CN117115731A - Dust shielding system and method for building construction - Google Patents

Dust shielding system and method for building construction Download PDF

Info

Publication number
CN117115731A
CN117115731A CN202311033366.4A CN202311033366A CN117115731A CN 117115731 A CN117115731 A CN 117115731A CN 202311033366 A CN202311033366 A CN 202311033366A CN 117115731 A CN117115731 A CN 117115731A
Authority
CN
China
Prior art keywords
feature
decoding
monitoring
image
matrixes
Prior art date
Legal status (The legal status 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 status listed.)
Pending
Application number
CN202311033366.4A
Other languages
Chinese (zh)
Inventor
施春雷
史祥生
刘琳
姜博瀚
唐超
郑新树
李春华
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China Overseas Construction Ltd
Original Assignee
China Overseas Construction Ltd
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 China Overseas Construction Ltd filed Critical China Overseas Construction Ltd
Priority to CN202311033366.4A priority Critical patent/CN117115731A/en
Publication of CN117115731A publication Critical patent/CN117115731A/en
Pending legal-status Critical Current

Links

Classifications

    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B01—PHYSICAL OR CHEMICAL PROCESSES OR APPARATUS IN GENERAL
    • B01D—SEPARATION
    • B01D47/00—Separating dispersed particles from gases, air or vapours by liquid as separating agent
    • B01D47/06—Spray cleaning
    • 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
    • G06N3/0455—Auto-encoder networks; Encoder-decoder 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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • 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
    • G06N3/048—Activation functions
    • 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
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/766—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using regression, e.g. by projecting features on hyperplanes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects

Abstract

A dust shielding system for building construction and a method thereof are disclosed. Firstly, carrying out image graying on monitoring images around a construction site to obtain a graying monitoring image, then carrying out image block division on the graying monitoring image, obtaining a plurality of monitoring local feature matrixes through a convolutional neural network model, then calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes to obtain a decoding feature map composed of a plurality of covariance matrixes, and finally carrying out decoding regression on the decoding feature map after feature distribution optimization so as to obtain a decoding value for representing a recommended valve opening value. In this way, dust removal efficiency and effect can be optimized.

