboratory and have no practical application in the actual concrete processing link.

[0006] Therefore, there is a need for a vision-based real-time control method and device for concrete formula, which can automatically adjust formula parameters such as water-cement ratio and admixture dosage through intelligent monitoring and big data analysis at different stages of the concrete production chain, timely avoid the production of non-compliant concrete, reduce waste, and ensure that the fluidity and plasticity of concrete can meet structural requirements. Summary of the invention

[0007] The invention discloses a vision-based real-time control method and device for concrete formula, which mainly solves the problem that it is impossible to monitor and intelligently control the raw material entry stage, production mixing and unloading stages in the concrete production process in real time.

[0008] To achieve the above object, the technical solution of the present invention is implemented as follows:

[0009] In a first aspect, the present invention provides a vision-based real-time control method for concrete formula, including a material property detection method, a mixing quality detection method, and a production quality adjustment formula method;

[0010] The material property detection method includes the steps of collecting aggregate images, segmenting aggregate images and calculating properties, analyzing the moisture content of aggregates, and adjusting the pre-formula. The mixing quality detection method includes the step of detecting the flow rate of the slurry during mixing. The production quality adjustment formula method includes the step of adjusting the formula according to the production quality.

[0011] An aggregate image acquisition step is to acquire an aggregate image acquired before mixing;

[0012] Aggregate image segmentation and characteristic calculation steps: the collected aggregate image is used to obtain masks of different particles using the instance segmentation algorithm MaskTransfiner, and then the masks of different particles are separated from the background by image processing methods, and the aggregate gradation is calculated based on the equivalent representation form of aggregates, and the aggregate porosity is inferred by multivariate regression based on the particle shape parameters corresponding to the calculated aggregate gradation;

[0013] Aggregate moisture content analysis step, obtaining the actual aggregate microwave moisture content collected before mixing;

[0014] Adjust the pre-formulation step, match the obtained aggregate characteristics with the pre-formulation adjustment rules, and automatically adjust the pre-formulation parameters of the concrete according to the matching results;

[0015] The step of detecting the flow velocity of the slurry during the mixing process is to obtain the video of the concrete mixing process, process it according to the preset frame rate, decompose it into an image frame sequence, and preprocess and optimize the decomposed image. The preprocessed image frame sequence is processed using the particle image velocimetry algorithm to calculate the velocity field of multiple consecutive frames. By analyzing the data of the velocity field, the average flow velocity of the concrete slurry is calculated. With the support of a large amount of order data, the regression analysis of slump and flow velocity is carried out, and the conversion relationship model of slump and flow velocity is established to obtain the predicted slump.

[0016] According to the production quality adjustment formula steps, the predicted slump is compared with the standard value on the order. When the slump is within the reasonable range of the standard value, it is judged that the quality of the concrete currently produced is good. When the slump deviates from the standard value range, the proportion of water, cement and admixture in the initial formula concrete is adjusted based on the slump deviation, and the material property detection method, mixing quality detection method and production quality adjustment formula method are repeated to achieve better slump control.

[0017] In one embodiment, the mixing quality detection method also includes the step of controlling the opening of the mixing tank discharge port, setting a flow rate threshold value related to the discharge port opening according to the calculated average flow rate of the concrete slurry, and controlling the discharge port opening according to the flow rate threshold value.

[0018] In one embodiment, the particle shape parameters in the aggregate image segmentation and property calculation steps include sphericity, shape factor and angularity.

[0019] In one embodiment, the aggregate properties include aggregate grading, particle shape parameters, aggregate porosity, and microwave moisture content.

[0020] In one embodiment, the preformulation adjustment rules include the water-cement ratio adjustment relationship corresponding to different gradations, the relationship between particle shape parameters and cohesiveness, and the relationship between porosity and sand content.

[0021] In one embodiment, the pre-formula parameters in the step of adjusting the pre-formula include water-cement ratio, cement dosage, and sand content.

