WO2023142262A1 - 基于深度学习的混凝土配方调整方法、装置及可读介质 - Google Patents

基于深度学习的混凝土配方调整方法、装置及可读介质 Download PDF

Info

Publication number
WO2023142262A1
WO2023142262A1 PCT/CN2022/084286 CN2022084286W WO2023142262A1 WO 2023142262 A1 WO2023142262 A1 WO 2023142262A1 CN 2022084286 W CN2022084286 W CN 2022084286W WO 2023142262 A1 WO2023142262 A1 WO 2023142262A1
Authority
WO
WIPO (PCT)
Prior art keywords
concrete
aggregate
image
relationship
value
Prior art date
Application number
PCT/CN2022/084286
Other languages
English (en)
French (fr)
Inventor
杨建红
林柏宏
黄文景
房怀英
张宝裕
黄骁明
陈海生
Original Assignee
福建南方路面机械股份有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 福建南方路面机械股份有限公司 filed Critical 福建南方路面机械股份有限公司
Publication of WO2023142262A1 publication Critical patent/WO2023142262A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/40—Image enhancement or restoration using histogram techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T5/00—Image enhancement or restoration
    • G06T5/90—Dynamic range modification of images or parts thereof
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/11—Region-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/12—Edge-based segmentation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection
    • G06T2207/30132—Masonry; Concrete