Description

Dust shielding system and method for building construction
Technical Field
The present application relates to the field of dust shielding, and more particularly, to a dust shielding system for construction and a method thereof.
Background
In the building construction process, a great amount of dust, pollutants and other harmful gases are generated in a series of processes of construction material transportation, processing and the like, and if the dust is not effectively controlled, the physical health of workers and surrounding residents can be influenced, and even traffic accidents and other safety problems are caused. Therefore, the construction site needs to take corresponding emission reduction measures to shield dust, so that the environment is ensured to be good.
Many devices for controlling dust in building construction, such as spray systems, fan systems, etc., are currently on the market. In the spraying system, the spraying device is arranged at the periphery of a construction site, a main building, a tower crane and other parts, and water mist is sprayed into the air, so that the water mist is combined with dust particles and then is settled, and the purpose of reducing the dust concentration is achieved. However, the existing spray control scheme often controls the opening of a valve to be sprayed within a preset range, and does not pay attention to the suitability relationship between the valve and the surrounding environment, so that the dust settling effect cannot reach the original expectation or the waste of water resources is caused, and the environmental protection requirement is not met.
Accordingly, an optimized dust shielding system for construction is desired.
Disclosure of Invention
The present application has been made to solve the above-mentioned technical problems. The embodiment of the application provides a dust shielding system and a dust shielding method for building construction. Firstly, carrying out image graying on monitoring images around a construction site to obtain a graying monitoring image, then carrying out image block division on the graying monitoring image, obtaining a plurality of monitoring local feature matrixes through a convolutional neural network model, then calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes to obtain a decoding feature map composed of a plurality of covariance matrixes, and finally carrying out decoding regression on the decoding feature map after feature distribution optimization so as to obtain a decoding value for representing a recommended valve opening value. In this way, dust removal efficiency and effect can be optimized.
According to an aspect of the present application, there is provided a dust shielding system for construction, comprising:
the monitoring image acquisition module is used for acquiring monitoring images around the construction site acquired by the camera;
the gray processing module is used for carrying out image graying on the monitoring images around the building construction site to obtain grayed monitoring images;
the image dividing module is used for dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks;
the image feature extraction module is used for obtaining a plurality of monitoring local feature matrixes through a convolutional neural network model using a cavity convolutional kernel for each image block in the sequence of the monitoring image blocks;
the image feature association module is used for calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes to obtain a decoding feature diagram composed of a plurality of covariance matrixes;
the feature optimization module is used for carrying out feature distribution optimization on the decoding feature map so as to obtain an optimized decoding feature map; and
and the valve opening control module is used for carrying out decoding regression on the optimized decoding characteristic diagram through a decoder to obtain a decoding value, wherein the decoding value is used for representing the recommended valve opening value.
In the above dust shielding system for construction, the image dividing module is configured to:
and uniformly dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks.
In the above dust shielding system for construction, the image feature extraction module is configured to:
and respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on input data in forward transmission of layers by using each layer of the convolutional neural network model using the hole convolution kernel so as to output the plurality of monitoring local feature matrices by the last layer of the convolutional neural network model using the hole convolution kernel, wherein the input of the first layer of the convolutional neural network model using the hole convolution kernel is each image block in the sequence of the monitoring image blocks.
In the dust shielding system for building construction, the image feature correlation module is configured to:
calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes according to the following covariance formula to obtain the plurality of covariance matrixes;
wherein, the covariance formula is:
Wherein M is 1 And M 2 Representing each two monitored local feature matrices of the plurality of monitored local feature matrices, M representing a covariance matrix between each two monitored local feature matrices of the plurality of monitored local feature matrices,representing matrix multiplication; and
and arranging the covariance matrixes into the decoding characteristic diagram.
In the above dust shielding system for construction, the feature optimization module includes:
an optimization factor calculation unit, configured to calculate a piece-wise approximation factor of the convex-decomposition-based feature geometry metric for each of the decoding feature matrices along the channel dimension to obtain a plurality of piece-wise approximation factors of the convex-decomposition-based feature geometry metric;
the weighted optimization unit is used for carrying out weighted optimization on a plurality of decoding feature matrixes of the decoding feature graph along the channel dimension by taking the piece-by-piece approximation factors of the feature geometric metrics based on the convex decomposition as weighted coefficients so as to obtain a plurality of optimized decoding feature matrixes; and
and the dimension reconstruction unit is used for carrying out dimension reconstruction on the plurality of optimized decoding feature matrixes to obtain the optimized decoding feature map.
In the above dust shielding system for construction, the optimization factor calculating unit is configured to:
Calculating the piecewise approximation factors of the convex decomposition-based feature geometry metrics of each decoded feature matrix of the decoded feature map along the channel dimension with the following optimization formula to obtain the piecewise approximation factors of the plurality of convex decomposition-based feature geometry metrics;
wherein, the optimization formula is:
wherein V is ij Is the j-th row vector or column vector of the i-th decoding feature matrix, sigmoid (·) represents Sigmoid function, logSumExp (·) represents LogSumExp function, [:]representing concatenating individual vectors, anRepresenting the square of the two norms of a vector，w i Representing an ith one of the plurality of piecewise approximation factors of the convex-based feature geometry metric.
In the dust shielding system for building construction, the valve opening control module is used for:
performing decoding regression on the optimized decoding feature map using a plurality of fully connected layers of the decoder in a decoding regression formula to obtain the decoded values;
wherein, the decoding regression formula is:
wherein X is the optimized decoding feature map, Y is the decoding value, W is a weight matrix,representing a matrix multiplication.
According to another aspect of the present application, there is provided a dust shielding method for construction, comprising:
Acquiring monitoring images around a building construction site acquired by a camera;
performing image graying on the monitoring image around the building construction site to obtain a grayed monitoring image;
dividing the graying monitoring image into image blocks to obtain a sequence of monitoring image blocks;
each image block in the sequence of the monitoring image blocks is subjected to a convolutional neural network model by using a cavity convolutional kernel to obtain a plurality of monitoring local feature matrixes;
calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes to obtain a decoding feature diagram composed of a plurality of covariance matrixes;
performing feature distribution optimization on the decoding feature map to obtain an optimized decoding feature map; and
and carrying out decoding regression on the optimized decoding characteristic map through a decoder to obtain a decoding value, wherein the decoding value is used for representing the recommended valve opening value.
In the above dust shielding method for construction, the image block division of the graying monitoring image is performed to obtain a sequence of monitoring image blocks, including:
and uniformly dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks.
In the above dust shielding method for construction, the step of obtaining a plurality of monitoring local feature matrices by using a convolutional neural network model of a hole convolutional kernel for each image block in the sequence of monitoring image blocks includes:
and respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on input data in forward transmission of layers by using each layer of the convolutional neural network model using the hole convolution kernel so as to output the plurality of monitoring local feature matrices by the last layer of the convolutional neural network model using the hole convolution kernel, wherein the input of the first layer of the convolutional neural network model using the hole convolution kernel is each image block in the sequence of the monitoring image blocks.
Compared with the prior art, the dust shielding system for building construction and the method thereof provided by the application have the advantages that firstly, the monitoring image around the building construction site is subjected to image graying to obtain a graying monitoring image, then, the graying monitoring image is subjected to image block division and then is subjected to convolutional neural network model to obtain a plurality of monitoring local feature matrixes, then, covariance matrixes between every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes are calculated to obtain a decoding feature map composed of a plurality of covariance matrixes, and finally, the decoding feature map is subjected to feature distribution optimization and then is subjected to decoding regression through a decoder to obtain a decoding value for representing a recommended valve opening value. In this way, dust removal efficiency and effect can be optimized.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for the description of the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort to a person of ordinary skill in the art. The following drawings are not intended to be drawn to scale, emphasis instead being placed upon illustrating the principles of the application.