[0022] In one embodiment, in the step of controlling the opening of the discharge port of the mixing tank, when the flow rate is greater than a threshold value, the opening of the discharge port is controlled to decrease; when the flow rate is less than the threshold value, the opening of the discharge port is controlled to increase.

[0023] The advantages or beneficial effects of the above technical solution include at least: through the cooperation of material property detection method, mixing quality detection method and production quality adjustment formula method, it is possible to effectively predict the working performance of the concrete mixing process and detect aggregate grading and particle shape parameters in real time, reduce the necessary performance detection time in the production process, and improve production efficiency. When the performance of the current batch of concrete does not meet the requirements, the amount of aggregate to be supplemented is calculated through the current aggregate grading and particle shape parameters, and adjusted in real time to ensure that the performance of the concrete can meet the requirements when it is discharged, reduce resource waste, and improve the efficiency of adjusting the formula.

[0024] In a second aspect, the present invention provides a vision-based real-time control device for concrete formula, comprising:

[0025] Material property detection module, used to obtain the image of aggregate before mixing and the microwave moisture content of aggregate;

[0026] An image segmentation calculation module is used to segment the obtained aggregate image before mixing to obtain masks of different particles using the instance segmentation algorithm Mask Transfiner, and then separate the masks of different particles from the background through an image processing method, and calculate the aggregate gradation based on the equivalent representation form of the aggregate, and perform multiple regression based on the particle shape parameters corresponding to the calculated aggregate gradation to infer the aggregate porosity;

[0027] A preformula adjustment module is used to match the obtained aggregate properties with the preformula adjustment rules, and automatically adjust the preformula parameters of the concrete according to the matching results. The preformula adjustment rules include the water-cement ratio adjustment corresponding to different gradations, the relationship between the particle shape parameters and the amount of binder, and the relationship between the porosity and the sand content;

[0028] The mixing quality detection module is used to obtain the video of the concrete mixing process, process it according to the preset frame rate, decompose it into an image frame sequence, and preprocess and optimize the decomposed image. The preprocessed image frame sequence is processed using the particle image velocimetry algorithm to calculate the velocity field of multiple consecutive frames. By analyzing the data of the velocity field, the average flow velocity of the concrete paste is calculated. With the support of a large amount of order data, the slump and flow velocity regression analysis is performed, and a conversion relationship model between the slump and flow velocity is established to obtain the predicted slump.

[0029] The discharge port opening control module is used to set a flow rate threshold value according to the relationship between the average flow rate of the concrete slurry and the discharge port opening, and adjust the opening size of the discharge port according to the threshold value;

[0030] The production quality formula adjustment module is used to compare the predicted slump with the order standard value and judge the concrete quality based on the comparison result. When the predicted slump deviates from the standard value range, the ingredient ratio of the initial formula concrete is adjusted and tested again through the material property detection module, image segmentation calculation module, pre-formula adjustment module and mixing quality detection module.

[0031] The advantages or beneficial effects of the above technical solution include at least: through the cooperation of the material property detection module, the image segmentation calculation module, the pre-formula adjustment module, the mixing quality detection module, the discharge port opening control module and the production quality formula adjustment module, the working performance of the concrete mixing process can be predicted in real time and effectively, and the aggregate grading and particle shape parameters can be detected, thereby reducing the necessary performance detection time in the production process and improving production efficiency; when the performance of the current batch of concrete does not meet the requirements, the amount of aggregate to be supplemented can be calculated through the current aggregate grading and particle shape parameters, and adjusted in real time to ensure that the performance of the concrete can meet the requirements when it is discharged, reduce resource waste, and improve the efficiency of formula adjustment. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The accompanying drawings illustrate exemplary embodiments of the present invention and together with the description serve to explain the principles of the present invention. These drawings are included to provide a further understanding of the present invention and are incorporated in and constitute a part of this specification.

[0033] Figure 1 A schematic diagram showing an overall device according to an exemplary embodiment of the present invention is shown;

[0034] Figure 2 A schematic diagram of a segmentation model of an aggregate image according to an exemplary embodiment of the present invention is shown;

[0035] Figure 3 A schematic diagram showing the test results of aggregate moisture content detection according to an exemplary embodiment of the present invention is shown;

[0036] Figure 4 A schematic diagram of a concrete formula real-time control process according to an exemplary embodiment of the present invention is shown;

[0037] Figure 5 FIG. 4 is a schematic diagram showing estimation of an image slurry velocity field according to an exemplary embodiment of the present invention.

[0038] Description of reference numerals:

[0039] 1. Industrial camera; 2. LED light source; 3. Aggregate; 4. Belt; 5. Receiving probe of microwave detector; 6. Transmitting probe of microwave detector; 7. Camera; 8. Mixing tank; 9. Light source; 10. Computer. DETAILED DESCRIPTION

[0040] Embodiments of the present invention will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present invention are shown in the accompanying drawings, it should be understood that the present invention can be implemented in various forms and should not be construed as being limited to the embodiments described herein, which are instead provided for a more thorough and complete understanding of the present invention. It should be understood that the drawings and embodiments of the present invention are only for exemplary purposes and are not intended to limit the scope of protection of the present invention.

[0041] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0042] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0043] The term "including" and its variations used in this document are open inclusions, that is, "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one other embodiment"; the term "some embodiments" means "at least some embodiments". Relevant definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc. mentioned in the present invention are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0044] It should be noted that the modifications of "one" and "plurality" mentioned in the present invention are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0045] The names of the messages or information exchanged between multiple devices in the embodiments of the present invention are only used for illustrative purposes, and are not used to limit the scope of these messages or information.

[0046] See also Figures 1 to 5 , the embodiment of the present invention provides a vision-based real-time control method for concrete formula, including a material property detection method, a mixing quality detection method, and a production quality adjustment formula method;

[0047] The material property detection method includes the steps of collecting aggregate images, segmenting aggregate images and calculating properties, analyzing the moisture content of aggregates, and adjusting the pre-formula. The mixing quality detection method includes the steps of detecting the flow rate of the slurry during mixing and controlling the opening of the mixing tank discharge port. The production quality adjustment formula method includes the step of adjusting the formula according to the production quality.

[0048] S1. Aggregate image acquisition step, obtaining aggregate images collected before mixing;

[0049] For details, see Figure 1 An aggregate image detection device is arranged on a belt 4 for conveying aggregate 3. The image detection device includes an industrial camera 1 and an LED light source 2. The industrial camera 1 is sealed with a proportional partition to prevent long-term dust from damaging the camera lens. LED light sources 2 are arranged on both sides of the industrial camera 1 to illuminate the belt, which can avoid the aggregate above the belt 4 from generating shadows, so as to ensure the subsequent aggregate image segmentation and calculation accuracy. During the process of conveying aggregate by the belt 4, the industrial camera 1 collects aggregate images at high speed and transmits them to the computer 10 through the network card.

[0050] S2. Aggregate image segmentation and characteristic calculation step, using the instance segmentation algorithm MaskTransfiner to obtain masks of different particles from the collected aggregate image, and then separating the masks of different particles from the background through image processing methods, and calculating the aggregate gradation based on the equivalent representation form of the aggregate, and performing multivariate regression based on the particle shape parameters corresponding to the calculated aggregate gradation to infer the aggregate porosity; wherein the particle shape parameters include but are not limited to sphericity, shape factor and angularity;

[0051] S3. Aggregate moisture content analysis step, obtaining the actual aggregate microwave moisture content collected before mixing;

[0052] For details, see Figure 1 A microwave detector device is arranged on the belt 4. The microwave detector device includes a receiving probe 5 of the microwave detector and a transmitting probe 6 of the microwave detector. The microwave detector device calculates the instantaneous height according to the attenuation of the microwave energy value using an ultrasonic height compensation probe, and then calculates the moisture content of the material according to the ratio of the calculated total microwave energy loss and the instantaneous height value. The measurement range of the moisture content of the aggregate is the overall moisture content of the material, not just the surface layer, and has higher accuracy. The conventional accuracy of the microwave detector device is ±0.5%, and the maximum can reach ±0.3%. The test error is as follows: Figure 3 During the process of conveying aggregates by the belt 4, the microwave detector device detects the moisture content of the aggregates in real time and transmits it to the computer 10 through the network card.

[0053] S4. Adjust the preformula step, matching the obtained aggregate properties with the preformula adjustment rules, and automatically adjusting the preformula parameters of the concrete according to the matching results; wherein the aggregate properties include aggregate gradation, particle shape parameters, aggregate porosity, and microwave moisture content; the preformula adjustment rules include but are not limited to the water-cement ratio adjustment relationship corresponding to different gradations, the relationship between particle shape parameters and cohesiveness, and the relationship between porosity and sand content; the preformula parameters include but are not limited to water-cement ratio, cement dosage, and sand content;

[0054] For details, see Figure 1 , aggregate properties and pre-formula adjustment rules are formulated in the computer 10, and then the obtained aggregate properties are matched with the pre-formula adjustment rules through fitting analysis by the computer 10. According to the matching results, the pre-formula parameters of the concrete are automatically adjusted.

[0055] S5. The step of detecting the flow rate of the slurry during the mixing process is to obtain the video of the concrete mixing process, process it according to the preset frame rate, and decompose it into an image frame sequence; in order to eliminate the image blur caused by the flying of powder, the decomposed image is preprocessed and optimized to ensure the accuracy of subsequent analysis; the preprocessed image frame sequence is processed using the particle image velocimetry algorithm to calculate the velocity field of multiple consecutive frames, such as Figure 5 As shown in the figure, this process can accurately capture the movement of concrete paste; by analyzing the velocity field data, the average flow rate of concrete paste is calculated, and with the support of a large amount of order data, the slump and flow rate regression analysis is carried out, and the conversion relationship model between slump and flow rate is established to obtain the predicted slump, so as to evaluate the production quality of concrete in real time;

[0056] For details, see Figure 1 A mixing cylinder 8 is provided at the discharge end of the belt 4, and a camera 7 and a light source 9 are provided on the mixing cylinder 8. The camera 7 is used to capture a video of the concrete mixing process in the mixing cylinder 8, and transmit the video to a computer 10, and the computer 10 processes the video as in step S5.

[0057] S6. A step of controlling the opening of the discharging port of the mixing tank, wherein a flow rate threshold value related to the opening of the discharging port is set according to the calculated average flow rate of the concrete slurry, and the opening of the discharging port is controlled according to the flow rate threshold value; that is, when the flow rate is greater than the threshold value, the opening of the discharging port is controlled to be reduced to prevent overflow caused by a large amount of slurry being put into the receiving port of the mixer truck in a short time; when the flow rate is less than the threshold value, the opening of the discharging port is controlled to be increased to prevent blockage of the discharging port and improve the discharging efficiency;

[0058] S7. According to the production quality adjustment formula step, the predicted slump is compared with the standard value on the order. When the slump is within the reasonable range of the standard value, it is judged that the quality of the concrete currently produced is good; when the slump deviates from the standard value range, the initial formula of concrete water, cement, admixture ratio is adjusted based on the slump deviation, and the Figure 4 Repeat material property testing methods, mixing quality testing methods and production quality adjustment formula methods to achieve better slump control.

[0059] By adopting the above structure, through the cooperation of material property detection method, mixing quality detection method and production quality adjustment formula method, it is possible to effectively predict the working performance of the concrete mixing process and detect aggregate gradation and particle shape parameters in real time, reduce the necessary performance detection time in the production process, and improve production efficiency. When the performance of this batch of concrete does not meet the requirements, the amount of aggregate to be supplemented is calculated based on the current aggregate gradation and particle shape parameters, and adjusted in real time to ensure that the performance of the concrete can meet the requirements when it is discharged, reduce resource waste, and improve the efficiency of adjusting the formula.

[0060] See also Figure 1 and Figure 2 , an embodiment of the present invention provides a vision-based real-time control device for concrete formula, including a material property detection module for acquiring an image of aggregate before mixing and a microwave moisture content of the aggregate;

[0061] An image segmentation calculation module is used to segment the obtained aggregate image before mixing to obtain masks of different particles using the instance segmentation algorithm Mask Transfiner, and then separate the masks of different particles from the background through an image processing method, and calculate the aggregate gradation based on the equivalent representation form of the aggregate, and perform multiple regression based on the particle shape parameters corresponding to the calculated aggregate gradation to infer the aggregate porosity;

[0062] A preformula adjustment module is used to match the obtained aggregate properties with the preformula adjustment rules, and automatically adjust the preformula parameters of the concrete according to the matching results. The preformula adjustment rules include the water-cement ratio adjustment corresponding to different gradations, the relationship between the particle shape parameters and the amount of binder, and the relationship between the porosity and the sand content;

[0063] The mixing quality detection module is used to obtain the video of the concrete mixing process, process it according to the preset frame rate, decompose it into an image frame sequence, and preprocess and optimize the decomposed image. The preprocessed image frame sequence is processed using the particle image velocimetry algorithm to calculate the velocity field of multiple consecutive frames. By analyzing the data of the velocity field, the average flow velocity of the concrete paste is calculated. With the support of a large amount of order data, the slump and flow velocity regression analysis is performed, and a conversion relationship model between the slump and flow velocity is established to obtain the predicted slump.

[0064] The discharge port opening control module is used to set a flow rate threshold value according to the relationship between the average flow rate of the concrete slurry and the discharge port opening, and adjust the opening size of the discharge port according to the threshold value;

[0065] The production quality formula adjustment module is used to compare the predicted slump with the order standard value and judge the concrete quality based on the comparison result. When the predicted slump deviates from the standard value range, the ingredient ratio of the initial formula concrete is adjusted and tested again through the material property detection module, image segmentation calculation module, pre-formula adjustment module and mixing quality detection module.

[0066] In the description of the present invention, it should be noted that the terms "center", "up", "down", "left", "right", "vertical", "horizontal", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they cannot be understood as limitations on the present invention.

[0067] It should be understood by those skilled in the art that the above embodiments are only for the purpose of clearly illustrating the present invention, and are not intended to limit the scope of the present invention. For those skilled in the art, other changes or modifications may be made based on the above invention, and these changes or modifications are still within the scope of the present invention.

Claims

1. A real-time control method for concrete formula based on vision, characterized in that: Including material property testing methods, mixing quality testing methods, production quality adjustment formula methods; The material property detection method includes the steps of collecting aggregate images, segmenting aggregate images and calculating properties, analyzing the moisture content of aggregates, and adjusting the pre-formula. The mixing quality detection method includes the step of detecting the flow rate of the slurry during mixing. The production quality adjustment formula method includes the step of adjusting the formula according to the production quality. An aggregate image acquisition step is to acquire an aggregate image acquired before mixing; Aggregate image segmentation and characteristic calculation steps: the collected aggregate image is used to obtain masks of different particles using the instance segmentation algorithm MaskTransfiner, and then the masks of different particles are separated from the background by image processing methods, and the aggregate gradation is calculated based on the equivalent representation form of aggregates, and the aggregate porosity is inferred by multivariate regression based on the particle shape parameters corresponding to the calculated aggregate gradation; Aggregate moisture content analysis step, obtaining the actual aggregate microwave moisture content collected before mixing; Adjust the pre-formulation step, match the obtained aggregate characteristics with the pre-formulation adjustment rules, and automatically adjust the pre-formulation parameters of the concrete according to the matching results; The step of detecting the flow velocity of the slurry during the mixing process is to obtain the video of the concrete mixing process, process it according to the preset frame rate, decompose it into an image frame sequence, and preprocess and optimize the decomposed image. The preprocessed image frame sequence is processed using the particle image velocimetry algorithm to calculate the velocity field of multiple consecutive frames. By analyzing the data of the velocity field, the average flow velocity of the concrete slurry is calculated. With the support of a large amount of order data, the regression analysis of slump and flow velocity is carried out, and the conversion relationship model of slump and flow velocity is established to obtain the predicted slump. According to the production quality adjustment formula step, the predicted slump is compared with the standard value on the order. When the slump is within the reasonable range of the standard value, it is judged that the quality of the concrete currently produced is good; when the slump deviates from the standard value range, the initial formula concrete water, cement, admixture ratio is adjusted based on the slump deviation, and the material property detection method, mixing quality detection method and production quality adjustment formula method are repeated to achieve better slump control; The mixing quality detection method further comprises the step of controlling the opening of the mixing tank discharge port, setting a flow rate threshold value related to the discharge port opening according to the calculated average flow rate of the concrete slurry, and controlling the discharge port opening according to the flow rate threshold value; The aggregate properties include aggregate gradation, particle shape parameters, aggregate porosity, and microwave moisture content; The preformula adjustment rules include the water-cement ratio adjustment relationship corresponding to different gradations, the relationship between particle shape parameters and cohesiveness, and the relationship between porosity and sand content.

2. The method for real-time control of concrete formula based on vision according to claim 1, characterized in that: The particle shape parameters in the aggregate image segmentation and characteristic calculation steps include sphericity, shape factor and angularity.

3. The method for real-time control of concrete formula based on vision according to claim 1, characterized in that: The preformula parameters in the step of adjusting the preformula include water-cement ratio, cement dosage, and sand content.

4. The method for real-time control of concrete formula based on vision according to claim 1, characterized in that: In the step of controlling the opening of the mixing tank discharge port, when the flow rate is greater than a threshold value, the discharge port opening is controlled to decrease; when the flow rate is less than the threshold value, the discharge port opening is controlled to increase.

5. A real-time control device for concrete formula based on vision, characterized in that: include: Material property detection module, used to obtain the image of aggregate before mixing and the microwave moisture content of aggregate; An image segmentation calculation module is used to segment the obtained aggregate image before mixing to obtain masks of different particles using the instance segmentation algorithm Mask Transfiner, and then separate the masks of different particles from the background through an image processing method, and calculate the aggregate gradation based on the equivalent representation form of the aggregate, and perform multiple regression based on the particle shape parameters corresponding to the calculated aggregate gradation to infer the aggregate porosity; A preformula adjustment module is used to match the obtained aggregate properties with preformula adjustment rules, and automatically adjust the preformula parameters of concrete according to the matching results. The preformula adjustment rules include the water-cement ratio adjustment corresponding to different gradations, the relationship between particle shape parameters and cohesiveness, and the relationship between porosity and sand content. The aggregate properties include aggregate gradation, particle shape parameters, aggregate porosity, and microwave moisture content. The mixing quality detection module is used to obtain the video of the concrete mixing process, process it according to the preset frame rate, decompose it into an image frame sequence, and preprocess and optimize the decomposed image. The preprocessed image frame sequence is processed using the particle image velocimetry algorithm to calculate the velocity field of multiple consecutive frames. By analyzing the data of the velocity field, the average flow velocity of the concrete paste is calculated. With the support of a large amount of order data, the slump and flow velocity regression analysis is performed, and a conversion relationship model between the slump and flow velocity is established to obtain the predicted slump. The discharge port opening control module is used to set a flow rate threshold value according to the relationship between the average flow rate of the concrete slurry and the discharge port opening, and adjust the opening size of the discharge port according to the flow rate threshold value; The production quality formula adjustment module is used to compare the predicted slump with the order standard value and judge the concrete quality based on the comparison result. When the predicted slump deviates from the standard value range, the ingredient ratio of the initial formula concrete is adjusted and tested again through the material property detection module, image segmentation calculation module, pre-formula adjustment module and mixing quality detection module.

Images