Definitions

  • the invention relates to the field of industrial intelligent control, in particular to a method, device and readable medium for adjusting concrete formula based on deep learning.
  • Concrete mix design is usually carried out under laboratory conditions, and then fine-tuned when used on the construction site.
  • the ratio is measured under a condition of sand and stone, and sand and stone raw materials from different sources are often used in production.
  • the particle shape parameters and water content of sand and stone have a great influence on the working performance of concrete (including slump, slump expansion, etc.). Therefore, it is necessary to adjust the concrete ratio by measuring sand and stone parameters.
  • the working performance of concrete is measured by relevant testing personnel using testing instruments after the concrete is discharged from the machine.
  • work performance testing must be carried out after adjusting sand and stone materials.
  • the third is that the moisture content of the same batch of sand and stone raw materials will also be different, and the performance of concrete will be different. Therefore, there is an urgent need for a method that can detect sand, stone parameters and concrete work performance in real time during the concrete production process.
  • the purpose of the embodiments of the present application is to propose a method, device and readable medium for adjusting concrete formula based on deep learning, so as to solve the technical problems mentioned in the background technology section above.
  • the embodiments of the present application provide a method for adjusting concrete formula based on deep learning, including the following steps:
  • step S5 it is determined whether it is necessary to adjust the water consumption and/or the aggregate consumption during the concrete mixing process according to the predicted value of the concrete working performance in the mixing image, specifically including:
  • the working performance of concrete is the slump of concrete.
  • the slump is in one-to-one correspondence with the fluctuation range of the gray mean value in the gray mean value change curve.
  • the slump is The change value is in one-to-one correspondence with the change value of the water consumption and/or aggregate consumption.
  • step S5 the predicted value of the work performance of the concrete, the gradation of the aggregate in the concrete mixing process and the second relationship are adjusted during the concrete mixing process.
  • water usage and/or aggregate usage including:
  • the water consumption during the concrete mixing process is increased according to the relationship between the increased value of the slump and the increased value of the water consumption in the second relationship;
  • the predicted value of the slump of the concrete is higher than the preset threshold range, then according to the relationship between the decrease in the slump and the increase in the amount of the aggregate in the second relationship corresponding to the gradation of the aggregate in the concrete mixing process Adjust the amount of aggregate increase during concrete mixing.
  • the target detection model is a trained first Mask-Rcnn neural network
  • the instance segmentation model is a trained second Mask-Rcnn neural network
  • the first Mask-Rcnn neural network and the second Mask-Rcnn neural network
  • the backbone network of the network is Resnet50.
  • the concrete area image is preprocessed in step S1, specifically including:
  • the background part in the binarized concrete region image is filtered out to obtain the processed concrete region image.
  • step S2 the calcHist function in OpenCV is used to calculate the image grayscale histogram, and the image grayscale histogram records the number of pixels corresponding to different grayscale values in the processed concrete area image, and the grayscale The mean is the ratio of the sum of the gray values of all pixels in the processed concrete area image to the number of pixels.
  • the gradation of the aggregate in the concrete mixing process is determined based on the segmentation result in step S4, specifically including:
  • the fitting ellipse and the corresponding short diameter of the contour of each particle are calculated by using the fitEllipse function in OpenCV;
  • the particle size range of each particle is counted to obtain the gradation of the aggregate.
  • the embodiment of the present application provides a concrete formulation adjustment device based on deep learning, including:
  • the mixing image acquisition module is configured to acquire the mixing image during the concrete mixing process, extract the concrete area image in the mixing image through the target detection model, and preprocess the concrete area image to obtain the processed concrete area image;
  • the gray-scale average calculation module is configured to calculate the image gray-scale histogram based on the processed concrete area image, and calculate the gray-scale average value according to the image gray-scale histogram to obtain the gray-scale average value change curve;
  • the work performance prediction module is configured to establish a first relationship between the gray-scale mean change curve and the work performance of the concrete, and determine the predicted value of the work performance of the concrete in the mixing image according to the gray-scale mean change curve and the first relationship;
  • the aggregate grading calculation module is configured to obtain an aggregate image before concrete mixing, segment the aggregate image by using an instance segmentation model, obtain a segmentation result, and determine the aggregate gradation in the concrete mixing process based on the segmentation result;
  • the adjustment module is configured to establish a second relationship between the change value of the work performance of concrete corresponding to the gradation of different aggregates and the change value of water consumption and/or the change value of aggregate consumption, according to the mixing image
  • the predicted value of the working performance of concrete determines whether it is necessary to adjust the amount of water and/or aggregate used in the concrete mixing process, and based on the predicted value of the working performance of the concrete, the gradation of the aggregate in the concrete mixing process and the second relationship Adjust the water consumption and/or aggregate consumption during the concrete mixing process, and repeatedly execute the mixing image acquisition module to the adjustment module so that the working performance of the concrete meets the requirements.
  • the embodiment of the present application provides an electronic device, including one or more processors; a storage device for storing one or more programs, when one or more programs are executed by one or more processors , so that one or more processors implement the method described in any implementation manner of the first aspect.
  • the embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored.
  • the computer program is executed by a processor, the method described in any implementation manner in the first aspect is implemented.
  • the present invention has the following beneficial effects:
  • the present invention establishes a target detection model by collecting images of the concrete mixing process, and judges in real time whether the working performance of the concrete meets the requirements. At the same time, through the real-time collection of sand and stone images on the belt, the instance segmentation model is established, and the grain shape parameters are predicted online. And when the working performance of the concrete does not meet the requirements, the dosage can be adjusted according to the aggregate gradation and particle shape parameters at that moment, and the working performance of the concrete can be verified.
  • the present invention intelligently adjusts the amount of aggregate to meet the requirements of concrete work performance by detecting concrete work performance and particle shape parameters such as sand and stone aggregates in real time, reduces the necessary performance detection time in the production process, and improves production efficiency.
  • the present invention can calculate the amount of aggregate to be supplemented by calculating the aggregate gradation and grain shape at the current moment, and adjust it in real time to ensure the working performance of the concrete when it leaves the machine Meet the requirements, reduce the waste of resources, and improve the efficiency of formula adjustment.
  • Fig. 1 is an exemplary device architecture diagram to which an embodiment of the present application can be applied;
  • Fig. 2 is a schematic flow chart of a method for adjusting concrete formula based on deep learning according to an embodiment of the present invention
  • Fig. 3 is the schematic diagram of the overall equipment of the concrete formula adjustment method based on deep learning of the embodiment of the present invention
  • Fig. 4 is a schematic flow chart of concrete working performance prediction based on a deep learning-based concrete formula adjustment method according to an embodiment of the present invention
  • Fig. 5 is the result figure of the working performance prediction of the concrete of the concrete formulation adjustment method based on deep learning of the embodiment of the present invention
  • Fig. 6 is the input and segmentation result diagram of the instance segmentation model of the concrete formula adjustment method based on deep learning according to the embodiment of the present invention
  • FIG. 7 is a schematic diagram of a concrete formula adjustment device based on deep learning according to an embodiment of the present invention.
  • Fig. 8 is a schematic structural diagram of a computer device suitable for realizing the electronic equipment of the embodiment of the present application.
  • Fig. 1 shows an exemplary device architecture 100 to which the deep learning-based concrete formulation adjustment method or the deep learning-based concrete formulation adjustment device according to the embodiment of the present application can be applied.
  • the device architecture 100 may include terminal devices 101 , 102 , 103 , a network 104 and a server 105 .
  • the network 104 is used as a medium for providing communication links between the terminal devices 101 , 102 , 103 and the server 105 .
  • Network 104 may include various connection types, such as wires, wireless communication links, or fiber optic cables, among others.
  • Terminal devices 101 , 102 , 103 Users can use terminal devices 101 , 102 , 103 to interact with server 105 via network 104 to receive or send messages and the like.
  • Various applications may be installed on the terminal devices 101 , 102 , and 103 , such as data processing applications, file processing applications, and the like.
  • the terminal devices 101, 102, and 103 may be hardware or software.
  • the terminal devices 101, 102, 103 When the terminal devices 101, 102, 103 are hardware, they may be various electronic devices, including but not limited to smart phones, tablet computers, laptop computers, desktop computers and the like.
  • the terminal devices 101, 102, 103 When the terminal devices 101, 102, 103 are software, they can be installed in the electronic devices listed above. It can be implemented as multiple software or software modules (such as software or software modules for providing distributed services), or as a single software or software module. No specific limitation is made here.
  • the server 105 may be a server that provides various services, such as a background data processing server that processes files or data uploaded by the terminal devices 101 , 102 , and 103 .
  • the background data processing server can process the obtained files or data and generate processing results.
  • the deep learning-based concrete formula adjustment method can be executed by the server 105, or by the terminal devices 101, 102, and 103.
  • the deep learning-based concrete formula adjustment device can be It is set in the server 105, and can also be set in the terminal devices 101, 102, 103.
  • terminal devices, networks and servers in Fig. 1 are only illustrative. According to the implementation needs, there can be any number of terminal devices, networks and servers.
  • the above device architecture may not include a network, but only need a server or a terminal device.
  • Fig. 2 shows a kind of concrete formulation adjustment method based on deep learning provided by the embodiment of the present application, comprising the following steps:
  • an image acquisition device is arranged above the mixer, and light sources are arranged on both sides of the image acquisition device to collect image information during concrete mixing in real time.
  • a stand 2 is provided next to the concrete mixer 1 . Fix the light source 3 and the image acquisition device 4 on the support, the image acquisition device 4 can take the mixing image of the concrete mixing process through the feed port of the mixer, and transmit the mixing image to the computer 5 or server, so the concrete mixing can be acquired in real time Stirring image in process. Utilize the target detection model to extract the concrete region (ROI region) in the mixing image to obtain the concrete region image.
  • the target detection model is the trained first Mask-Rcnn neural network, the first Mask-Rcnn neural network
  • the backbone network of the network is Resnet50
  • Resnet50 is an existing neural network model, so its structure will not be repeated here.
  • the target detection model is constructed based on the deep learning neural network model, which identifies and extracts the image of the concrete area, and then predicts the working performance of the concrete according to steps S2 and S3.
  • step S1 the image of the concrete area is preprocessed in step S1, specifically including:
  • the background part in the binarized concrete region image is filtered out to obtain the processed concrete region image.
  • the value of each pixel in the binarized concrete area image is 0 or 1
  • the remaining background interference can be filtered out by setting the pixel value of the background part to 0, which is convenient for subsequent calculation of the average gray value.
  • the calcHist function in OpenCV is used to calculate the grayscale histogram of the image.
  • the image grayscale histogram records the number of pixels corresponding to different grayscale values in the processed concrete area image, the abscissa is the grayscale value, and the ordinate is the number of pixels.
  • the image gray histogram can be extracted by using OpenCV for calculation.
  • the mean gray value is the ratio between the sum of the gray values of all pixels in the processed concrete area image and the number of pixels, that is:
  • Mean gray value the sum of the gray values of all pixels in the image of the processed concrete area/the number of pixels.
  • the average gray value change curve indicates the fluctuation of the average gray value within a certain time range, and the fluctuation range of the average gray value can be obtained.
  • the working performance of the concrete is the slump of the concrete, and in the first relationship, the slump is in one-to-one correspondence with the fluctuation range of the gray average in the gray average change curve.
  • the slump of the concrete is adjusted respectively, and the fluctuation interval of the gray mean value in the gray mean value change curve is obtained according to the numerical range of the concrete slump, for example: slump
  • the corresponding gray value change curve fluctuates from 164 to 166
  • the corresponding gray value change curve fluctuates from 162 to 164
  • the slump is 120
  • the change curve fluctuates between 160 and 162.
  • Concrete with different working properties can be distinguished according to the average value of the gray value during the mixing process, so that it can be judged whether the working performance of the concrete meets the requirements at this moment.
  • the instance segmentation model is a trained second Mask-Rcnn neural network, and the backbone network of the second Mask-Rcnn neural network is Resnet50.
  • Resnet50 is an existing neural network model, so its structure will not be repeated here.
  • the gradation of the aggregate in the concrete mixing process is determined based on the segmentation result, specifically including:
  • the fitting ellipse and the corresponding short diameter of the contour of each particle are calculated by using the fitEllipse function in OpenCV;
  • the particle size range of each particle is counted to obtain the gradation of the aggregate.