Fig. 1 is an application scenario diagram of a dust shielding system for construction according to an embodiment of the present application.
Fig. 2 is a block diagram schematically illustrating a dust shielding system for construction according to an embodiment of the present application.
Fig. 3 is a block diagram schematic of the feature optimization module in the dust shielding system for construction according to the embodiment of the application.
Fig. 4 is a flowchart of a dust shielding method for construction according to an embodiment of the present application.
Fig. 5 is a schematic diagram of a system architecture of a dust shielding method for construction according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some, but not all embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are also within the scope of the application.
As used in the specification and in the claims, the terms "a," "an," "the," and/or "the" are not specific to a singular, but may include a plurality, unless the context clearly dictates otherwise. In general, the terms "comprises" and "comprising" merely indicate that the steps and elements are explicitly identified, and they do not constitute an exclusive list, as other steps or elements may be included in a method or apparatus.
Although the present application makes various references to certain modules in a system according to embodiments of the present application, any number of different modules may be used and run on a user terminal and/or server. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.
A flowchart is used in the present application to describe the operations performed by a system according to embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed in order precisely. Rather, the various steps may be processed in reverse order or simultaneously, as desired. Also, other operations may be added to or removed from these processes.
Hereinafter, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. It should be apparent that the described embodiments are only some embodiments of the present application and not all embodiments of the present application, and it should be understood that the present application is not limited by the example embodiments described herein.
As described above, in the spraying system, the spraying device is arranged at the periphery of the construction site, the main building, the tower crane and other parts, so that the water mist is sprayed into the air, and the water mist is combined with dust particles and then settled, thereby achieving the purpose of reducing the dust concentration. However, the existing spray control scheme often controls the opening of a valve to be sprayed within a preset range, and does not pay attention to the suitability relationship between the valve and the surrounding environment, so that the dust settling effect cannot reach the original expectation or the waste of water resources is caused, and the environmental protection requirement is not met. Accordingly, an optimized dust shielding system for construction is desired.
Accordingly, in order to avoid wasting resources on the basis of ensuring dust removal efficiency and effect in the process of using the spraying system to shield dust during construction of a building, self-adaptive control of the opening value of the spraying valve is required based on actual environmental conditions around the construction site so as to meet the environmental protection requirement. Accordingly, in the present application, it is desirable to perform real-time control of the valve opening value by capturing dust state characteristics of the surrounding environment based on image analysis of the monitoring image around the construction site. However, since a large amount of information exists in the monitoring image around the construction site, and dust is hidden characteristic information of a small scale in the image, it is difficult to sufficiently capture and describe the dust characteristic in the image, resulting in lower control accuracy for the valve opening. Therefore, in the process, the difficulty is how to fully express the dust state hidden characteristic distribution information about the periphery of the construction site in the monitoring image, so that the self-adaptive control of the opening value of the spraying valve is performed based on the dust environment condition about the periphery of the construction site, thereby optimizing the dust removal efficiency and effect and avoiding the waste of water resources.
In recent years, deep learning and neural networks have been widely used in the fields of computer vision, natural language processing, text signal processing, and the like. The development of deep learning and neural networks provides new solutions and schemes for mining dust state implicit feature distribution information about construction sites in the monitored images.
Specifically, in the technical scheme of the application, firstly, the monitoring image around the construction site is collected through the camera. Then, the problems that noise interference and poor image quality exist in the process of collecting the monitoring image due to the complex and diverse environments around the building construction site are considered. Therefore, in the technical scheme of the application, the monitoring image around the construction site is further subjected to image graying so as to obtain a graying monitoring image. It should be appreciated that image graying is one of the commonly used image processing techniques that is capable of converting color images into gray-scale images to reduce the amount of image data, simplify computation, and increase computational efficiency. That is, the monitoring image can be processed and converted into a monochromatic gray image by image graying, the subsequent processing steps are simplified, and the problems of accuracy, noise and the like in the monitoring image are considered, the characteristics in the image can be better highlighted by using the graying technology, and the problems of image misjudgment and the like caused by color difference are reduced. In addition, the graying operation can also help to improve the utilization rate of the machine learning model to data, reduce the calculation time and the resource consumption required by model training, provide better input for the subsequent steps, and improve the performance of the model.
Then, considering that the hidden characteristic about the dust distribution state in the grayscale monitoring image is fine characteristic information of a small scale, in order to improve the expression capability about the dust distribution state characteristic in the grayscale monitoring image, the accuracy of valve opening value control is improved. It should be appreciated that the dimensions of each monitoring image block in the sequence of monitoring image blocks are reduced compared to the original image, and therefore, the small-scale implicit features in the grayscale monitoring image with respect to the dust distribution state are no longer small-scale objects in each monitoring image block, so as to facilitate the subsequent feature extraction of the dust distribution state and the control of the valve opening.
Further, feature mining of each image block in the sequence of monitoring image blocks is performed using a convolutional neural network model having excellent performance in terms of implicit feature extraction, and in particular, in order to be able to focus on associated feature information between dust distribution states of each local region in a spatial position in a construction site, considering that dust distribution conditions in each local spatial region of the construction site are different when dust detection is actually performed, it is necessary to pass each image block through the convolutional neural network model using a hole convolution kernel to obtain a plurality of monitoring local feature matrices. That is, the feature mining of each image block is performed by using the convolutional neural network model of the hole convolutional kernel so as to extract the local implicit associated feature information about the dust distribution state in each image block, and meanwhile, the convolutional neural network model of the hole convolutional kernel can reduce the background interference, thereby being beneficial to the expression of the feature about the dust distribution state in the construction site.
Next, it is considered that, at the time of actually performing dust state distribution detection at the construction site, the distribution state characteristics concerning dust in the respective image blocks have an association relationship, and such association relationship exhibits different association characteristics in different dust distribution states. Therefore, if it is desired to further improve the expression of the dust distribution state feature, in the technical solution of the present application, the covariance matrix between every two monitoring local feature matrices in the plurality of monitoring local feature matrices is further calculated to obtain a decoding feature map composed of a plurality of covariance matrices, so as to represent the correlation feature distribution information about the dust distribution feature in every two image blocks.
And then, carrying out decoding regression processing on the decoding characteristic map through a decoder to obtain a decoding value for representing the recommended valve opening value. That is, the method decodes the relevant characteristic information about the dust distribution characteristic in each image block, so as to accurately analyze the environment dust state condition around the construction site, and thus, the self-adaptive control of the opening value of the spraying valve is performed, thereby optimizing the dust removal efficiency and effect and avoiding the waste of water resources.