  • aggregate gradation refers to the proportion relationship of particles of different particle sizes that make up aggregate; aggregate gradation is mainly divided into continuous gradation and discontinuous gradation (single-grain grade), and continuous gradation mainly refers to the particle size below the maximum particle size. , there are other corresponding particle sizes in sequence, without interruption, in order to fully fill the gaps between aggregates.
  • Discontinuous grading refers to the absence of one or several intermediate particle sizes in continuous grading.
  • Concrete aggregate is an important part of concrete, which plays the role of skeleton and filling in concrete. Usually divided into fine aggregate and coarse aggregate. In concrete, the particle size between 0.155mm and 5mm is generally called fine aggregate; the particle size greater than 5mm is called coarse aggregate.
  • Coarse aggregate is divided into pebbles, crushed stones, crushed pebbles, and a mixture of pebbles and crushed stones according to types.
  • the aggregates are all too fine, which may increase the slump of the concrete material and increase the bleeding, while the coarse aggregate will cause the cohesion of the concrete material to deteriorate and the slump to decrease. Therefore, in the actual production process, it needs to be handled properly.
  • Grading optimization combination of aggregates According to the requirements of aggregate grading, the proportion of aggregate meeting the slump can be calculated, and then the amount of aggregate can be calculated.
  • the instance segmentation model can be used to realize aggregate grading and online prediction of particle shape parameters.
  • Light sources 7 are set on both sides of the collection device to collect aggregate images in real time and transmit the aggregate images to the computer 5 or server.
  • Figure 6(a) is the aggregate image of the input instance segmentation model
  • Figure 6(b) is the output segmentation result after the instance segmentation model is segmented.
  • the outer contour of each particle can be obtained.
  • the fitEllipse function in OpenCV the particle contour fitting ellipse and the corresponding short diameter can be obtained. According to the size of the short diameter, it can be judged which grade of ingredients the particle belongs to, and all particles can be judged. The gradation of all aggregates in this batch can be obtained.
  • the second relationship there is a second relationship corresponding to the grading of each aggregate, and in the second relationship, the change value of the slump is in one-to-one correspondence with the change value of the water consumption and/or aggregate consumption.
  • the second relationship It is determined based on multiple experiments.
  • a model of the influence of gradation, aggregate content, and slump is established through multiple experiments.
  • the model is a corresponding relationship established according to the experimental data, which reflects the relationship between the gradation of aggregate, the increase in aggregate consumption (percentage) and the decrease in slump, or the relationship between the increase in water consumption and the increase in slump.
  • the increase in the amount of aggregate or the increase in water consumption is an increase relative to the original mixing amount.
  • step S5 it is judged whether it is necessary to adjust the water consumption and/or the aggregate consumption in the concrete mixing process according to the predicted value of the concrete working performance in the mixing image, specifically including:
  • step S5 the water consumption and/or aggregate consumption during the concrete mixing process are adjusted based on the predicted value of the working performance of the concrete, the gradation of the aggregate during the concrete mixing process, and the second relationship, specifically including:
  • the predicted value of the slump of the concrete is higher than the preset threshold range, then according to the relationship between the decrease in the slump and the increase in the amount of the aggregate in the second relationship corresponding to the gradation of the aggregate in the concrete mixing process Adjust the amount of aggregate increase during concrete mixing.
  • each aggregate gradation corresponds to a relationship between a decrease in slump and an increase in the amount of aggregate. If the concrete slump is required to be 150 ⁇ 20, and the predicted value of the concrete slump is 110, the water consumption needs to be increased accordingly. If the predicted value of the slump of the concrete is 190, then according to the gradation of the aggregate measured in step S4, a certain amount of aggregate should be supplemented. Modeling of the effect of slump and thus the amount of water and/or aggregate used during concrete mixing.
  • the value of increase or decrease will not fluctuate greatly, and the final adjusted slump can be within the allowable range of error.
  • the required slump is 150
  • the error is ⁇ 20.
  • the corresponding relationship of aggregate grading can be adjusted to make the slump closer to 150.
  • the invention can effectively predict the working performance of the concrete mixing process in real time and detect the aggregate gradation and particle shape parameters of sand and gravel, reduce the necessary performance detection time in the production process, and improve the production efficiency. If the working performance of this batch of concrete does not meet the requirements, calculate the amount of aggregate to be supplemented based on the aggregate grading and particle shape at the current moment, and make real-time adjustments to ensure that the working performance of the concrete meets the requirements and reduce waste of resources , improve the efficiency of formula adjustment.
  • the present application provides an embodiment of a concrete formula adjustment device based on deep learning, which corresponds to the method embodiment shown in Figure 2 , the device can be specifically applied to various electronic devices.
  • the embodiment of the present application provides a concrete formulation adjustment device based on deep learning, including:
  • the mixing image acquisition module 1 is configured to acquire the mixing image during the concrete mixing process, extract the concrete area image in the mixing image through the target detection model, and preprocess the concrete area image to obtain the processed concrete area image;
  • the gray-scale average calculation module 2 is configured to calculate the image gray-scale histogram based on the processed concrete area image, and calculate the gray-scale average value according to the image gray-scale histogram to obtain the gray-scale average value change curve;
  • the work performance prediction module 3 is configured to establish a first relationship between the gray-scale mean change curve and the work performance of the concrete, and determine the predicted value of the work performance of the concrete in the mixing image according to the gray-scale mean change curve and the first relationship;
  • the aggregate grading calculation module 4 is configured to obtain an aggregate image before concrete mixing, segment the aggregate image by using an instance segmentation model, obtain a segmentation result, and determine the aggregate gradation in the concrete mixing process based on the segmentation result;
  • the adjustment module 5 is configured to establish a second relationship between the change value of concrete work performance corresponding to the gradation of different aggregates and the change value of water consumption and/or the change value of aggregate consumption, according to the mixing image
  • the predicted value of the working performance of concrete determines whether it is necessary to adjust the amount of water and/or aggregate used in the concrete mixing process, and based on the predicted value of the working performance of the concrete, the gradation of the aggregate in the concrete mixing process and the second relationship Adjust the water consumption and/or aggregate consumption during the concrete mixing process, and repeatedly execute the mixing image acquisition module to the adjustment module so that the working performance of the concrete meets the requirements.
  • described agitation image acquisition module comprises the image acquisition device 4 that is arranged on mixer 1 feeding port, is used for the agitation image in real-time acquisition concrete mixing process;
  • Described aggregate grading calculation module comprises The image acquisition device 6 is arranged on the top of the mixer conveyor belt 9 for real-time collection of aggregate images; the adjustment module includes a control device (not shown in the figure) to control and adjust the water consumption and/or aggregate in the concrete mixing process dosage.
  • FIG. 8 shows a schematic structural diagram of a computer device 800 suitable for implementing an electronic device (such as the server or terminal device shown in FIG. 1 ) in the embodiment of the present application.
  • the electronic device shown in FIG. 8 is only an example, and should not limit the functions and application scope of the embodiment of the present application.
  • a computer device 800 includes a central processing unit (CPU) 801 and a graphics processing unit (GPU) 802, which can be randomly accessed according to a program stored in a read-only memory (ROM) 803 or loaded from a storage section 809. Various appropriate actions and processes are executed by programs in the memory (RAM) 804 . In the RAM 804, various programs and data necessary for the operation of the device 800 are also stored.
  • the CPU 801, GPU 802, ROM 803, and RAM 804 are connected to each other through a bus 805.
  • An input/output (I/O) interface 806 is also connected to the bus 805 .
  • the following components are connected to the I/O interface 806: an input section 807 including a keyboard, a mouse, etc.; an output section 808 including a liquid crystal display (LCD), etc., and a speaker; a storage section 809 including a hard disk, etc.;
  • the communication section 810 performs communication processing via a network such as the Internet.
  • Drive 811 may also be connected to I/O interface 806 as desired.
  • a removable medium 812, such as a magnetic disk, optical disk, magneto-optical disk, semiconductor memory, etc., is mounted on the drive 811 as necessary so that a computer program read therefrom is installed into the storage section 809 as necessary.
  • embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, where the computer program includes program codes for executing the methods shown in the flowcharts.
  • the computer program may be downloaded and installed from a network via communications portion 810 and/or installed from removable media 812 .
  • CPU central processing unit
  • GPU graphics processing unit
  • the computer-readable medium described in this application may be a computer-readable signal medium or a computer-readable medium or any combination of the above two.
  • a computer readable medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor device, device, or device, or a combination of any of the above. More specific examples of computer readable media may include, but are not limited to, electrical connections with one or more conductors, portable computer diskettes, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable Read Only Memory (EPROM or Flash), fiber optics, portable compact disk read only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above.
  • a computer-readable medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution device, device, or device.
  • a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, in which computer-readable program codes are carried. Such propagated data signals may take many forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the foregoing.
  • a computer-readable signal medium may also be any computer-readable medium other than a computer-readable medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution apparatus, apparatus, or device.
  • Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.
  • Computer program code for carrying out the operations of this application may be written in one or more programming languages, or combinations thereof, including object-oriented programming languages—such as Java, Smalltalk, C++, and conventional A procedural programming language—such as "C" or a similar programming language.
  • the program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server.
  • the remote computer can be connected to the user computer through any kind of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computer (such as through an Internet service provider). Internet connection).
  • LAN local area network
  • WAN wide area network
  • Internet service provider such as AT&T, MCI, Sprint, EarthLink, MSN, GTE, etc.
  • each block in a flowchart or block diagram may represent a module, program segment, or portion of code that contains one or more logical functions for implementing specified executable instructions.
  • the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or they may sometimes be executed in the reverse order, depending upon the functionality involved.
  • each block in the block diagrams and/or flowcharts, and combinations of blocks in the block diagrams and/or flowcharts can be implemented by a dedicated hardware-based device that performs the specified functions or operations , or may be implemented by a combination of dedicated hardware and computer instructions.
  • the modules involved in the embodiments described in the present application may be implemented by means of software or hardware.
  • the described modules may also be provided in a processor.
  • the present application also provides a computer-readable medium.
  • the computer-readable medium may be included in the electronic device described in the above embodiments; it may also exist independently without being assembled into the electronic device. middle.
  • the above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed by the electronic device, the electronic device: acquires a mixing image during the concrete mixing process, and uses a target detection model to detect The concrete area image is extracted, and the concrete area image is preprocessed to obtain the processed concrete area image; the image gray histogram is calculated based on the processed concrete area image, and the gray value is calculated according to the image gray histogram, Obtain the gray-scale mean change curve; establish the first relationship between the gray-scale mean change curve and the working performance of the concrete, and determine the predicted value of the concrete work performance in the mixing image according to the gray-scale mean change curve and the first relationship; obtain the concrete mixing Based on the previous aggregate image, the instance segmentation model is used to segment the aggregate image, and the
  • the aggregate gradation in the concrete mixing process is determined; the concrete working performance corresponding to the different aggregate gradation is established
  • the second relationship between the change value of the change value and the change value of the water consumption and/or the change value of the amount of aggregate judge whether it is necessary to adjust the water consumption and/or the change value of the aggregate in the mixing image according to the predicted value of the concrete working performance in the concrete mixing process or the amount of aggregate, and adjust the water consumption and/or the amount of aggregate in the concrete mixing process based on the predicted value of the working performance of the concrete, the gradation of the aggregate in the concrete mixing process, and the second relationship, and repeat the above steps to Make the working performance of concrete meet the requirements.
  • the present invention is a concrete formula adjustment method, device and readable medium based on deep learning.
  • a target detection model is established to judge whether the working performance of the concrete meets the requirements in real time, and at the same time obtain the aggregate image before mixing in real time. Adjusting the amount of water and/or the amount of aggregate based on the predicted value of work performance, the grading of the aggregate and the second relationship, repeating the above to make the work performance meet the requirements.
  • the invention has wide application range and good industrial applicability.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Software Systems (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Quality & Reliability (AREA)
  • Image Analysis (AREA)
  • Preparation Of Clay, And Manufacture Of Mixtures Containing Clay Or Cement (AREA)

Abstract

一种基于深度学习的混凝土配方调整方法、装置及可读介质，通过目标检测模型对实时采集的搅拌图像中的混凝土区域图像进行提取和预处理，得到处理后的混凝土区域图像，并计算出图像灰度直方图，得到灰度均值变化曲线；建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，并确定混凝土的工作性能的预测值；采用实例分割模型对骨料图像进行分割，得到分割结果，基于分割结果确定骨料的级配；建立混凝土的工作性能的变化值与用水量的变化值和/或骨料的用量的变化值之间的第二关系，基于工作性能的预测值、骨料的级配和第二关系调整用水量和/或骨料的用量，重复以上以使工作性能满足要求。该混凝土配方调整方法能对混凝土配方实时调整，提高效率。

Description

基于深度学习的混凝土配方调整方法、装置及可读介质 技术领域
本发明涉及工业化智能控制领域，具体涉及一种基于深度学习的混凝土配方调整方法、装置及可读介质。
背景技术
混凝土配合比设计通常在实验室条件下进行，施工现场使用时再微调。然而该配比是在一种砂、石情况下测定，往往在生产时会使用不同来源的砂、石原料。砂、石的粒形参数、含水率等对混凝土工作性能(包括坍落度、坍落扩展度等等)的影响较大。因此，需要通过测定砂、石参数来调整混凝土配比。
混凝土的工作性能在混凝土出机后由相关检测人员使用检测仪器进行相应指标的测量。通常，调整砂、石用料后必须进行工作性能检测。这带来三类问题：一是调整砂、石用量使混凝土工作性能满足要求可能经过循环往复的多轮调整，浪费人工与时间；二是混凝土出机后，一旦检测不合格，那么整盘混凝土只能废弃，造成资源浪费。三是同一批次的砂、石原料表里的含水率也会有一定不同，混凝土工作性能会有所差异。因此，目前迫切需要一种可以在混凝土生产过程中实时检测砂、石参数以及混凝土工作性能的方法。
发明内容
针对上述提到的技术问题。本申请的实施例的目的在于提出了一种基于深度学习的混凝土配方调整方法、装置及可读介质，来解决以上背景技术部分提到的技术问题。
第一方面，本申请的实施例提供了一种基于深度学习的混凝土配方调整方法，包括以下步骤：
S1，获取混凝土搅拌过程中的搅拌图像，通过目标检测模型对搅拌图像中的混凝土区域图像进行提取，并对混凝土区域图像进行预处理，得到处理后的混凝土区域图像；
S2，基于处理后的混凝土区域图像计算出图像灰度直方图，并根据图像灰度直方图计算灰度均值，得到灰度均值变化曲线；
S3，建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，根据灰度均值变化曲线和第一关系确定搅拌图像中混凝土的工作性能的预测值；
S4，获取混凝土搅拌前的骨料图像，采用实例分割模型对骨料图像进行分割，得到分割结果，基于分割结果确定混凝土搅拌过程中的骨料的级配；
S5，建立不同骨料的级配所对应的混凝土的工作性能的变化值与用水量的变化值和/或骨料的用量的变化值之间的第二关系，根据搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，并基于混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，重复步骤S1-S5以使混凝土的工作性能满足要求。
在一些实施例中，步骤S5中根据搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，具体包括：
判断搅拌图像中混凝土的工作性能的预测值是否超过预设阈值范围，若是，则需要调整混凝土搅拌过程中的用水量和/或骨料的用量；否则不需要调整混凝土搅拌过程中的用水量和/或骨料的用量。
在一些实施例中，混凝土的工作性能为混凝土的坍落度，第一关系中坍落度与灰度均值变化曲线中的灰度均值的波动区间一一对应，第二关系中坍落度的变化值与用水量和/或骨料的用量的变化值一一对应，步骤S5中基于混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，具体包括：
若混凝土的坍落度的预测值低于预设阈值范围，则根据第二关系中坍落度的增加值与用水量增加值的关系增加混凝土搅拌过程中的用水量；
若混凝土的坍落度的预测值高于预设阈值范围，则根据混凝土搅拌过程中的骨料的级配所对应的第二关系中坍落度的降低值与骨料用量的增加量的关系调整凝土搅拌过程中的骨料用量的增加量。
在一些实施例中，目标检测模型为经训练的第一Mask-Rcnn神经网络，实例分割模型为经训练的第二Mask-Rcnn神经网络，第一Mask-Rcnn神经网络和 第二Mask-Rcnn神经网络的骨干网络为Resnet50。
在一些实施例中，步骤S1中对混凝土区域图像进行预处理，具体包括：
对混凝土区域图像进行二值化处理，得到二值化后的混凝土区域图像；
将二值化后的混凝土区域图像中的背景部分进行滤除，得到处理后的混凝土区域图像。
在一些实施例中，步骤S2中采用OpenCV中calcHist函数计算出图像灰度直方图，图像灰度直方图记录的是处理后的混凝土区域图像中不同的灰度值所对应的像素数量，灰度均值为处理后的混凝土区域图像中所有像素点灰度值的总和与像素个数之间的比值。
在一些实施例中，步骤S4中基于分割结果确定混凝土搅拌过程中的骨料的级配，具体包括：
根据分割结果得到骨料图像中每个颗粒的轮廓；
基于每个颗粒的轮廓采用OpenCV中fitEllipse函数计算得出每个颗粒的轮廓的拟合椭圆以及相应短径；
根据短径大小分别判断每个颗粒所属的粒径范围；
对每个颗粒所属的粒径范围进行统计，获得骨料的级配。
第二方面，本申请的实施例提供了一种基于深度学习的混凝土配方调整装置，包括：
搅拌图像获取模块，被配置为获取混凝土搅拌过程中的搅拌图像，通过目标检测模型对搅拌图像中的混凝土区域图像进行提取，并对混凝土区域图像进行预处理，得到处理后的混凝土区域图像；
灰度均值计算模块，被配置为基于处理后的混凝土区域图像计算出图像灰度直方图，并根据图像灰度直方图计算灰度均值，得到灰度均值变化曲线；
工作性能预测模块，被配置为建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，根据灰度均值变化曲线和第一关系确定搅拌图像中混凝土的工作性能的预测值；
骨料级配计算模块，被配置为获取混凝土搅拌前的骨料图像，采用实例分割模型对骨料图像进行分割，得到分割结果，基于分割结果确定混凝土搅拌过程中的骨料的级配；
调整模块，被配置为建立不同骨料的级配所对应的混凝土的工作性能的变化值与用水量的变化值和/或骨料的用量的变化值之间的第二关系，根据搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，并基于混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，重复执行搅拌图像获取模块至调整模块以使混凝土的工作性能满足要求。
第三方面，本申请的实施例提供了一种电子设备，包括一个或多个处理器；存储装置，用于存储一个或多个程序，当一个或多个程序被一个或多个处理器执行，使得一个或多个处理器实现如第一方面中任一实现方式描述的方法。
第四方面，本申请的实施例提供了一种计算机可读存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现如第一方面中任一实现方式描述的方法。
相比于现有技术，本发明具有以下有益效果：
(1)本发明通过采集混凝土搅拌过程图像，建立目标检测模型，实时判断混凝土工作性能是否满足要求。同时，通过实时采集皮带上砂、石图像，建立实例分割模型，在线预测粒形参数等。并在混凝土工作性能不满足要求时能够根据该时刻骨料级配以及粒形参数调整用量，并验证混凝土工作性能。
(2)本发明通过实时检测混凝土工作性能以及砂、石骨料等粒形参数，智能调整骨料用量以满足混凝土工作性能要求，减少生产过程中必要的性能检测时间，提高生产效率。
(3)本发明能够在该批次混凝土工作性能不满足要求时，通过当前时刻骨料级配、粒形计算出所要补充的骨料用量，并实时进行调整，以保证混凝土出机时工作性能满足要求，减少资源浪费，提高调整配方效率。
附图说明
为了更清楚地说明本发明实施例中的技术方案，下面将对实施例描述中所需要使用的附图作简要介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域的普通技术人员来讲，在不付出创造性劳动的 前提下，还可以根据这些附图获得其他的附图。
图1是本申请的一个实施例可以应用于其中的示例性装置架构图；
图2为本发明的实施例的基于深度学习的混凝土配方调整方法的流程示意图；
图3为本发明的实施例的基于深度学习的混凝土配方调整方法的整体设备的示意图；
图4为本发明的实施例的基于深度学习的混凝土配方调整方法的混凝土的工作性能预测的流程示意图；
图5为本发明的实施例的基于深度学习的混凝土配方调整方法的混凝土的工作性能预测的结果图；
图6为本发明的实施例的基于深度学习的混凝土配方调整方法的实例分割模型的输入和分割结果图；
图7为本发明的实施例的基于深度学习的混凝土配方调整装置的示意图；
图8是适于用来实现本申请实施例的电子设备的计算机装置的结构示意图。
具体实施方式