In particular, in the technical solution of the present application, the decoding feature map is obtained by arranging the covariance matrices along a channel dimension, and each covariance matrix in the covariance matrices expresses covariance features of block image semantics of image blocks of the grayscale monitoring image expressed by the monitoring local feature matrix, so that, due to source image semantic differences among the image blocks of the grayscale monitoring image, distribution among the image semantic features expressed by the monitoring local feature matrix may exhibit larger non-uniformity, thereby resulting in non-uniformity of manifold geometric representations with high-dimensional feature manifolds among the covariance matrices obtained by calculating covariance matrices between every two monitoring local feature matrices. Therefore, when the covariance matrixes are arranged along the channel to form the decoding feature map, the convergence difficulty of the decoding feature map in decoding regression through the decoder is improved, and therefore the training speed and the accuracy of the converged decoding value are reduced.
The applicant of the present application therefore calculates a piece-wise approximation factor of the convex decomposition-based feature geometry metric for each covariance matrix of the decoded feature map, expressed as:
Wherein V is ij Is each covariance matrix M i Is the j-th row vector or column vector of [:]representing concatenating individual vectors, anRepresenting the square of the two norms of the vector.
In particular, the piece-wise approximation factor of the convex decomposition-based feature geometry metric may define a symbolized distance metric between local geometries of high-dimensional feature manifolds of each covariance matrix by a smooth maximum function of LogSumExp, obtain a microscopic convex indicator (convex indicator) of each convex polyhedron object based on convex polyhedron (convex polytope) decomposition of the high-dimensional feature manifolds, and further determine a learnable piece-wise (piece-wise) convex decomposition hyperplane distance parameter for expressing the high-dimensional feature manifolds by a Sigmoid function to approximately measure feature geometries. In this way, by weighting the covariance matrix by the slice-by-slice approximation factor based on the feature geometric measure of the convex decomposition, the consistency of manifold geometric representation of the high-dimensional feature manifold among different covariance matrices under different channels of the decoding feature map obtained by arranging the covariance matrix according to the channels can be improved, so that the convergence effect of the decoding feature map when decoding regression is carried out by a decoder is improved, and the training speed of the model and the accuracy of decoding values are improved. Therefore, the self-adaptive control of the opening value of the spraying valve can be performed based on the dust environment conditions around the building construction site, so that the dust removal efficiency and effect are optimized, and the waste of water resources is avoided.
Fig. 1 is an application scenario diagram of a dust shielding system for construction according to an embodiment of the present application. As shown in fig. 1, in this application scenario, first, a monitoring image around a construction site acquired by a camera (e.g., D illustrated in fig. 1) is acquired, and then, the monitoring image around the construction site is input to a server where a dust shielding algorithm for construction is deployed (e.g., S illustrated in fig. 1), wherein the server can process the monitoring image around the construction site using the dust shielding algorithm for construction to obtain a decoded value for representing a recommended valve opening value.
Having described the basic principles of the present application, various non-limiting embodiments of the present application will now be described in detail with reference to the accompanying drawings.
Fig. 2 is a block diagram schematically illustrating a dust shielding system for construction according to an embodiment of the present application. As shown in fig. 2, the dust shielding system 100 for construction according to an embodiment of the present application includes: a monitoring image acquisition module 110 for acquiring monitoring images around the construction site acquired by the camera; the gray processing module 120 is configured to perform image graying on the monitoring image around the construction site to obtain a grayed monitoring image; an image dividing module 130, configured to perform image block division on the graying monitoring image to obtain a sequence of monitoring image blocks; an image feature extraction module 140, configured to obtain a plurality of monitoring local feature matrices by using a convolutional neural network model of a hole convolutional kernel for each image block in the sequence of monitoring image blocks; the image feature association module 150 is configured to calculate a covariance matrix between every two monitored local feature matrices in the plurality of monitored local feature matrices to obtain a decoded feature map composed of a plurality of covariance matrices; the feature optimization module 160 is configured to perform feature distribution optimization on the decoding feature map to obtain an optimized decoding feature map; and a valve opening control module 170, configured to perform decoding regression on the optimized decoding feature map through a decoder to obtain a decoded value, where the decoded value is used to represent a recommended valve opening value.
More specifically, in the embodiment of the present application, the monitoring image acquisition module 110 is configured to acquire a monitoring image around a construction site acquired by a camera. In the process of using a spraying system to shield dust during construction of a building, in order to avoid wasting resources on the basis of ensuring dust removal efficiency and effect, self-adaptive control of the opening value of a spraying valve is required based on actual environmental conditions around a construction site so as to meet the environmental protection requirement. Therefore, the real-time control of the valve opening value can be performed by capturing the dust state characteristics of the surrounding environment based on the image analysis of the monitoring image around the construction site.
More specifically, in the embodiment of the present application, the gray-scale processing module 120 is configured to perform image graying on the monitoring image around the construction site to obtain a grayed monitoring image. The environments around the construction sites are complex and various, so that the problems of noise interference and poor image quality of the monitoring images in the process of acquisition can exist. Therefore, in the technical scheme of the application, the monitoring image around the construction site is further subjected to image graying so as to obtain a graying monitoring image. It should be understood that the monitoring image can be processed and converted into a monochromatic gray image by image graying, so that the subsequent processing steps are simplified, and the problems of precision, noise and the like in the monitoring image are considered, and the characteristics in the image can be better highlighted and the problems of image misjudgment and the like caused by color difference can be reduced by using the graying technology.
More specifically, in the embodiment of the present application, the image dividing module 130 is configured to perform image block division on the graying monitoring image to obtain a sequence of monitoring image blocks. In order to improve the expression capability of the dust distribution state characteristics in the grayscale monitoring image, the accuracy of valve opening value control is improved, and in the technical scheme of the application, the grayscale monitoring image is subjected to image block division to obtain a sequence of monitoring image blocks. It should be appreciated that the dimensions of each monitoring image block in the sequence of monitoring image blocks are reduced compared to the original image, and therefore, the small-scale implicit features in the grayscale monitoring image with respect to the dust distribution state are no longer small-scale objects in each monitoring image block, so as to facilitate the subsequent feature extraction of the dust distribution state and the control of the valve opening.
Accordingly, in one specific example, the image dividing module 130 is configured to: and uniformly dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks.
More specifically, in the embodiment of the present application, the image feature extraction module 140 is configured to obtain a plurality of monitoring local feature matrices by using a convolutional neural network model of a hole convolutional kernel for each image block in the sequence of monitoring image blocks. Feature mining of each image block in the sequence of monitoring image blocks is performed using a convolutional neural network model having excellent performance in terms of implicit feature extraction, particularly, in order to be able to pay attention to associated feature information between dust distribution states of each local region in a spatial position in a construction site due to difference in dust distribution situation in each local spatial region of the construction site when dust detection is actually performed, it is necessary to perform feature mining of each image block using a convolutional neural network model of a hole convolutional kernel to extract locally implicit associated feature information about dust distribution states in each image block, respectively, while the convolutional neural network model using a hole convolutional kernel can also reduce background interference, which is advantageous for expression of features about dust distribution states in the construction site.
It should be appreciated that convolutional neural network (Convolutional Neural Network, CNN) is an artificial neural network and has wide application in the fields of image recognition and the like. The convolutional neural network may include an input layer, a hidden layer, and an output layer, where the hidden layer may include a convolutional layer, a pooling layer, an activation layer, a full connection layer, etc., where the previous layer performs a corresponding operation according to input data, outputs an operation result to the next layer, and obtains a final result after the input initial data is subjected to a multi-layer operation.
Accordingly, in one specific example, the image feature extraction module 140 is configured to: and respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on input data in forward transmission of layers by using each layer of the convolutional neural network model using the hole convolution kernel so as to output the plurality of monitoring local feature matrices by the last layer of the convolutional neural network model using the hole convolution kernel, wherein the input of the first layer of the convolutional neural network model using the hole convolution kernel is each image block in the sequence of the monitoring image blocks.
More specifically, in the embodiment of the present application, the image feature correlation module 150 is configured to calculate a covariance matrix between each two monitored local feature matrices in the plurality of monitored local feature matrices to obtain a decoded feature map composed of a plurality of covariance matrices. When the dust state distribution detection of the building construction site is actually performed, the distribution state characteristics of dust in each image block have an association relationship, and the association relationship shows different association characteristics under different dust distribution states. Therefore, if it is desired to further improve the expression of the dust distribution state feature, in the technical solution of the present application, the covariance matrix between every two monitoring local feature matrices in the plurality of monitoring local feature matrices is further calculated to obtain a decoding feature map composed of a plurality of covariance matrices, so as to represent the correlation feature distribution information about the dust distribution feature in every two image blocks.
Accordingly, in one specific example, the image feature association module 150 is configured to: calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes according to the following covariance formula to obtain the plurality of covariance matrixes; wherein, the covariance formula is:
wherein M is 1 And M 2 Representing each two monitoring local feature matrices in the plurality of monitoring local feature matrices, M representing a space between each two monitoring local feature matrices in the plurality of monitoring local feature matricesA covariance matrix is formed by the sum of the squares,representing matrix multiplication; and arranging the covariance matrices into the decoding feature map.
More specifically, in the embodiment of the present application, the feature optimization module 160 is configured to perform feature distribution optimization on the decoding feature map to obtain an optimized decoding feature map.
Accordingly, in one specific example, as shown in fig. 3, the feature optimization module 160 includes: an optimization factor calculation unit 161 for calculating a piece-wise approximation factor of the convex-decomposition-based feature geometry metric for each of the decoding feature matrices along the channel dimension to obtain a plurality of piece-wise approximation factors of the convex-decomposition-based feature geometry metric; a weighted optimization unit 162, configured to perform weighted optimization on a plurality of decoding feature matrices of the decoding feature map along a channel dimension by using the piece-by-piece approximation factors of the feature geometric metrics based on the convex decomposition as weighting coefficients to obtain a plurality of optimized decoding feature matrices; and a dimension reconstruction unit 163, configured to perform dimension reconstruction on the plurality of optimized decoding feature matrices to obtain the optimized decoding feature map.
In particular, in the technical solution of the present application, the decoding feature map is obtained by arranging the covariance matrices along a channel dimension, and each covariance matrix in the covariance matrices expresses covariance features of block image semantics of image blocks of the grayscale monitoring image expressed by the monitoring local feature matrix, so that, due to source image semantic differences among the image blocks of the grayscale monitoring image, distribution among the image semantic features expressed by the monitoring local feature matrix may exhibit larger non-uniformity, thereby resulting in non-uniformity of manifold geometric representations with high-dimensional feature manifolds among the covariance matrices obtained by calculating covariance matrices between every two monitoring local feature matrices. Therefore, when the covariance matrixes are arranged along the channel to form the decoding feature map, the convergence difficulty of the decoding feature map in decoding regression through the decoder is improved, and therefore the training speed and the accuracy of the converged decoding value are reduced. The present application thus calculates a slice-wise approximation factor of the convex decomposition-based feature geometry metric for each covariance matrix of the decoded feature map.
Accordingly, in a specific example, the optimization factor calculating unit 161 is configured to: calculating the piecewise approximation factors of the convex decomposition-based feature geometry metrics of each decoded feature matrix of the decoded feature map along the channel dimension with the following optimization formula to obtain the piecewise approximation factors of the plurality of convex decomposition-based feature geometry metrics; wherein, the optimization formula is:
wherein V is ij Is the j-th row vector or column vector of the i-th decoding feature matrix, sigmoid (·) represents Sigmoid function, logSumExp (·) represents LogSumExp function, [:]representing concatenating individual vectors, anRepresenting the square of the two norms of the vector, w i Representing an ith one of the plurality of piecewise approximation factors of the convex-based feature geometry metric.
In particular, the piece-wise approximation factor of the convex decomposition-based feature geometry metric may define a symbolized distance metric between local geometries of the high-dimensional feature manifold of each covariance matrix by a smooth maximum function of LogSumExp to obtain a minutely convex indicator of each convex polyhedron object based on the convex polyhedron decomposition of the high-dimensional feature manifold, and further determine a hyperplane distance parameter for expressing the learnable piece-wise convex decomposition of the high-dimensional feature manifold by a Sigmoid function to approximately measure feature geometry. In this way, by weighting the covariance matrix by the slice-by-slice approximation factor based on the feature geometric measure of the convex decomposition, the consistency of manifold geometric representation of the high-dimensional feature manifold among different covariance matrices under different channels of the decoding feature map obtained by arranging the covariance matrix according to the channels can be improved, so that the convergence effect of the decoding feature map when decoding regression is carried out by a decoder is improved, and the training speed of the model and the accuracy of decoding values are improved. Therefore, the self-adaptive control of the opening value of the spraying valve can be performed based on the dust environment conditions around the building construction site, so that the dust removal efficiency and effect are optimized, and the waste of water resources is avoided.
More specifically, in the embodiment of the present application, the valve opening control module 170 is configured to perform decoding regression on the optimized decoding feature map through a decoder to obtain a decoded value, where the decoded value is used to represent the recommended valve opening value. That is, the method decodes the relevant characteristic information about the dust distribution characteristic in each image block, so as to accurately analyze the environment dust state condition around the construction site, and thus, the self-adaptive control of the opening value of the spraying valve is performed, thereby optimizing the dust removal efficiency and effect and avoiding the waste of water resources.
Accordingly, in one specific example, the valve opening control module 170 is configured to: performing decoding regression on the optimized decoding feature map using a plurality of fully connected layers of the decoder in a decoding regression formula to obtain the decoded values; wherein, the decoding regression formula is:
wherein X is the optimized decoding feature map, Y is the decoding value, W is a weight matrix,representing a matrix multiplication.
In summary, the dust shielding system 100 for building construction according to the embodiment of the present application is illustrated, firstly, a monitoring image around a building construction site is subjected to image graying to obtain a graying monitoring image, then, the graying monitoring image is subjected to image block division, a plurality of monitoring local feature matrices are obtained through a convolutional neural network model, covariance matrices between every two monitoring local feature matrices in the plurality of monitoring local feature matrices are calculated to obtain a decoding feature map composed of a plurality of covariance matrices, and finally, after feature distribution optimization is performed on the decoding feature map, decoding regression is performed through a decoder to obtain a decoding value for representing a recommended valve opening value. In this way, dust removal efficiency and effect can be optimized.
As described above, the dust-shielding system 100 for construction according to the embodiment of the present application can be implemented in various terminal devices, for example, a server or the like having the dust-shielding algorithm for construction according to the embodiment of the present application. In one example, the dust shielding system 100 for construction according to an embodiment of the present application may be integrated into a terminal device as one software module and/or hardware module. For example, the dust-shielding system 100 for construction according to the embodiment of the present application may be a software module in the operating system of the terminal device, or may be an application program developed for the terminal device; of course, the dust shielding system 100 for construction according to the embodiment of the present application may be one of a plurality of hardware modules of the terminal device.
Alternatively, in another example, the dust shielding system 100 for construction according to an embodiment of the present application and the terminal device may be separate devices, and the dust shielding system 100 for construction may be connected to the terminal device through a wired and/or wireless network and transmit interactive information in an agreed data format.