为了使本发明的目的、技术方案和优点更加清楚，下面将结合附图对本发明作进一步地详细描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其它实施例，都属于本发明保护的范围。
图1示出了可以应用本申请实施例的基于深度学习的混凝土配方调整方法或基于深度学习的混凝土配方调整装置的示例性装置架构100。
如图1所示，装置架构100可以包括终端设备101、102、103，网络104和服务器105。网络104用以在终端设备101、102、103和服务器105之间提供通信链路的介质。网络104可以包括各种连接类型，例如有线、无线通信链路或者光纤电缆等等。
用户可以使用终端设备101、102、103通过网络104与服务器105交互， 以接收或发送消息等。终端设备101、102、103上可以安装有各种应用，例如数据处理类应用、文件处理类应用等。
终端设备101、102、103可以是硬件，也可以是软件。当终端设备101、102、103为硬件时，可以是各种电子设备，包括但不限于智能手机、平板电脑、膝上型便携计算机和台式计算机等等。当终端设备101、102、103为软件时，可以安装在上述所列举的电子设备中。其可以实现成多个软件或软件模块(例如用来提供分布式服务的软件或软件模块)，也可以实现成单个软件或软件模块。在此不做具体限定。
服务器105可以是提供各种服务的服务器，例如对终端设备101、102、103上传的文件或数据进行处理的后台数据处理服务器。后台数据处理服务器可以对获取的文件或数据进行处理，生成处理结果。
需要说明的是，本申请实施例所提供的基于深度学习的混凝土配方调整方法可以由服务器105执行，也可以由终端设备101、102、103执行，相应地，基于深度学习的混凝土配方调整装置可以设置于服务器105中，也可以设置于终端设备101、102、103中。
应该理解，图1中的终端设备、网络和服务器的数目仅仅是示意性的。根据实现需要，可以具有任意数目的终端设备、网络和服务器。在所处理的数据不需要从远程获取的情况下，上述装置架构可以不包括网络，而只需服务器或终端设备。
图2示出了本申请的实施例提供的一种基于深度学习的混凝土配方调整方法，包括以下步骤：
S1，获取混凝土搅拌过程中的搅拌图像，通过目标检测模型对搅拌图像中的混凝土区域图像进行提取，并对混凝土区域图像进行预处理，得到处理后的混凝土区域图像。
在具体的实施例中，参考图3，在搅拌机上方设置图像采集设备，在图像采集设备两侧设置光源，实时采集混凝土搅拌过程中的图像信息。在混凝土搅拌机1旁边设置支架2。在支架上固定光源3以及图像采集设备4，图像采集设备4可以通过搅拌机投料口拍摄混凝土搅拌过程的搅拌图像，并将该搅拌图像传输至电脑5或服务器端，因此可以实时采集获取到混凝土搅拌过程中的搅拌图像。利 用目标检测模型对搅拌图像中混凝土区域(ROI区域)进行提取，得到混凝土区域图像，在优选的实施例中，目标检测模型为经训练的第一Mask-Rcnn神经网络，第一Mask-Rcnn神经网络的骨干网络为Resnet50，Resnet50为现有的神经网络模型，因此其结构在此不再赘述。该目标检测模型基于深度学习神经网络模型构建，识别并提取出混凝土区域图像，后续再根据步骤S2和S3对混凝土的工作性能进行预测。
在具体的实施例中，步骤S1中对混凝土区域图像进行预处理，具体包括：
对混凝土区域图像进行二值化处理，得到二值化后的混凝土区域图像；
将二值化后的混凝土区域图像中的背景部分进行滤除，得到处理后的混凝土区域图像。
具体地，二值化后的混凝土区域图像中每个像素点的值为0或1，通过将背景部分的像素值设为0可以滤除其余背景干扰，方便后续进行灰度均值的计算。
S2，基于处理后的混凝土区域图像计算出图像灰度直方图，并根据图像灰度直方图计算灰度均值，得到灰度均值变化曲线。
在具体的实施例中，步骤S2中采用OpenCV中calcHist函数计算出图像灰度直方图。参考图4，图像灰度直方图记录的是处理后的混凝土区域图像中不同的灰度值所对应的像素数量，横坐标为灰度值，纵坐标为像素数量，从处理后的混凝土区域图像中采用OpenCV进行计算就可以提取出图像灰度直方图。灰度均值为处理后的混凝土区域图像中所有像素点灰度值的总和与像素个数之间的比值，即：
灰度均值＝处理后的混凝土区域图像中所有像素点灰度值的总和/像素个数。
根据灰度均值随着时间的变化建立灰度均值变化曲线，如图5所示，灰度均值变化曲线表示灰度均值在一定时间范围内的波动情况，并可以得到灰度均值的波动区间。
S3，建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，根据灰度均值变化曲线和第一关系确定搅拌图像中混凝土的工作性能的预测值。
在具体的实施例中，混凝土的工作性能为混凝土的坍落度，第一关系中坍落度与灰度均值变化曲线中的灰度均值的波动区间一一对应。
具体地，在确定第一关系时，分别调整混凝土的坍落度，并根据在此混凝土 的坍落度的数值范围下得到灰度均值变化曲线中的灰度均值的波动区间，例如：坍落度为180时所对应的灰度均值变化曲线在164～166波动；坍落度为150时所对应的灰度均值变化曲线在162～164波动；坍落度为120时所对应的灰度均值变化曲线在160～162波动。根据搅拌过程灰度均值大小区分不同工作性能(坍落度)的混凝土，由此可以判断该时刻混凝土工作性能是否满足要求。
S4，获取混凝土搅拌前的骨料图像，采用实例分割模型对骨料图像进行分割，得到分割结果，基于分割结果确定混凝土搅拌过程中的骨料的级配。
在具体的实施例中，实例分割模型为经训练的第二Mask-Rcnn神经网络，第二Mask-Rcnn神经网络的骨干网络为Resnet50。Resnet50为现有的神经网络模型，因此其结构在此不再赘述。步骤S4中基于分割结果确定混凝土搅拌过程中的骨料的级配，具体包括：
根据分割结果得到骨料图像中每个颗粒的轮廓；
基于每个颗粒的轮廓采用OpenCV中fitEllipse函数计算得出每个颗粒的轮廓的拟合椭圆以及相应短径；
根据短径大小分别判断每个颗粒所属的粒径范围；
对每个颗粒所属的粒径范围进行统计，获得骨料的级配。
具体地，骨料级配为组成骨料的不同粒径颗粒的比例关系；骨料级配主要分为连续级配和间断级配(单粒级)，连续级配主要指在最大粒径以下，依次序有其它相应粒级，不得间断，以期能充分填充骨料间的空隙。间断级配指在连续级配中缺少其中一级或几级中间粒级。混凝土骨料是混凝土的重要组成部分，在混凝土中起骨架和填充的作用。通常分为细骨料和粗骨料两类。在混凝土中一般将粒径在0.155～5mm之间的称为细骨料；将粒径大于5mm的称为粗骨料。粗骨料按种类分为卵石、碎石、破碎卵石、卵石和碎石的拌合物。骨料都偏细，有可能使砼料坍损变大、泌水增加，而偏粗的则会使砼料粘聚性变差、坍落度减小，所以实际生产过程中，需要处理好骨料的级配优化组合。根据骨料的级配的要求可以计算出符合坍落度的骨料的比例，进而计算出骨料的用量。在本申请的实施例中采用实例分割模型可实现骨料级配、粒形参数在线预测，参考图3，在传送带9上传输骨料8，在传送带9两侧设置图像采集设备6，在图像采集设备两侧设置光源7，实时采集骨料图像，并将该骨料图像传输至电脑5或服务器端中。 参考图6，图6(a)为输入实例分割模型的骨料图像，图6(b)为实例分割模型进行分割后输出的分割结果。根据分割结果可以得到每个颗粒的外轮廓，利用OpenCV中fitEllipse函数可以得出颗粒轮廓拟合椭圆以及相应短径，根据短径大小判断该颗粒属于哪一档级配料，对所有颗粒进行判断，即可得到该批次所有骨料的级配。
S5，建立不同骨料的级配所对应的混凝土的工作性能的变化值与用水量的变化值和/或骨料的用量的变化值之间的第二关系，根据搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，并基于混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，重复步骤S1-S5以使混凝土的工作性能满足要求。
具体地，每一个骨料的级配对应存在一个第二关系，第二关系中坍落度的变化值与用水量和/或骨料的用量的变化值一一对应，具体地，第二关系是根据多次实验确定的，在确定第二关系时，通过多次实验建立级配与骨料掺量以及坍落度影响的模型。该模型为根据实验数据建立的对应关系，体现出骨料的级配、骨料用量增加量(百分比)与坍落度减少值的关系或者用水量的增加量与坍落度增加值的关系，该骨料用量增加量或用水量的增加量是相对原先的搅拌量而言的增加量。
在具体的实施例中，步骤S5中根据搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，具体包括：
判断搅拌图像中混凝土的工作性能的预测值是否超过预设阈值范围，若是，则需要调整混凝土搅拌过程中的用水量和/或骨料的用量；否则不需要调整混凝土搅拌过程中的用水量和/或骨料的用量。
因此通过本申请的实施例不仅可以实时预测出混凝土的工作性能，还能够根据混凝土的工作性能、骨料的级配等关系对混凝土搅拌过程中的用水量和/或骨料的用量进行调整，实现混凝土配方的智能化调整。
步骤S5中基于混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，具体包括：
若混凝土的坍落度的预测值低于预设阈值范围，则根据第二关系中坍落度的 增加值与用水量增加值的关系增加凝土搅拌过程中的用水量；
若混凝土的坍落度的预测值高于预设阈值范围，则根据混凝土搅拌过程中的骨料的级配所对应的第二关系中坍落度的降低值与骨料用量的增加量的关系调整混凝土搅拌过程中的骨料用量的增加量。
具体地，在每一个骨料的级配分别对应一个坍落度的降低值与骨料用量的增加量的关系。如果要求混凝土坍落度为150±20，而混凝土的坍落度的预测值为110，则需要相应增加用水量。若混凝土的坍落度的预测值为190，则根据步骤S4所测出的骨料的级配，补充一定量的骨料，该量需要通过多次实验建立级配与骨料掺量以及坍落度影响的模型，从而得到混凝土搅拌过程中的用水量和/或骨料的用量。
在一个坍落度区间内，增加或者减少的值波动不会很大，最终调整后的坍落度只要在误差允许范围内就可以，例如要求坍落度150，误差是±20，只是说根据骨料级配这个对应关系可以调整使得坍落度更接近150。
本发明能够实时有效的预测混凝土搅拌过程工作性能以及检测砂石骨料级配、粒形参数，减少生产过程中必要的性能检测时间，提高生产效率。若该批次混凝土工作性能不满足要求时，通过当前时刻骨料级配、粒形计算出所要补充的骨料用量，并实时进行调整，以保证混凝土出机时工作性能满足要求，减少资源浪费，提高调整配方效率。
进一步参考图7，作为对上述各图所示方法的实现，本申请提供了一种基于深度学习的混凝土配方调整装置的一个实施例，该装置实施例与图2所示的方法实施例相对应，该装置具体可以应用于各种电子设备中。
本申请实施例提供了一种基于深度学习的混凝土配方调整装置，包括：
搅拌图像获取模块1，被配置为获取混凝土搅拌过程中的搅拌图像，通过目标检测模型对搅拌图像中的混凝土区域图像进行提取，并对混凝土区域图像进行预处理，得到处理后的混凝土区域图像；
灰度均值计算模块2，被配置为基于处理后的混凝土区域图像计算出图像灰度直方图，并根据图像灰度直方图计算灰度均值，得到灰度均值变化曲线；
工作性能预测模块3，被配置为建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，根据灰度均值变化曲线和第一关系确定搅拌图像中混凝土的工 作性能的预测值；
骨料级配计算模块4，被配置为获取混凝土搅拌前的骨料图像，采用实例分割模型对骨料图像进行分割，得到分割结果，基于分割结果确定混凝土搅拌过程中的骨料的级配；
调整模块5，被配置为建立不同骨料的级配所对应的混凝土的工作性能的变化值与用水量的变化值和/或骨料用量的变化值之间的第二关系，根据搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，并基于混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，重复执行搅拌图像获取模块至调整模块以使混凝土的工作性能满足要求。
参考图3，在具体实施例中，所述搅拌图像获取模块包括设置在搅拌机1投料口的图像采集设备4，用于实时采集混凝土搅拌过程中的搅拌图像；所述骨料级配计算模块包括设置在搅拌机传送带9上方图像采集设备6，用于实时采集骨料图像；所述调整模块包括控制装置(图中未示出)，用以控制调整混凝土搅拌过程中的用水量和/或骨料的用量。
下面参考图8，其示出了适于用来实现本申请实施例的电子设备(例如图1所示的服务器或终端设备)的计算机装置800的结构示意图。图8示出的电子设备仅仅是一个示例，不应对本申请实施例的功能和使用范围带来任何限制。
如图8所示，计算机装置800包括中央处理单元(CPU)801和图形处理器(GPU)802，其可以根据存储在只读存储器(ROM)803中的程序或者从存储部分809加载到随机访问存储器(RAM)804中的程序而执行各种适当的动作和处理。在RAM 804中，还存储有装置800操作所需的各种程序和数据。CPU 801、GPU802、ROM 803以及RAM 804通过总线805彼此相连。输入/输出(I/O)接口806也连接至总线805。
以下部件连接至I/O接口806：包括键盘、鼠标等的输入部分807；包括诸如、液晶显示器(LCD)等以及扬声器等的输出部分808；包括硬盘等的存储部分809；以及包括诸如LAN卡、调制解调器等的网络接口卡的通信部分810。通信部分810经由诸如因特网的网络执行通信处理。驱动器811也 可以根据需要连接至I/O接口806。可拆卸介质812，诸如磁盘、光盘、磁光盘、半导体存储器等等，根据需要安装在驱动器811上，以便于从其上读出的计算机程序根据需要被安装入存储部分809。
特别地，根据本公开的实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的实施例中，该计算机程序可以通过通信部分810从网络上被下载和安装，和/或从可拆卸介质812被安装。在该计算机程序被中央处理单元(CPU)801和图形处理器(GPU)802执行时，执行本申请的方法中限定的上述功能。
需要说明的是，本申请所述的计算机可读介质可以是计算机可读信号介质或者计算机可读介质或者是上述两者的任意组合。计算机可读介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的装置、装置或器件，或者任意以上的组合。计算机可读介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本申请中，计算机可读介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行装置、装置或者器件使用或者与其结合使用。而在本申请中，计算机可读的信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读的信号介质还可以是计算机可读介质以外的任何计算机可读介质，该计算机可读介质可以发送、传播或者传输用于由指令执行装置、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：无线、电线、光缆、RF等等，或者上述的任意合适的组合。
可以以一种或多种程序设计语言或其组合来编写用于执行本申请的操作的计算机程序代码，所述程序设计语言包括面向对象的程序设计语言—诸 如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图，图示了按照本申请各种实施例的装置、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的装置来实现，或者可以用专用硬件与计算机指令的组合来实现。
描述于本申请实施例中所涉及到的模块可以通过软件的方式实现，也可以通过硬件的方式来实现。所描述的模块也可以设置在处理器中。
作为另一方面，本申请还提供了一种计算机可读介质，该计算机可读介质可以是上述实施例中描述的电子设备中所包含的；也可以是单独存在，而未装配入该电子设备中。上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该电子设备执行时，使得该电子设备：获取混凝土搅拌过程中的搅拌图像，通过目标检测模型对搅拌图像中的混凝土区域图像进行提取，并对混凝土区域图像进行预处理，得到处理后的混凝土区域图像；基于处理后的混凝土区域图像计算出图像灰度直方图，并根据图像灰度直方图计算灰度均值，得到灰度均值变化曲线；建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，根据灰度均值变化曲线和第一关系确定搅拌图像中混凝土的工作性能的预测值；获取混凝土搅拌前的骨料图像，采用实例分割模型对骨料图像进行 分割，得到分割结果，基于分割结果确定混凝土搅拌过程中的骨料的级配；建立不同骨料的级配所对应的混凝土的工作性能的变化值与用水量的变化值和/或骨料的用量的变化值之间的第二关系，根据搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，并基于混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，重复上述步骤以使混凝土的工作性能满足要求。
以上描述仅为本申请的较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本申请中所涉及的发明范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述发明构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本申请中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。
工业实用性
本发明一种基于深度学习的混凝土配方调整方法、装置及可读介质，通过采集混凝土搅拌过程图像，建立目标检测模型，实时判断混凝土工作性能是否满足要求，同时实时获取搅拌前的骨料图像，基于工作性能的预测值、骨料的级配和第二关系调整用水量和/或骨料的用量，重复以上以使工作性能满足要求。本发明适用范围广，具有良好的工业实用性。

Claims (11)

  1. 一种基于深度学习的混凝土配方调整方法，其特征在于，包括以下步骤：
    S1，获取混凝土搅拌过程中的搅拌图像，通过目标检测模型对所述搅拌图像中的混凝土区域图像进行提取，并对所述混凝土区域图像进行预处理，得到处理后的混凝土区域图像；
    S2，基于所述处理后的混凝土区域图像计算出图像灰度直方图，并根据所述图像灰度直方图计算灰度均值，得到灰度均值变化曲线；
    S3，建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，根据所述灰度均值变化曲线和所述第一关系,确定所述搅拌图像中混凝土的工作性能的预测值；
    S4，获取混凝土搅拌前的骨料图像，采用实例分割模型对所述骨料图像进行分割，得到分割结果，基于所述分割结果确定混凝土搅拌过程中的骨料的级配；
    S5，建立不同骨料的级配所对应的混凝土的工作性能的变化值与用水量的变化值和/或骨料的用量的变化值之间的第二关系；根据所述搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，并基于所述混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和所述第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量；
    重复步骤S1-S5以使所述混凝土的工作性能满足要求。
  2. 根据权利要求1所述的基于深度学习的混凝土配方调整方法，其特征在于，所述步骤S5中根据所述搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，具体包括：
    判断所述搅拌图像中混凝土的工作性能的预测值是否超过预设阈值范围，若是，则调整混凝土搅拌过程中的用水量和/或骨料的用量。
  3. 根据权利要求2所述的基于深度学习的混凝土配方调整方法，其特征在于，所述混凝土的工作性能为混凝土的坍落度，所述第一关系中所述坍落度与灰度均值变化曲线中的灰度均值的波动区间一一对应，所述第二关系中所述坍落度的变化值与所述用水量和/或骨料的用量的变化值一一对应，所述步骤S5中基于所述混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和所述第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，具体包括：
    若混凝土的坍落度的预测值低于所述预设阈值范围，则根据所述第二关系中所述坍落度的增加值与用水量增加值的关系增加混凝土搅拌过程中的用水量；
    若混凝土的坍落度的预测值高于所述预设阈值范围，则根据所述混凝土搅拌过程中的骨料的级配所对应的第二关系中所述坍落度的降低值与骨料用量的增加量的关系调整凝土搅拌过程中的骨料用量的增加量。
  4. 根据权利要求1所述的基于深度学习的混凝土配方调整方法，其特征在于，所述目标检测模型为经训练的第一Mask-Rcnn神经网络，所述实例分割模型为经训练的第二Mask-Rcnn神经网络，所述第一Mask-Rcnn神经网络和所述第二Mask-Rcnn神经网络的骨干网络为Resnet50。
  5. 根据权利要求1所述的基于深度学习的混凝土配方调整方法，其特征在于，所述步骤S1中对所述混凝土区域图像进行预处理，具体包括：
    对所述混凝土区域图像进行二值化处理，得到二值化后的混凝土区域图像；
    将所述二值化后的混凝土区域图像中的背景部分进行滤除，得到所述处理后的混凝土区域图像。
  6. 根据权利要求1所述的基于深度学习的混凝土配方调整方法，其特征在于，所述步骤S2中采用OpenCV中calcHist函数计算出图像灰度直方图，所述图像灰度直方图记录的是所述处理后的混凝土区域图像中不同的灰度值所对应的像素数量，灰度均值为所述处理后的混凝土区域图像中所有像素点灰度值的总和与像素个数之间的比值。
  7. 根据权利要求1所述的基于深度学习的混凝土配方调整方法，其特征在于，所述步骤S4中基于所述分割结果确定混凝土搅拌过程中的骨料的级配，具体包括：
    根据所述分割结果得到所述骨料图像中每个颗粒的轮廓；
    基于所述每个颗粒的轮廓采用OpenCV中fitEllipse函数计算得出每个颗粒的轮廓的拟合椭圆以及相应短径；
    根据短径大小分别判断每个颗粒所属的粒径范围；
    对所述每个颗粒所属的粒径范围进行统计，获得所述骨料的级配。
  8. 一种基于深度学习的混凝土配方调整装置，其特征在于，包括：
    搅拌图像获取模块，被配置为获取混凝土搅拌过程中的搅拌图像，通过目标 检测模型对所述搅拌图像中的混凝土区域图像进行提取，并对所述混凝土区域图像进行预处理，得到处理后的混凝土区域图像；
    灰度均值计算模块，被配置为基于所述处理后的混凝土区域图像计算出图像灰度直方图，并根据所述图像灰度直方图计算灰度均值，得到灰度均值变化曲线；
    工作性能预测模块，被配置为建立灰度均值变化曲线与混凝土的工作性能之间的第一关系，根据所述灰度均值变化曲线和所述第一关系确定所述搅拌图像中混凝土的工作性能的预测值；
    骨料级配计算模块，被配置为获取混凝土搅拌前的骨料图像，采用实例分割模型对所述骨料图像进行分割，得到分割结果，基于所述分割结果确定混凝土搅拌过程中的骨料的级配；
    调整模块，被配置为建立不同骨料的级配所对应的混凝土的工作性能的变化值与用水量的变化值和/或骨料的用量的变化值之间的第二关系，根据所述搅拌图像中混凝土的工作性能的预测值判断是否需要调整混凝土搅拌过程中的用水量和/或骨料的用量，并基于所述混凝土的工作性能的预测值、混凝土搅拌过程中的骨料的级配和所述第二关系调整混凝土搅拌过程中的用水量和/或骨料的用量，重复执行搅拌图像获取模块至调整模块以使所述混凝土的工作性能满足要求。
  9. 根据权利要求8所述的基于深度学习的混凝土配方调整装置，其特征在于，所述搅拌图像获取模块包括设置在搅拌机投料口的图像采集设备，用于实时采集混凝土搅拌过程中的搅拌图像；
    所述骨料级配计算模块包括设置在搅拌机传送带上方图像采集设备，用于实时采集骨料图像；
    所述调整模块包括控制装置，用以调整混凝土搅拌过程中的用水量和/或骨料的用量。
  10. 一种电子设备，包括：
    一个或多个处理器；
    存储装置，用于存储一个或多个程序，
    当所述一个或多个程序被所述一个或多个处理器执行，使得所述一个或多个处理器实现如权利要求1-7中任一所述的方法。
  11. 一种计算机可读存储介质，其上存储有计算机程序，其特征在于，该程 序被处理器执行时实现如权利要求1-7中任一所述的方法。
PCT/CN2022/084286 2022-01-26 2022-03-31 基于深度学习的混凝土配方调整方法、装置及可读介质 WO2023142262A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210093471.6 2022-01-26
CN202210093471.6A CN114565561A (zh) 2022-01-26 2022-01-26 基于深度学习的混凝土配方调整方法、装置及可读介质

Publications (1)

Publication Number Publication Date
WO2023142262A1 true WO2023142262A1 (zh) 2023-08-03

Family

ID=81713327

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/084286 WO2023142262A1 (zh) 2022-01-26 2022-03-31 基于深度学习的混凝土配方调整方法、装置及可读介质

Country Status (2)

Country Link
CN (1) CN114565561A (zh)
WO (1) WO2023142262A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117195083A (zh) * 2023-11-08 2023-12-08 福建南方路面机械股份有限公司 基于电流曲线的坍落度预测方法、装置及可读介质
CN117892603A (zh) * 2024-03-15 2024-04-16 江西省水利科学院(江西省大坝安全管理中心、江西省水资源管理中心) 一种用于渡槽混凝土底板的数值模拟方法

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114971092B (zh) * 2022-07-29 2022-10-18 华北电力大学 一种变电站主控楼混凝土用量预测方法和装置
CN115294127A (zh) * 2022-10-08 2022-11-04 南通钰昇智能科技有限公司 一种混凝土制品生产异常预警系统
CN115592819B (zh) * 2022-12-15 2023-03-17 福建南方路面机械股份有限公司 基于深度学习的搅拌站卸料口尺寸调整方法及装置
CN116973553B (zh) * 2023-06-08 2024-04-12 山东博硕自动化技术有限公司 基于深度学习的在线物料识别与含水率测量方法及装置
CN117218118B (zh) * 2023-11-07 2024-03-12 福建南方路面机械股份有限公司 基于图像序列的坍落度监测方法、装置及可读介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111751253A (zh) * 2020-07-06 2020-10-09 重庆理工大学 一种混凝土骨料检测模型的形成方法及质量检测方法
CN112873533A (zh) * 2021-01-15 2021-06-01 清华大学 混凝土自适应智能拌合方法及装置
CN112927184A (zh) * 2021-01-15 2021-06-08 重庆交通大学 基于深度学习的自密实混凝土性能检测方法及装置
WO2021116279A1 (de) * 2019-12-11 2021-06-17 Peri Gmbh Computergestütztes verfahren sowie einrichtung zur steuerung einer beton-mischanlage
CN113942121A (zh) * 2021-09-03 2022-01-18 湖南中联重科混凝土机械站类设备有限公司 用于搅拌站骨料含水率的控制方法、处理器及装置

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2021116279A1 (de) * 2019-12-11 2021-06-17 Peri Gmbh Computergestütztes verfahren sowie einrichtung zur steuerung einer beton-mischanlage
CN111751253A (zh) * 2020-07-06 2020-10-09 重庆理工大学 一种混凝土骨料检测模型的形成方法及质量检测方法
CN112873533A (zh) * 2021-01-15 2021-06-01 清华大学 混凝土自适应智能拌合方法及装置
CN112927184A (zh) * 2021-01-15 2021-06-08 重庆交通大学 基于深度学习的自密实混凝土性能检测方法及装置
CN113942121A (zh) * 2021-09-03 2022-01-18 湖南中联重科混凝土机械站类设备有限公司 用于搅拌站骨料含水率的控制方法、处理器及装置

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117195083A (zh) * 2023-11-08 2023-12-08 福建南方路面机械股份有限公司 基于电流曲线的坍落度预测方法、装置及可读介质
CN117195083B (zh) * 2023-11-08 2024-03-12 福建南方路面机械股份有限公司 基于电流曲线的坍落度预测方法、装置及可读介质
CN117892603A (zh) * 2024-03-15 2024-04-16 江西省水利科学院(江西省大坝安全管理中心、江西省水资源管理中心) 一种用于渡槽混凝土底板的数值模拟方法
CN117892603B (zh) * 2024-03-15 2024-05-17 江西省水利科学院(江西省大坝安全管理中心、江西省水资源管理中心) 一种用于渡槽混凝土底板的数值模拟方法

Also Published As

Publication number Publication date
CN114565561A (zh) 2022-05-31

Similar Documents

Publication Publication Date Title
WO2023142262A1 (zh) 基于深度学习的混凝土配方调整方法、装置及可读介质
WO2023168899A1 (zh) 圆锥破碎机排料口尺寸智能调整方法、装置及可读介质
CN114266989A (zh) 一种混凝土拌合物和易性测定方法和装置
CN110118775A (zh) 厂拌水泥稳定碎石骨料组成快速检测方法
Quan et al. The method of the road surface crack detection by the improved Otsu threshold
CN112419397B (zh) 一种基于图像和深度神经网络的矿石粒度分级方法及系统
WO2023240776A1 (zh) 基于深度学习的渣土分类处理引导方法、装置及可读介质
Wei et al. Automated bughole detection and quality performance assessment of concrete using image processing and deep convolutional neural networks
CN113743424B (zh) 全尾砂动态絮凝沉降中絮体动态沉降分散程度定量表征方法
CN117218118B (zh) 基于图像序列的坍落度监测方法、装置及可读介质
CN114496083A (zh) 细胞类型确定方法、装置、设备以及存储介质
CN112329782A (zh) 一种原料粒度确定方法、系统、终端及介质
CN113592792A (zh) 堆石图像粒径识别方法、装置、电子设备及存储介质
CN115890915A (zh) 一种混凝土搅拌过程图像识别方法
CN113822110B (zh) 一种目标检测的方法和装置
WO2024066664A1 (zh) 一种混凝土可泵性类别识别方法、装置及电子设备
CN115731519A (zh) 一种非接触式全尾砂膏体搅拌均匀度监测方法及系统
CN115375685B (zh) 用于混凝土原料中的砂石粒径异常检测方法
CN114758184B (zh) 基于深度学习的渣土分类处理引导方法、装置及可读介质
CN114565838A (zh) 一种基于渣土图像的配方调节控制方法、装置及可读介质
CN116689133B (zh) 基于深度学习的再生骨料质量控制方法、装置及可读介质
CN112904813B (zh) 基于5g和物联网的隧洞衬砌智能化控制系统及方法
CN114196799A (zh) 一种转炉溢渣的检测方法、装置及计算机设备
CN104299002B (zh) 一种基于监控系统的塔吊图像检测方法
CN102507665B (zh) 一种浊液搅拌效果的判断装置及方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 22923080

Country of ref document: EP

Kind code of ref document: A1