Fig. 4 is a flowchart of a dust shielding method for construction according to an embodiment of the present application. As shown in fig. 4, the dust shielding method for construction according to an embodiment of the present application includes: s110, acquiring monitoring images around a building construction site acquired by a camera; s120, carrying out image graying on the monitoring image around the building construction site to obtain a grayed monitoring image; s130, dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks; s140, each image block in the sequence of the monitoring image blocks is subjected to a convolutional neural network model by using a hole convolutional kernel to obtain a plurality of monitoring local feature matrixes; s150, calculating covariance matrixes among every two monitoring local feature matrixes in the monitoring local feature matrixes to obtain a decoding feature diagram composed of the covariance matrixes; s160, optimizing the feature distribution of the decoding feature map to obtain an optimized decoding feature map; and S170, carrying out decoding regression on the optimized decoding characteristic map through a decoder to obtain a decoding value, wherein the decoding value is used for representing the recommended valve opening value.
Fig. 5 is a schematic diagram of a system architecture of a dust shielding method for construction according to an embodiment of the present application. As shown in fig. 5, in the system architecture of the dust shielding method for construction, first, a monitoring image around a construction site collected by a camera is acquired; then, carrying out image graying on the monitoring image around the building construction site to obtain a grayed monitoring image; then, dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks; then, each image block in the sequence of the monitoring image blocks is subjected to a convolutional neural network model by using a hole convolutional kernel to obtain a plurality of monitoring local feature matrixes; then, calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes to obtain a decoding feature diagram composed of a plurality of covariance matrixes; then, carrying out feature distribution optimization on the decoding feature map to obtain an optimized decoding feature map; and finally, carrying out decoding regression on the optimized decoding characteristic diagram through a decoder to obtain a decoding value, wherein the decoding value is used for representing the recommended valve opening value.
In a specific example, in the above dust shielding method for construction, the image block division of the graying monitoring image to obtain a sequence of monitoring image blocks includes: and uniformly dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks.
In a specific example, in the dust shielding method for construction, the step of obtaining a plurality of monitoring local feature matrices by using a convolutional neural network model of a hole convolutional kernel for each image block in the sequence of monitoring image blocks includes: and respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on input data in forward transmission of layers by using each layer of the convolutional neural network model using the hole convolution kernel so as to output the plurality of monitoring local feature matrices by the last layer of the convolutional neural network model using the hole convolution kernel, wherein the input of the first layer of the convolutional neural network model using the hole convolution kernel is each image block in the sequence of the monitoring image blocks.
In a specific example, in the dust shielding method for construction described above, calculating a covariance matrix between each two monitoring local feature matrices among the plurality of monitoring local feature matrices to obtain a decoded feature map composed of a plurality of covariance matrices includes: calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes according to the following covariance formula to obtain the plurality of covariance matrixes; wherein, the covariance formula is:
Wherein M is 1 And M 2 Representing each two monitored local feature matrices of the plurality of monitored local feature matrices, M representing a covariance matrix between each two monitored local feature matrices of the plurality of monitored local feature matrices,representing matrix multiplication; and arranging the covariance matrices into the decoding feature map.
In a specific example, in the dust shielding method for building construction, the optimizing the feature distribution of the decoding feature map to obtain an optimized decoding feature map includes: calculating the piece-wise approximation factors of the convex-decomposition-based feature geometry metrics of each decoding feature matrix of the decoding feature graph along the channel dimension to obtain a plurality of piece-wise approximation factors of the convex-decomposition-based feature geometry metrics; taking the piece-by-piece approximation factors of the feature geometric metrics based on convex decomposition as weighting coefficients to carry out weighted optimization on a plurality of decoding feature matrixes of the decoding feature graph along the channel dimension so as to obtain a plurality of optimized decoding feature matrixes; and performing dimension reconstruction on the optimized decoding feature matrixes to obtain the optimized decoding feature map.
In a specific example, in the dust shielding method for construction described above, calculating the piece-wise approximation factors of the convex-decomposition-based feature geometry metrics of the respective decoding feature matrices of the decoding feature map along the channel dimension to obtain a plurality of piece-wise approximation factors of the convex-decomposition-based feature geometry metrics includes: calculating the piecewise approximation factors of the convex decomposition-based feature geometry metrics of each decoded feature matrix of the decoded feature map along the channel dimension with the following optimization formula to obtain the piecewise approximation factors of the plurality of convex decomposition-based feature geometry metrics; wherein, the optimization formula is:
Wherein V is ij Is the j-th row vector or column vector of the i-th decoding feature matrix, sigmoid (·) represents Sigmoid function, logSumExp (·) represents LogSumExp function, [:]representing concatenating individual vectors, anRepresenting the square of the two norms of the vector, w i Representing an ith one of the plurality of piecewise approximation factors of the convex-based feature geometry metric.
In a specific example, in the dust shielding method for construction, the optimizing decoding feature map is decoded by a decoder and returned to obtain a decoded value, and the decoded value is used for representing a recommended valve opening value, and the method includes: performing decoding regression on the optimized decoding feature map using a plurality of fully connected layers of the decoder in a decoding regression formula to obtain the decoded values; wherein, the decoding regression formula is:
wherein X is the optimized decoding feature map, Y is the decoding value, W is a weight matrix,representing a matrix multiplication.
Here, it will be understood by those skilled in the art that the specific operations of the respective steps in the above-described dust shielding method for construction have been described in detail in the above description of the dust shielding system 100 for construction with reference to fig. 1 to 3, and thus, repetitive descriptions thereof will be omitted.
According to another aspect of the present application there is also provided a non-volatile computer readable storage medium having stored thereon computer readable instructions which when executed by a computer can perform a method as described above.
Program portions of the technology may be considered to be "products" or "articles of manufacture" in the form of executable code and/or associated data, embodied or carried out by a computer readable medium. A tangible, persistent storage medium may include any memory or storage used by a computer, processor, or similar device or related module. Such as various semiconductor memories, tape drives, disk drives, or the like, capable of providing storage functionality for software.
The application uses specific words to describe embodiments of the application. Reference to "a first/second embodiment," "an embodiment," and/or "some embodiments" means that a particular feature, structure, or characteristic is associated with at least one embodiment of the application. Thus, it should be emphasized and should be appreciated that two or more references to "an embodiment" or "one embodiment" or "an alternative embodiment" in various positions in this specification are not necessarily referring to the same embodiment. Furthermore, certain features, structures, or characteristics of one or more embodiments of the application may be combined as suitable.
Furthermore, those skilled in the art will appreciate that the various aspects of the application are illustrated and described in the context of a number of patentable categories or circumstances, including any novel and useful procedures, machines, products, or materials, or any novel and useful modifications thereof. Accordingly, aspects of the application may be performed entirely by hardware, entirely by software (including firmware, resident software, micro-code, etc.) or by a combination of hardware and software. The above hardware or software may be referred to as a "data block," module, "" engine, "" unit, "" component, "or" system. Furthermore, aspects of the application may take the form of a computer product, comprising computer-readable program code, embodied in one or more computer-readable media.
Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.
The foregoing is illustrative of the present invention and is not to be construed as limiting thereof. Although a few exemplary embodiments of this invention have been described, those skilled in the art will readily appreciate that many modifications are possible in the exemplary embodiments without materially departing from the novel teachings and advantages of this invention. Accordingly, all such modifications are intended to be included within the scope of this invention as defined in the following claims. It is to be understood that the foregoing is illustrative of the present invention and is not to be construed as limited to the specific embodiments disclosed, and that modifications to the disclosed embodiments, as well as other embodiments, are intended to be included within the scope of the appended claims. The invention is defined by the claims and their equivalents.

Claims (10)

1. A dust shielding system for construction, comprising:
the monitoring image acquisition module is used for acquiring monitoring images around the construction site acquired by the camera;
the gray processing module is used for carrying out image graying on the monitoring images around the building construction site to obtain grayed monitoring images;
the image dividing module is used for dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks;
The image feature extraction module is used for obtaining a plurality of monitoring local feature matrixes through a convolutional neural network model using a cavity convolutional kernel for each image block in the sequence of the monitoring image blocks;
the image feature association module is used for calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes to obtain a decoding feature diagram composed of a plurality of covariance matrixes;
the feature optimization module is used for carrying out feature distribution optimization on the decoding feature map so as to obtain an optimized decoding feature map; and
and the valve opening control module is used for carrying out decoding regression on the optimized decoding characteristic diagram through a decoder to obtain a decoding value, wherein the decoding value is used for representing the recommended valve opening value.
2. The dust shielding system for construction according to claim 1, wherein the image dividing module is configured to:
and uniformly dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks.
3. The dust shielding system for construction according to claim 2, wherein the image feature extraction module is configured to:
and respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on input data in forward transmission of layers by using each layer of the convolutional neural network model using the hole convolution kernel so as to output the plurality of monitoring local feature matrices by the last layer of the convolutional neural network model using the hole convolution kernel, wherein the input of the first layer of the convolutional neural network model using the hole convolution kernel is each image block in the sequence of the monitoring image blocks.
4. A dust shielding system for construction according to claim 3, wherein the image feature correlation module is configured to:
calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes according to the following covariance formula to obtain the plurality of covariance matrixes;
wherein, the covariance formula is:
wherein M is 1 And M 2 Representing each two monitored local feature matrices of the plurality of monitored local feature matrices, M representing a covariance matrix between each two monitored local feature matrices of the plurality of monitored local feature matrices,representing matrix multiplication; and
and arranging the covariance matrixes into the decoding characteristic diagram.
5. The dust shielding system for construction of claim 4, wherein the feature optimization module comprises:
an optimization factor calculation unit, configured to calculate a piece-wise approximation factor of the convex-decomposition-based feature geometry metric for each of the decoding feature matrices along the channel dimension to obtain a plurality of piece-wise approximation factors of the convex-decomposition-based feature geometry metric;
the weighted optimization unit is used for carrying out weighted optimization on a plurality of decoding feature matrixes of the decoding feature graph along the channel dimension by taking the piece-by-piece approximation factors of the feature geometric metrics based on the convex decomposition as weighted coefficients so as to obtain a plurality of optimized decoding feature matrixes; and
And the dimension reconstruction unit is used for carrying out dimension reconstruction on the plurality of optimized decoding feature matrixes to obtain the optimized decoding feature map.
6. The dust shielding system for construction according to claim 5, wherein the optimization factor calculation unit is configured to:
calculating the piecewise approximation factors of the convex decomposition-based feature geometry metrics of each decoded feature matrix of the decoded feature map along the channel dimension with the following optimization formula to obtain the piecewise approximation factors of the plurality of convex decomposition-based feature geometry metrics;
wherein, the optimization formula is:
wherein V is ij Is the j-th row vector or column vector of the i-th decoding feature matrix, sigmoid (·) represents Sigmoid function, logSumExp (·) represents LogSumExp function, [:]representing concatenating individual vectors, anRepresenting the square of the two norms of the vector, w i Representing an ith one of the plurality of piecewise approximation factors of the convex-based feature geometry metric.
7. The dust shielding system for construction of claim 6, wherein the valve opening control module is configured to:
performing decoding regression on the optimized decoding feature map using a plurality of fully connected layers of the decoder in a decoding regression formula to obtain the decoded values;
Wherein, the decoding regression formula is:
wherein X is the optimized decoding feature map, Y is the decoding value, W is a weight matrix,representing a matrix multiplication.
8. A dust shielding method for construction, comprising:
acquiring monitoring images around a building construction site acquired by a camera;
performing image graying on the monitoring image around the building construction site to obtain a grayed monitoring image;
dividing the graying monitoring image into image blocks to obtain a sequence of monitoring image blocks;
each image block in the sequence of the monitoring image blocks is subjected to a convolutional neural network model by using a cavity convolutional kernel to obtain a plurality of monitoring local feature matrixes;
calculating covariance matrixes among every two monitoring local feature matrixes in the plurality of monitoring local feature matrixes to obtain a decoding feature diagram composed of a plurality of covariance matrixes;
performing feature distribution optimization on the decoding feature map to obtain an optimized decoding feature map; and
and carrying out decoding regression on the optimized decoding characteristic map through a decoder to obtain a decoding value, wherein the decoding value is used for representing the recommended valve opening value.
9. The dust shielding method for construction according to claim 8, wherein image block dividing the grayscaled monitoring image to obtain a sequence of monitoring image blocks, comprises:
And uniformly dividing the gray monitoring image into image blocks to obtain a sequence of monitoring image blocks.
10. The dust shielding method for construction according to claim 9, wherein the step of obtaining a plurality of monitoring local feature matrices from each image block in the sequence of monitoring image blocks by using a convolutional neural network model of a hole convolution kernel, comprises:
and respectively carrying out convolution processing, pooling processing along a channel dimension and nonlinear activation processing on input data in forward transmission of layers by using each layer of the convolutional neural network model using the hole convolution kernel so as to output the plurality of monitoring local feature matrices by the last layer of the convolutional neural network model using the hole convolution kernel, wherein the input of the first layer of the convolutional neural network model using the hole convolution kernel is each image block in the sequence of the monitoring image blocks.
CN202311033366.4A 2023-08-16 2023-08-16 Dust shielding system and method for building construction Pending CN117115731A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311033366.4A CN117115731A (en) 2023-08-16 2023-08-16 Dust shielding system and method for building construction

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311033366.4A CN117115731A (en) 2023-08-16 2023-08-16 Dust shielding system and method for building construction

Publications (1)

Publication Number Publication Date
CN117115731A true CN117115731A (en) 2023-11-24

Family

ID=88801392

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311033366.4A Pending CN117115731A (en) 2023-08-16 2023-08-16 Dust shielding system and method for building construction

Country Status (1)

Country Link
CN (1) CN117115731A (en)

Similar Documents

Publication Publication Date Title
Labach et al. Survey of dropout methods for deep neural networks
She et al. Wear indicator construction of rolling bearings based on multi-channel deep convolutional neural network with exponentially decaying learning rate
CN114707227A (en) Dam safety early warning and warning method and system based on digital twins
CN104484890B (en) Video target tracking method based on compound sparse model
CN109033513B (en) Power transformer fault diagnosis method and power transformer fault diagnosis device
CN112115419A (en) System state estimation method and system state estimation device
CN116832588B (en) Acid regeneration flue gas purifying device and method thereof
CN112580784A (en) Intelligent early warning method for equipment based on multi-input multi-output convolutional neural network
CN113051130A (en) Mobile cloud load prediction method and system of LSTM network combined with attention mechanism
CN108537825B (en) Target tracking method based on transfer learning regression network
CN115294563A (en) 3D point cloud analysis method and device based on Transformer and capable of enhancing local semantic learning ability
Wu et al. Msstn: Multi-scale spatial temporal network for air pollution prediction
CN116227689A (en) Photovoltaic panel pollution degree state evaluation and cleaning decision system
CN115658672A (en) KPIs (Kernel Key information systems) missing value interpolation method based on improved Transformer and masking training
Zhang et al. Gated recurrent unit-enhanced deep convolutional neural network for real-time industrial process fault diagnosis
CN110852522A (en) Short-term power load prediction method and system
CN112580798A (en) Intelligent early warning method for equipment based on multi-input multi-output ResNet
CN117115731A (en) Dust shielding system and method for building construction
CN116842827A (en) Electromagnetic performance boundary model construction method for unmanned aerial vehicle flight control system
CN116645395A (en) Track prediction method, device, equipment and medium for robot inspection target
CN115982555A (en) Sedimentation prediction method based on fuzzy information granulation and dynamic neural network
CN113435780B (en) Emergency communication sensing equipment system efficiency evaluation method based on neural network
Ding et al. A rolling bearing fault diagnosis method based on Markov transition field and multi-scale Runge-Kutta residual network
CN115081609A (en) Acceleration method in intelligent decision, terminal equipment and storage medium
Mateus et al. Comparing LSTM and GRU Models to Predict the Condition of a Pulp Paper Press. Energies 2021, 14, 6958

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination