Intelligent sensor monitoring system and method in building construction process

Abstract

The invention relates to the technical field of intelligent monitoring, and more specifically discloses an intelligent sensor monitoring system and method for a building construction process, and the method comprises the steps: extracting a concrete pouring parameter input matrix from pouring temperature values, pouring pressure values and pouring humidity values at a plurality of preset time points; and obtaining a concrete pouring parameter feature vector through a first convolutional neural network model of spatial attention, and obtaining a classification result through a classifier, thereby representing whether the concrete pouring process is abnormal or not.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring technology, and more specifically, to an intelligent sensor monitoring system and method for the construction process of a building. Background Art

[0002] A construction site is a location where a building project is under development and civil engineering is being carried out. Its scope is enclosed by hoardings, wire fences or walls, restricting the entry and exit of personnel, materials, machinery and vehicles. There are often various hazard sources at a construction site. The hazard sources mainly come from physical factors and behavioral factors. Physical factors refer to defects in construction equipment and protective equipment, and behavioral factors refer to abnormal working states of workers.

[0003] Nowadays, the demand monitoring positions are mainly determined based on construction standards, the construction conditions are monitored based on established construction monitoring modes, and the condition judgment and early warning are carried out for the monitoring data based on expert experience combined with auxiliary analysis tools. The current construction monitoring methods have obvious drawbacks and need further technological innovation. In the prior art, the flexibility and rigor of the monitoring methods for building construction are insufficient, the monitoring dimensions are not comprehensive enough, and the monitoring control parameters do not fit well with the actual construction conditions, resulting in inaccurate and incomplete monitoring data, and thus there are drawbacks in the construction risk management.

[0004] Therefore, an intelligent sensor monitoring system and method for the construction process of a building are desired. Summary of the Invention

[0005] To solve the above technical problems, this application is proposed. Embodiments of this application provide an intelligent sensor monitoring system and method for the construction process of a building, which realize intelligent monitoring of abnormalities in the concrete pouring process by extracting concrete pouring parameters and using a deep learning model for feature extraction and classification.

[0006] Correspondingly, according to one aspect of this application, an intelligent sensor monitoring system for the construction process of a building is provided, which includes:

[0007] A building construction data acquisition module, configured to obtain pouring temperature values, pouring pressure values and pouring humidity values at a plurality of predetermined time points including the current time point;

[0008] A building construction data processing module, configured to extract a concrete pouring parameter input matrix from the pouring temperature values, pouring pressure values and pouring humidity values at the plurality of predetermined time points;

[0009] A building construction data coding module, configured to obtain a concrete pouring parameter feature vector by passing the input matrix of the concrete pouring parameters through a first convolutional neural network model with spatial attention, and perform bitwise variational entropy optimization of global regression on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector;

[0010] A building construction data control module, configured to obtain a classification result by passing the optimized concrete pouring parameter feature vector through a classifier, where the classification result is used to indicate whether there is an abnormality in the concrete pouring process.

[0011] According to another aspect of the present application, there is also provided an intelligent sensor monitoring method for a building construction process, which includes:

[0012] Obtain pouring temperature values, pouring pressure values, and pouring humidity values at multiple predetermined time points including the current time point;

[0013] Extract an input matrix of concrete pouring parameters from the pouring temperature values, pouring pressure values, and pouring humidity values at the multiple predetermined time points;

[0014] Pass the input matrix of the concrete pouring parameters through a first convolutional neural network model with spatial attention to obtain a concrete pouring parameter feature vector, and perform bitwise variational entropy optimization of global regression on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector;

[0015] Pass the optimized concrete pouring parameter feature vector through a classifier to obtain a classification result, where the classification result is used to indicate whether there is an abnormality in the concrete pouring process.

[0016] Compared with the prior art, an intelligent sensor monitoring system and method for a building construction process provided by the present application extract an input matrix of concrete pouring parameters from the pouring temperature values, pouring pressure values, and pouring humidity values at multiple predetermined time points, obtain a concrete pouring parameter feature vector through a first convolutional neural network model with spatial attention, and then obtain a classification result through a classifier, thereby indicating whether there is an abnormality in the concrete pouring process. Description of the Drawings

[0017] By describing the embodiments of the present application in more detail in conjunction with the drawings, the above and other objects, features, and advantages of the present application will become more obvious. The drawings are used to provide a further understanding of the embodiments of the present application, and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application, and do not constitute a limitation to the present application. In the drawings, the same reference numerals generally represent the same components or steps.

[0018] Figure 1Schematic block diagram of an intelligent sensor monitoring system for a building construction process according to an embodiment of the present application.

[0019] Figure 2 Schematic block diagram of a building construction data processing module in an intelligent sensor monitoring system for a building construction process according to an embodiment of the present application.

[0020] Figure 3 Schematic block diagram of a temperature processing unit in an intelligent sensor monitoring system for a building construction process according to an embodiment of the present application.

[0021] Figure 4 Flowchart of an intelligent sensor monitoring method for a building construction process according to an embodiment of the present application. Detailed implementation manners

[0022] Various exemplary embodiments, features, and aspects of the present application will be described in detail below with reference to the accompanying drawings. Identical reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0023] The special term "exemplary" here means "serving as an example, embodiment, or illustration". Any embodiment described as "exemplary" here is not necessarily to be construed as superior to or better than other embodiments.

[0024] In addition, for a better illustration of the present application, numerous specific details are given in the following detailed implementation manners. Those skilled in the art should understand that the present application can be implemented without some specific details. In some instances, methods, means, elements, and circuits well known to those skilled in the art are not described in detail so as to highlight the gist of the present application.

[0025] Furthermore, the terms "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. In the description of the present application, "a plurality of" means two or more unless otherwise specifically defined.

[0026] Figure 1 The schematic block diagram of an intelligent sensor monitoring system for a building construction process according to an embodiment of the present application is illustrated. As Figure 1As shown, the intelligent sensor monitoring system 100 for the building construction process according to an embodiment of the present application includes: a building construction data acquisition module 110, configured to obtain the pouring temperature values, pouring pressure values, and pouring humidity values at a plurality of predetermined time points including the current time point; a building construction data processing module 120, configured to extract a concrete pouring parameter input matrix from the pouring temperature values, pouring pressure values, and pouring humidity values at the plurality of predetermined time points; a building construction data encoding module 130, configured to obtain a concrete pouring parameter feature vector by passing the concrete pouring parameter input matrix through a first convolutional neural network model with spatial attention, and perform global regression bitwise variational entropy optimization on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector; and a building construction data control module 140, configured to obtain a classification result by passing the optimized concrete pouring parameter feature vector through a classifier, where the classification result is used to indicate whether there is an abnormality in the concrete pouring process.

[0027] In an embodiment of the present application, the building construction data acquisition module 110 is configured to obtain the pouring temperature values, pouring pressure values, and pouring humidity values at a plurality of predetermined time points including the current time point. It should be understood that the concrete pouring process is a dynamic process, and its parameters such as temperature, pressure, and humidity change over time. By obtaining the pouring parameter values at a plurality of predetermined time points, time series data can be constructed to facilitate time series modeling and analysis using a deep learning model. Abnormalities in the concrete pouring process usually manifest as abnormal changes in parameters such as temperature, pressure, or humidity. By obtaining the pouring parameter values at a plurality of predetermined time points, the characteristics of these parameters changing over time can be extracted to facilitate feature extraction and classification using a deep learning model. The sensor data in the concrete pouring process may be affected by noise and interference. By obtaining the pouring parameter values at a plurality of predetermined time points, the data can be smoothed to improve the robustness of the data, thereby improving the accuracy and reliability of the deep learning model. By obtaining the pouring parameter values at a plurality of predetermined time points including the current time point, the concrete pouring process can be predicted and warned using a deep learning model. For example, when the deep learning model detects an abnormal change in the pouring parameter value, a warning can be issued in a timely manner so that relevant personnel can take measures to prevent accidents.

[0028] In the embodiment of the present application, the building construction data processing module 120 is configured to extract a concrete pouring parameter input matrix from the pouring temperature values, pouring pressure values, and pouring humidity values at the plurality of predetermined time points. It should be understood that parameters such as temperature, pressure, and humidity during the concrete pouring process are interrelated and jointly affect the quality and performance of the concrete. By extracting these parameters into a matrix, they can be analyzed and processed as a whole, thereby more comprehensively reflecting the state of the concrete pouring process. Parameters such as temperature, pressure, and humidity each contain different information. By fusing these parameters into a matrix, the information of different parameters can be fused to extract richer features. These fused features can better reflect the overall state of the concrete pouring process and improve the classification accuracy of the deep learning model. The units and dimensions of parameters such as temperature, pressure, and humidity may be different. By extracting these parameters into a matrix, they can be standardized to have the same unit and dimension. This can eliminate the differences between different parameters and enable the deep learning model to better learn and process the data. The deep learning model usually requires a matrix or tensor as input. By extracting the pouring temperature values, pouring pressure values, and pouring humidity values at the plurality of predetermined time points into a concrete pouring parameter input matrix, it can be used as the input of the deep learning model for training and prediction.

[0029] Specifically, in one embodiment of the present application, Figure 2 The figure shows a block diagram of the building construction data processing module in the intelligent sensor monitoring system of the building construction process according to the embodiment of the present application. As Figure 2 shown, in the intelligent sensor monitoring system 100 of the above building construction process, the building construction data processing module 120 includes: a temperature processing unit 121 configured to arrange the pouring temperature values at the plurality of predetermined time points as a vector along the time dimension and then obtain a pouring temperature feature vector through feature extraction; a pressure processing unit 122 configured to arrange the pouring pressure values at the plurality of predetermined time points as a vector along the time dimension and then obtain a pouring pressure feature vector through feature extraction; a humidity processing unit 123 configured to arrange the pouring humidity values at the plurality of predetermined time points as a vector along the time dimension and then obtain a pouring humidity feature vector through feature extraction; a fusion parameter unit 124 configured to arrange the pouring temperature feature vector, the pouring pressure feature vector, and the pouring humidity feature vector along the parameter dimension to obtain the concrete pouring parameter input matrix.

[0030] Correspondingly, in a specific example of the present application, the temperature processing unit 121 is configured to arrange the pouring temperature values at the plurality of predetermined time points into a vector according to the time dimension and then perform feature extraction to obtain a pouring temperature feature vector. It should be understood that the temperature during the concrete pouring process is a dynamic process, and its temperature value changes over time. By arranging the pouring temperature values at a plurality of predetermined time points into a vector according to the time dimension, time series data can be constructed to facilitate time series modeling and analysis using a deep learning model. Abnormalities during the concrete pouring process usually manifest as abnormal changes in temperature values. By performing feature extraction on the pouring temperature vector, features of the temperature value changing over time can be extracted to facilitate feature extraction and classification using a deep learning model. The sensor data during the concrete pouring process may be affected by noise and interference. By performing feature extraction on the pouring temperature vector, the data can be smoothed to improve the robustness of the data, thereby improving the accuracy and reliability of the deep learning model. The pouring temperature vector may contain a large number of data points. By performing feature extraction on the pouring temperature vector, the data can be dimensionally reduced to extract more representative features, thereby reducing the computational complexity of the deep learning model and improving the training speed of the model. A deep learning model usually requires a vector or tensor as input. By arranging the pouring temperature values at the plurality of predetermined time points into a vector according to the time dimension and then performing feature extraction to obtain a pouring temperature feature vector, it can be used as the input of the deep learning model for training and prediction.

[0031] Furthermore, Figure 3 FIG. illustrates a block diagram of a temperature processing unit in an intelligent sensor monitoring system for a building construction process according to an embodiment of the present application. As Figure 3 shown, in the building construction data processing module 120 of the intelligent sensor monitoring system 100 for the building construction process described above, the temperature processing unit 121 includes: a temperature arrangement vector sub-unit 1211 configured to arrange the pouring temperature values at the plurality of predetermined time points into a pouring temperature input vector according to the time dimension; a temperature feature extraction sub-unit 1212 configured to obtain the pouring temperature feature vector by passing the pouring temperature input vector through a temperature extractor based on a multi-scale neighborhood feature extraction module.

[0032] Specifically, the temperature arrangement vector sub-unit 1211 is used to arrange the pouring temperature values at the multiple predetermined time points into a pouring temperature input vector in the time dimension. It should be understood that during the concrete pouring process, the temperature changes over time and has an obvious time correlation. Arranging the pouring temperature values at multiple predetermined time points into a vector in the time dimension helps to retain the information of temperature change over time, enabling the model to better capture the laws and trends of temperature change. Arranging the pouring temperature values into a vector in the time dimension can form a time series data, enabling the deep learning model to better understand and utilize the time series relationship between temperature values. This helps to improve the model's prediction ability for temperature change. After arranging the pouring temperature values at multiple time points into a vector, various feature extraction techniques, such as sliding window, time series feature extraction, etc., can be applied to extract more meaningful features from it. These features can more comprehensively describe the pattern of temperature change over time and help to improve the performance of the model. Deep learning models usually require the input to be a vector or tensor with a fixed dimension. Arranging the pouring temperature values into a pouring temperature input vector in the time dimension enables the data to be directly used as the input of the model, simplifying the data processing and model construction processes. Integrating the pouring temperature values at multiple time points into a vector can maintain the dimensional consistency of the data, making the data easier to manage and operate during the processing. Therefore, arranging the pouring temperature values at the multiple predetermined time points into a pouring temperature input vector in the time dimension helps to retain time series information, extract key features, simplify the model input, and improve the model performance. This way can more effectively utilize the temperature data and provide richer and more accurate information for the deep learning model, thereby improving the model's prediction and analysis ability during the concrete pouring process.

[0033] Specifically, the temperature feature extraction subunit 1212 is configured to obtain the pouring temperature feature vector by passing the pouring temperature input vector through a temperature extractor based on a multi-scale neighborhood feature extraction module. It should be understood that during the concrete pouring process, there may be feature information at different scales in the temperature change. By using a multi-scale neighborhood feature extraction module, the features of temperature data can be extracted at different scales, capturing the diversity and complexity of temperature changes, so as to more comprehensively describe the law of temperature changes. Multi-scale feature extraction can help the model understand temperature data from different levels and perspectives, and extract richer and more detailed feature information. This helps to enhance the model's representation ability of temperature changes, improve the generalization performance and accuracy of the model. The temperature extractor based on multi-scale neighborhood feature extraction can learn the feature representations at different scales, making the model more generalizable, capable of adapting to temperature change patterns at different scales, and improving the model's performance on unknown data. Multi-scale feature extraction can effectively reduce the loss of information because the features at different scales can complement and correct each other, improving the model's understanding and modeling ability of temperature data. By using a temperature extractor based on a multi-scale neighborhood feature extraction module, the model's ability to abstract and represent temperature data can be improved, thereby improving the performance of the deep learning model on pouring temperature data, including prediction accuracy and generalization ability. Therefore, by passing the pouring temperature input vector through a temperature extractor based on a multi-scale neighborhood feature extraction module to obtain a pouring temperature feature vector, the multi-scale feature information of temperature data can be more fully utilized, the representation ability and performance of the model can be improved, and thus the temperature changes during the concrete pouring process can be analyzed and predicted more effectively.

[0034] Correspondingly, the temperature feature extraction subunit includes: a first-scale temperature encoding secondary subunit, configured to perform one-dimensional convolutional encoding on the pouring temperature input vector with a one-dimensional convolutional kernel of a first scale using a first convolutional layer of the temperature extractor based on the multi-scale neighborhood feature extraction module to obtain a first-scale pouring temperature feature vector, wherein the first convolutional layer has a first one-dimensional convolutional kernel of a first length; a second-scale temperature encoding secondary subunit, configured to perform one-dimensional convolutional encoding on the pouring temperature input vector with a one-dimensional convolutional kernel of a second scale using a second convolutional layer of the temperature extractor based on the multi-scale neighborhood feature extraction module to obtain a second-scale pouring temperature feature vector, wherein the second convolutional layer has a second one-dimensional convolutional kernel of a second length, and the first length is different from the second length; a multi-scale temperature concatenation secondary subunit, configured to concatenate the first-scale pouring temperature feature vector and the second-scale pouring temperature feature vector to obtain the pouring temperature feature vector.

[0035] Further, the first-scale temperature encoding secondary unit is configured to: perform one-dimensional convolution encoding on the pouring temperature input vector using the first convolutional layer of the temperature extractor of the multi-scale neighborhood feature extraction module according to the following first-scale convolution formula to obtain a first-scale pouring temperature feature vector;

[0036] wherein, the first-scale convolution formula is:

[0037]

[0038] where a is the width of the first convolutional kernel in the x direction, F(a) is the first convolutional kernel parameter vector, G(x - a) is the local vector matrix for operation with the first convolutional kernel function, w is the size of the first convolutional kernel, X represents the pouring temperature input vector, and Cov1(X) represents the first-scale pouring temperature feature vector;

[0039] The second-scale feature extraction unit is configured to: perform one-dimensional convolution encoding on the pouring temperature input vector using the second convolutional layer of the temperature extractor of the multi-scale neighborhood feature extraction module according to the following second-scale convolution formula to obtain the second-scale pouring temperature feature vector;

[0040] wherein, the second-scale convolution formula is:

[0041]

[0042] where b is the width of the second convolutional kernel in the x direction, F(b) is the second convolutional kernel parameter vector, G(x - b) is the local vector matrix for operation with the second convolutional kernel function, m is the size of the second-dimensional convolutional kernel, X represents the pouring temperature input vector, and Cov2(X) represents the second-scale pouring temperature feature vector.

[0043] Correspondingly, in a specific example of the present application, the pressure processing unit 122 is configured to arrange the pouring pressure values at the plurality of predetermined time points into a vector according to the time dimension and then obtain a pouring pressure feature vector through feature extraction. It should be understood that arranging the pouring pressure values at multiple time points into a vector can effectively compress the original data, reduce the data dimension, and facilitate subsequent processing and analysis. Through the feature extraction method, more representative and discriminative features can be extracted from the pouring pressure data, helping the model better understand and utilize the data. The pouring pressure feature vector obtained after feature extraction can be used as the input of the deep learning model, simplifying the input data of the model and reducing the complexity of model training. Feature extraction can help identify patterns and regularities in the pouring pressure data, improve the model's understanding and prediction ability of the pressure changes during the pouring process. Through feature extraction, the dimension and redundant information of the data can be reduced, the computational complexity of the model can be lowered, and the training and inference efficiency of the model can be improved. Feature extraction can help the model learn more abstract and generalized feature representations, improve the model's generalization ability on unseen data, and enable it to better adapt to the pouring pressure changes in different situations.

[0044] Specifically, the pressure processing unit includes: a pressure arrangement vector sub-unit, configured to arrange the pouring pressure values at the plurality of predetermined time points into a pouring pressure input vector according to the time dimension; and a pressure feature extraction sub-unit, configured to obtain the pouring pressure feature vector by passing the pouring pressure input vector through a pressure extractor based on a multi-scale neighborhood feature extraction module.

[0045] Further, the pressure arrangement vector sub-unit is configured to arrange the pouring pressure values at the plurality of predetermined time points into a pouring pressure input vector according to the time dimension. It should be understood that the pressure values during the pouring process are usually time series data that change with time. Arranging these pressure values according to the time dimension can retain the time information and help capture the trends and regularities of the pressure change over time. Arranging the pressure values at multiple time points into a vector can retain the temporal relationship between them, which is beneficial for the model to understand the correlation and influence between different time points and improve the model's performance on time series data. Many deep learning models and time series analysis methods require the input data to be in a specific format. Arranging the pressure values at time points into a vector can meet the input requirements of these models and facilitate model training and prediction. The time series data arranged into a vector can extract higher-level features through the feature extraction method, helping the model better understand the data and discover the patterns and features hidden behind the data. Arranging the pressure values into a vector according to the time dimension helps in the processing and management of the data, simplifying the processes of data storage, processing, and analysis.

[0046] Furthermore, the pressure feature extraction sub-unit is configured to obtain the pouring pressure feature vector by passing the pouring pressure input vector through a pressure extractor based on a multi-scale neighborhood feature extraction module. It should be understood that the multi-scale neighborhood feature extraction module can capture feature information in pressure data from different scales, including local details and global structures, which helps to improve the richness and expressiveness of features. The pressure extractor based on multi-scale neighborhood feature extraction can extract features from pressure data at different levels, improving the feature representation ability, making the features more discriminative and representative. Multi-scale feature extraction can help the model better understand the structure and pattern in pressure data, contributing to improving the model's understanding and prediction ability of data. Multi-scale feature extraction helps to extract more abstract and generalized feature representations, which can improve the model's generalization ability on unseen data and make the model more versatile. Multi-scale feature extraction can effectively reduce the risk of overfitting of the model, improve the model's generalization ability, and make the model more robust. The feature vector obtained through the pressure extractor based on multi-scale neighborhood feature extraction is more representative and discriminative, which helps to improve the prediction performance and accuracy of the model on pouring pressure data.

[0047] Correspondingly, in a specific example of the present application, the humidity processing unit 123 is configured to arrange the pouring humidity values at the plurality of predetermined time points into a vector according to the time dimension and then obtain a pouring humidity feature vector through feature extraction. It should be understood that the pouring humidity value usually changes with time. Arranging it into a vector according to the time dimension can retain the time series information, which helps to capture the trend and law of humidity change over time. Through feature extraction, more representative and discriminative features can be extracted from the original humidity data, which helps to improve the model's understanding and prediction ability of humidity data. Feature extraction can help reduce the dimension of the data, remove noise and redundant information in the data, improve the data's expression ability and the model's generalization ability. The pouring humidity feature vector obtained after feature extraction can better meet the input requirements of the model, simplify the input data format of the model, and facilitate model training and prediction. The humidity feature vector obtained through feature extraction is more representative and discriminative, which helps to improve the prediction performance and accuracy of the model on humidity data. Feature extraction can help extract the key features in the data, making the model more interpretable and able to better understand the laws and features in the data.

[0048] Specifically, the humidity processing unit includes: a humidity arrangement vector sub-unit configured to arrange the pouring humidity values at the plurality of predetermined time points into a pouring humidity input vector according to the time dimension; and a humidity feature extraction sub-unit configured to obtain the pouring humidity feature vector by passing the pouring humidity input vector through a humidity extractor based on a multi-scale neighborhood feature extraction module.

[0049] Further, the humidity arrangement vector sub-unit is used to arrange the pouring humidity values at the plurality of predetermined time points into a pouring humidity input vector along the time dimension. It should be understood that by arranging the humidity values along the time dimension, the humidity change information at different time points can be retained, which helps to capture the change trend and law of humidity over time. Arranging the humidity values into a vector along the time dimension can organize the data structure, making the data easier to process and analyze, which is beneficial to the subsequent feature extraction and modeling processes. After arranging the humidity values into a vector, it can be used as the input of the model, which meets the requirements of many machine learning models for data formats and simplifies the input processing process of the model. The humidity data arranged as a vector can be more conveniently subjected to feature extraction operations. For example, models such as one-dimensional convolutional neural networks (CNNs) can be directly applied for feature extraction, improving the expressive ability of the data. Arranging the humidity values into a vector helps to reduce the dimension of the data, remove redundant information, simplify the data representation, and is beneficial to the training and prediction of the model. By arranging the humidity values into a vector along the time dimension, the change law in the time dimension can be better reflected, which helps to improve the training effect and prediction accuracy of the model on time series data.

[0050] Furthermore, the humidity feature extraction sub-unit is used to obtain the pouring humidity feature vector by passing the pouring humidity input vector through a humidity extractor based on a multi-scale neighborhood feature extraction module. It should be understood that the multi-scale neighborhood feature extraction module can capture humidity information at different scales. The multi-scale feature extraction from local to global helps to comprehensively understand the spatial structure and change law of humidity data. Through multi-scale neighborhood feature extraction, rich and diverse humidity features can be extracted, enhancing the expressive ability of humidity data and making the obtained humidity feature vector more representative and discriminative. Multi-scale neighborhood feature extraction helps to improve the model's understanding ability of humidity data, thereby improving the performance and accuracy of the model in humidity prediction or analysis tasks. Through multi-scale neighborhood feature extraction, the humidity information at different spatial scales can be fused together to better understand the spatial distribution characteristics of humidity data. Multi-scale neighborhood feature extraction helps to reduce information loss, retain important features in humidity data, and improve the sensitivity of the model to humidity data. Through multi-scale neighborhood feature extraction, the generalization ability of the model can be improved, enabling the model to better adapt to humidity data of different scales and complexities.

[0051] Accordingly, in a specific example of the present application, the fusion parameter unit 124 is configured to arrange the pouring temperature feature vector, the pouring pressure feature vector, and the pouring humidity feature vector according to the parameter dimension to obtain the concrete pouring parameter input matrix. It should be understood that arranging the temperature, pressure, and humidity feature vectors according to the parameter dimension can make them correspond to the same dimension in the same matrix, facilitating subsequent processing and analysis. Integrating the feature vectors of different parameters into the same matrix helps to integrate the correlation information between different parameters, improving the comprehensiveness and integrity of the data. Arranging the feature vectors into a matrix form according to the parameter dimension can be used as the input of the model, meeting the requirements of many machine learning models for the data format and simplifying the input processing process of the model. Arranging the feature vectors of different parameters into a matrix according to the parameter dimension is beneficial for the model to learn the correlation and interaction between different parameters, improving the model's ability to model the complex relationships between parameters. Arranging the feature vectors into a matrix according to the parameter dimension can reduce the complexity of data processing and calculation, simplifying the model training and prediction processes. Integrating the feature vectors of different parameters into a matrix form helps to improve the feature expression ability, enabling the model to better understand and utilize the information between different parameters.

[0052] In an embodiment of the present application, the construction data encoding module 130 is configured to obtain a concrete pouring parameter feature vector by passing the concrete pouring parameter input matrix through a first convolutional neural network model with spatial attention, and perform global regression bitwise variational entropy optimization on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector. It should be understood that by passing through the first convolutional neural network model with spatial attention, features, including spatial structure and correlation, can be effectively extracted from the concrete pouring parameter input matrix, which helps to capture important patterns and information in the data. The spatial attention mechanism helps the model focus on the importance of different positions in the input data, enabling the model to better understand the spatial relationship of the concrete pouring parameter data and improving the effect of feature extraction.

[0053] Accordingly, in an embodiment of the present application, the construction data encoding module 130 includes: an attention unit configured to obtain a concrete pouring parameter feature vector by passing the concrete pouring parameter input matrix through a first convolutional neural network model with spatial attention; and an optimization unit configured to perform global regression bitwise variational entropy optimization on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector.

[0054] Specifically, the attention unit is used to obtain the concrete pouring parameter feature vector by passing the concrete pouring parameter input matrix through the first convolutional neural network model with spatial attention. It should be understood that using the spatial attention mechanism can help the model pay more attention to important spatial positions during the learning process, thereby improving the effect of feature extraction and enabling the model to better understand the spatial structure and correlation of the concrete pouring parameter data. Through the spatial attention mechanism, the convolutional feature map and the spatial attention score matrix are multiplied element-wise at each position, realizing the weighted position correlation between the feature maps, thus enhancing the feature interaction between different positions and contributing to improving the representation ability of the model. Processing the feature matrix along the channel dimension can retain the information interaction between different channels, contribute to comprehensively utilizing the information of different features, and improve the feature expression ability. Pooling the spatial attention feature map along the channel dimension helps reduce the dimension of the features, extract more significant features, while retaining important information and simplifying the complexity of the model. Nonlinear activation of the pooled feature map can introduce nonlinear factors and increase the expression ability of the model, enabling the model to better learn complex data patterns and features. Through the multi-level structure of the convolutional neural network, the model can gradually learn the abstract feature representation of the data, and the step-by-step extraction of features from low-level to high-level helps improve the model's understanding and modeling ability of the concrete pouring parameters. Therefore, by processing the concrete pouring parameter input matrix through the first convolutional neural network model with spatial attention and using operations such as convolution, spatial attention, pooling, and nonlinear activation, the feature extraction ability and expression ability of the model for the concrete pouring parameter data can be improved, so as to better understand and utilize the concrete pouring parameter data and provide a better basis for subsequent concrete pouring parameter analysis and prediction tasks.

[0055] Specifically, the attention unit is used for: respectively performing, in the forward pass of each layer of the first convolutional neural network model on the input data: performing convolutional processing on the input data based on the convolutional kernel to obtain a convolutional feature map; passing the convolutional feature map through the spatial attention module to obtain the spatial attention score matrix; multiplying the spatial attention score matrix and each feature matrix of the convolutional feature map along the channel dimension element-wise to obtain the spatial attention feature map; performing pooling processing on the spatial attention feature map to obtain a pooled feature map; performing nonlinear activation on the pooled feature map to obtain an activated feature map; wherein, the input of the first layer of the first convolutional neural network model is the concrete pouring parameter input matrix, and the output of the last layer of the first convolutional neural network model is the concrete pouring parameter feature vector.

[0056] Further, the optimization unit is configured to perform bit-by-bit variational entropy optimization of global regression on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector. Specifically, in the above technical solution, on the one hand, the concrete pouring parameter feature vector is a feature extracted based on parameters such as the pouring temperature value, pouring pressure value, and pouring humidity value. After feature extraction and processing, a high-dimensional feature matrix is formed. The high-dimensional concrete pouring parameter feature vector can more comprehensively describe the parameter features in the concrete pouring process, improve the feature representation ability, and help to more accurately judge whether there are abnormal conditions in the concrete pouring process. However, at the same time, the high-dimensional concrete pouring parameter feature vector will also increase the computational complexity, require more computing resources and time to process and train the model, and may lead to a decline in system performance. Excessively high dimensions are also prone to overfitting problems, that is, the model performs well on the training set but poorly on the test set, affecting the generalization ability of the model. On the other hand, in this application scenario, it is desired to improve the monotonicity of the concrete pouring parameter feature vector to better understand and distinguish the relationships between different parameters, so as to accurately judge whether there are abnormal conditions in the concrete pouring process. By improving the monotonicity of the concrete pouring parameter feature vector, the correlations between parameters can be captured more clearly, which helps to reduce the overfitting risk, improve the generalization ability and prediction accuracy of the model. At the same time, improving the monotonicity also helps to simplify the expression of the feature matrix, reduce the complexity of the model, increase the interpretability of the model, and thus better judge whether there are abnormal conditions in the concrete pouring process. Based on this, in the technical solution of this application, bit-by-bit variational entropy optimization of global regression is performed on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector.

[0057] Specifically, the optimization unit is configured to: calculate the autocovariance matrix of the concrete pouring parameter feature vector, and perform principal component dimensionality reduction on the autocovariance matrix of the concrete pouring parameter feature vector to obtain a set of concrete pouring parameter feature basis factor coding vectors, which is expressed by the formula:

[0058]

[0059]

[0060] where V represents the concrete pouring parameter feature vector, T represents the transpose of the matrix, M z represents the autocovariance matrix, U represents the set of concrete pouring parameter feature basis factor coding vectors, v1, v2, v m respectively represent the first, second, and mth concrete pouring parameter feature basis factor coding vectors, Λ represents the concrete pouring parameter feature diagonal matrix after principal component dimensionality reduction, λ1, λ mThey respectively represent the first and mth eigenvalues on the diagonal of the concrete pouring parameter characteristic diagonal matrix.

[0061] That is, by calculating the autocovariance matrix, the periodic fluctuation law and interdependence strength of the features of each dimension can be quantified, revealing the potential correlation pattern hidden in high-frequency vibration. The essence of further implementing PCA is to reconstruct the low-dimensional embedding space through orthogonal transformation, which not only compresses redundant information but also highlights the dominant variable combination that is sensitive to abnormalities in the direction of the principal component that retains the maximum variance contribution. This dual processing mechanism enables the generated set of basic factor encoding vectors of concrete pouring parameter features to effectively distinguish the boundary distribution of normal / abnormal working conditions, and improve the generalization ability of the classifier by reducing the redundancy of the feature space, thereby achieving accurate capture and robust recognition of weak abnormal signals.

[0062] Specifically, the optimization unit is further used to: input the set of the concrete pouring parameter feature basic factor encoding vectors into a sequence encoder based on a forward LSTM model to obtain a set of concrete pouring parameter feature basic factor context-related encoding vectors, which is expressed as follows:

[0063] F=LSTM([v1,v2,…,v m ])=[s1,s2,…,s m ]

[0064] Among them, LSTM represents the forward LSTM model, F represents the set of context-related encoding vectors of the basic factors of concrete pouring parameter features, s1, s2, s m Represents the first, second, and mth concrete pouring parameter feature-based factor context-associated encoding vector.

[0065] That is, through the sequence modeling of LSTM, the temporal dependency of the feature basic factor encoding vector of concrete pouring parameters can be deeply analyzed, so that the system can explore the synergy or conflict rules of different feature dimensions in the spatiotemporal evolution, thereby converting the discrete feature basic factor encoding vector of concrete pouring parameters into a set of context-related encoding vectors of concrete pouring parameter feature basic factors with strong semantic associations, which significantly improves the robustness and generalization of the anomaly detection model.

[0066] Specifically, the optimization unit is further used to: calculate the bitwise variational entropy between each corresponding set of concrete pouring parameter feature basic factor context-associated coding vectors and concrete pouring parameter feature basic factor coding vectors in the set of the concrete pouring parameter feature basic factor coding vectors to obtain a set of bitwise variational entropies, which is expressed as follows:

[0067]

[0068] Among them, represents a logical operator, w represents a bit-by-bit variational comparison function, and v i represents the encoding vector of the i-th concrete pouring parameter characteristic basic factor, represents the characteristic value at the j-th position of the encoding vector of the i-th concrete pouring parameter characteristic basic factor, s i represents the context-related encoding vector of the i-th concrete pouring parameter characteristic basic factor, represents the characteristic value at the j-th position of the context-related encoding vector of the i-th concrete pouring parameter characteristic basic factor, ε represents a predetermined threshold, and r i represents the i-th bit-by-bit variational matching feature vector, represents the characteristic value at the j-th position of the i-th bit-by-bit variational matching feature vector, L represents the length of the i-th bit-by-bit variational matching feature vector, and e i represents the i-th bit-by-bit variational entropy.

[0069] That is, through the quantization of the bit-by-bit variational entropy, the system can reveal the pattern transfer law of features in binary coding from the perspective of information entropy. Its essence is to map the feature difference from the geometric space to the information theory space. By calculating the distribution deviation of each binary bit before and after context-related modeling, a sensitivity index for the increase or decrease of feature information, noise injection, or pattern reconstruction is formed. The set of bit-by-bit variational entropy generated thereby not only provides a quantization threshold for the classifier to distinguish "normal fluctuations" from "abnormal mutations", but also guides the weight allocation of key binary bits in subsequent feature fusion by exposing the dynamic reconstruction mechanism of the information layer, breaking through the dull perception of information perturbation by traditional metrics, achieving a leapfrog discrimination from "numerical similarity" to "semantic consistency", thus significantly improving the system's early warning ability for concealed construction anomalies.

[0070] Specifically, the optimization unit is further configured to: perform a weighting process on the set of bit-by-bit variational entropy to obtain a set of bit-by-bit variational entropy optimization factors, which is expressed by the formula:

[0071] a i = softmax(e i )

[0072] Among them, softmax represents the normalized exponential function, and a i represents the i-th bit-by-bit variational entropy optimization factor.

[0073] That is, through the exponential property of the Softmax function, the entropy value difference is transformed into a dominance measure with clear probability significance. For high-entropy components, since their entropy values are significantly higher than those of other dimensions, the corresponding weights of the optimization factors are exponentially amplified, forming a strong focusing effect on abnormally sensitive features. At the same time, the smoothness constraint of Softmax avoids numerical instability caused by extreme weight distribution, enabling the optimization process to adaptively identify which change patterns of binary bits should be preferentially strengthened and which local perturbations should be suppressed, thereby achieving directional amplification of key information gain and dynamic filtering of noise interference during the feature fusion stage.

[0074] Specifically, the optimization unit is further configured to: based on the set of bit-by-bit variational entropy optimization factors, fuse the set of encoded vectors of the concrete pouring parameter feature basic factors to obtain an optimized concrete pouring parameter feature vector, which is expressed by the formula:

[0075]

[0076] where v f represents the optimized concrete pouring parameter feature vector.

[0077] That is, by introducing the set of bit-by-bit variational entropy optimization factors, the system can dynamically allocate feature weights according to the degree of information change, transforming feature adjustment from "experience-driven global equilibrium" to "data-driven directional strengthening", enabling the generated optimized concrete pouring parameter feature vector to achieve a dynamic balance among dimensional compression rate, abnormal sensitivity, and classification robustness, significantly enhancing the generalization ability of the construction risk warning system and the false alarm rate control level.

[0078] In the embodiment of the present application, the construction data control module 140 is configured to obtain a classification result by passing the optimized concrete pouring parameter feature vector through a classifier, and the classification result is used to indicate whether there is an abnormality in the concrete pouring process. It should be understood that classifying the concrete pouring parameters through a classifier can help identify abnormal situations. The classifier can learn the characteristic patterns of the normal concrete pouring process. When an abnormality occurs, the classifier may output different categories, thereby helping to monitor and detect abnormal situations in the concrete pouring process. Using a classifier for anomaly detection can achieve an automated anomaly detection process, reduce manual intervention, and improve detection efficiency and accuracy. The classification result obtained through the classifier can be used to monitor abnormal situations in the concrete pouring process in real time, and timely measures can be taken to avoid losses caused by potential problems. When the classification result shows an abnormality, preventive measures can be taken in a timely manner to avoid possible problems in the concrete pouring process, ensuring construction quality and safety. The classification result can provide decision-making support for relevant personnel, helping them take corresponding actions according to the abnormal situation to ensure the smooth progress of the project construction.

[0079] Correspondingly, in an embodiment of the present application, the building construction data control module 140 includes: a pouring parameter matrix expansion unit for expanding the optimized concrete pouring parameter feature vector into a concrete pouring parameter classification feature vector according to a row vector or a column vector; a pouring parameter fully connected encoding unit for using the fully connected layer of the classifier to perform fully connected encoding on the concrete pouring parameter classification feature vector to obtain a concrete pouring parameter fully connected encoding feature vector; a pouring parameter probability obtaining unit for passing the concrete pouring parameter fully connected encoding feature vector through the Softmax classification function of the classifier to obtain a first probability belonging to an abnormal concrete pouring process and a second probability belonging to a normal concrete pouring process; and a pouring parameter classification result determination unit for determining the classification result based on a comparison between the first probability and the second probability.

[0080] In summary, based on the intelligent sensor monitoring system and method for the building construction process according to the embodiments of the present application, it extracts a concrete pouring parameter input matrix from the pouring temperature values, pouring pressure values, and pouring humidity values at multiple predetermined time points, obtains a concrete pouring parameter feature vector through the first convolutional neural network model with spatial attention, and then obtains a classification result through a classifier, thereby indicating whether there is an abnormality in the concrete pouring process.

[0081] As described above, the intelligent sensor monitoring system 100 for the building construction process according to the embodiments of the present application can be implemented in various terminal devices, such as the server of the intelligent sensor monitoring system for the building construction process, etc. In one example, the intelligent sensor monitoring system 100 for the building construction process can be integrated into the terminal device as a software module and / or a hardware module. For example, the intelligent sensor monitoring system 100 for the building construction process can be a software module in the operating system of the terminal device, or can be an application program developed for the terminal device; of course, the intelligent sensor monitoring system 100 for the building construction process can also be one of the numerous hardware modules of the terminal device.

[0082] Alternatively, in another example, the intelligent sensor monitoring system 100 for the building construction process and the terminal device can also be separate devices, and the intelligent sensor monitoring system 100 for the building construction process can be connected to the terminal device through a wired and / or wireless network and transmit interaction information according to a predefined data format.

[0083] Figure 4 It is a flowchart of the intelligent sensor monitoring method for the building construction process according to the embodiments of the present application. As Figure 4As shown, the intelligent sensor monitoring method for the building construction process according to an embodiment of the present application includes the steps of: S110, obtaining the pouring temperature values, pouring pressure values, and pouring humidity values at a plurality of predetermined time points including the current time point; S120, extracting a concrete pouring parameter input matrix from the pouring temperature values, pouring pressure values, and pouring humidity values at the plurality of predetermined time points; S130, passing the concrete pouring parameter input matrix through a first convolutional neural network model with spatial attention to obtain a concrete pouring parameter feature vector, and performing global regression bitwise variational entropy optimization on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector; S140, passing the optimized concrete pouring parameter feature vector through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the concrete pouring process.

[0084] Here, those skilled in the art can understand that the specific operations of each step in the above intelligent sensor monitoring method for the building construction process have been described in detail above with reference to Figures 1 to 3 the description of the intelligent sensor monitoring system for the building construction process, and therefore, the repeated description thereof will be omitted.

[0085] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present invention.

[0086] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the above-described systems, devices, and units can refer to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0087] In several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be in an electrical, mechanical, or other form.

[0088] The unit described as a separation component may or may not be physically separated. The component shown as a unit may or may not be a physical unit, that is, it may be located in one place or may be distributed across multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0089] In addition, in each embodiment of the present invention, each functional unit can be integrated in a processing unit, can also exist physically separately for each unit, or two or more units can be integrated in one unit.

[0090] If the above function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or part of this technical solution can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0091] The above is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present invention can easily think of changes or substitutions, which should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

1. An intelligent sensor monitoring system for the building construction process, characterized in that, Including: A building construction data acquisition module, configured to obtain pouring temperature values, pouring pressure values, and pouring humidity values at multiple predetermined time points including the current time point; A building construction data processing module, configured to extract a concrete pouring parameter input matrix from the pouring temperature values, pouring pressure values, and pouring humidity values at the multiple predetermined time points; A building construction data encoding module, configured to obtain a concrete pouring parameter feature vector by passing the concrete pouring parameter input matrix through a first convolutional neural network model with spatial attention, and perform global regression bitwise variational entropy optimization on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector; A building construction data control module, configured to obtain a classification result by passing the optimized concrete pouring parameter feature vector through a classifier, where the classification result is used to indicate whether there is an abnormality in the concrete pouring process.

2. The intelligent sensor monitoring system for the building construction process according to claim 1, wherein The building construction data processing module includes: A temperature processing unit, configured to arrange the pouring temperature values at the multiple predetermined time points into a vector according to the time dimension and then obtain a pouring temperature feature vector through feature extraction; A pressure processing unit, configured to arrange the pouring pressure values at the multiple predetermined time points into a vector according to the time dimension and then obtain a pouring pressure feature vector through feature extraction; A humidity processing unit, configured to arrange the pouring humidity values at the multiple predetermined time points into a vector according to the time dimension and then obtain a pouring humidity feature vector through feature extraction; A fusion parameter unit, configured to arrange the pouring temperature feature vector, the pouring pressure feature vector, and the pouring humidity feature vector according to the parameter dimension to obtain the concrete pouring parameter input matrix.

3. The intelligent sensor monitoring system for the building construction process according to claim 2, wherein The temperature processing unit includes: A temperature arrangement vector sub-unit, configured to arrange the pouring temperature values at the multiple predetermined time points into a pouring temperature input vector according to the time dimension; A temperature feature extraction sub-unit, configured to obtain the pouring temperature feature vector by passing the pouring temperature input vector through a temperature extractor of a multi-scale neighborhood feature extraction module.

4. The intelligent sensor monitoring system for the building construction process according to claim 3, wherein, The temperature feature extraction sub-unit includes: A first-scale temperature encoding second-level sub-unit, configured to perform one-dimensional convolutional encoding on the pouring temperature input vector using a first convolutional layer of the temperature extractor of the multi-scale neighborhood feature extraction module with a one-dimensional convolutional kernel of a first scale to obtain a first-scale pouring temperature feature vector, where the first convolutional layer has a first one-dimensional convolutional kernel of a first length; A second-scale temperature encoding second-level sub-unit, configured to perform one-dimensional convolutional encoding on the pouring temperature input vector using a second convolutional layer of the temperature extractor of the multi-scale neighborhood feature extraction module with a one-dimensional convolutional kernel of a second scale to obtain a second-scale pouring temperature feature vector, where the second convolutional layer has a second one-dimensional convolutional kernel of a second length, and the first length is different from the second length; A multi-scale temperature cascade second-level sub-unit, configured to cascade the first-scale pouring temperature feature vector and the second-scale pouring temperature feature vector to obtain the pouring temperature feature vector.

5. The intelligent sensor monitoring system for the building construction process according to claim 4, characterized in that, The first-scale temperature encoding secondary unit is configured to: perform one-dimensional convolutional encoding on the pouring temperature input vector using the first convolutional layer of the temperature extractor of the multi-scale neighborhood feature extraction module according to the following first-scale convolutional formula to obtain a first-scale pouring temperature feature vector; Wherein, the first-scale convolutional formula is: Where a is the width of the first convolutional kernel in the x direction, F(a) is the first convolutional kernel parameter vector, G(x - a) is the local vector matrix for operation with the first convolutional kernel function, w is the size of the first convolutional kernel, X represents the pouring temperature input vector, and Cov1(X) represents the first-scale pouring temperature feature vector; The second-scale feature extraction unit is configured to: perform one-dimensional convolutional encoding on the pouring temperature input vector using the second convolutional layer of the temperature extractor of the multi-scale neighborhood feature extraction module according to the following second-scale convolutional formula to obtain the second-scale pouring temperature feature vector; Wherein, the second-scale convolutional formula is: Where b is the width of the second convolutional kernel in the x direction, F(b) is the second convolutional kernel parameter vector, G(x - b) is the local vector matrix for operation with the second convolutional kernel function, m is the size of the second-dimensional convolutional kernel, X represents the pouring temperature input vector, and Cov2(X) represents the second-scale pouring temperature feature vector.

6. The intelligent sensor monitoring system for the building construction process according to claim 5, characterized in that The pressure processing unit includes: The pressure arrangement vector sub-unit is configured to arrange the pouring pressure values at the plurality of predetermined time points into a pouring pressure input vector according to the time dimension; The pressure feature extraction sub-unit is configured to obtain the pouring pressure feature vector by passing the pouring pressure input vector through the pressure extractor of the multi-scale neighborhood feature extraction module.

7. The intelligent sensor monitoring system for the building construction process according to claim 6, characterized in that, The humidity processing unit includes: The humidity arrangement vector sub-unit is configured to arrange the pouring humidity values at the plurality of predetermined time points into a pouring humidity input vector according to the time dimension; The humidity feature extraction sub-unit is configured to obtain the pouring humidity feature vector by passing the pouring humidity input vector through the humidity extractor of the multi-scale neighborhood feature extraction module.

8. The intelligent sensor monitoring system for the building construction process according to claim 7, characterized in that The building construction data encoding module includes: The attention unit is configured to obtain a concrete pouring parameter feature vector by passing the concrete pouring parameter input matrix through the first convolutional neural network model of spatial attention; The optimization unit is configured to perform per-bit variational entropy optimization of global regression on the concrete pouring parameter feature vector to obtain an optimized concrete pouring parameter feature vector.

9. The intelligent sensor monitoring system for the construction process according to claim 8, characterized in that, The optimization unit is configured to: Calculate the autocovariance matrix of the concrete pouring parameter feature vector, and perform principal component dimensionality reduction on the autocovariance matrix of the concrete pouring parameter feature vector to obtain a set of concrete pouring parameter feature basic factor encoding vectors; Input the set of concrete pouring parameter feature basic factor encoding vectors into the sequence encoder based on the forward LSTM model to obtain a set of concrete pouring parameter feature basic factor context-associated encoding vectors; Calculate the bitwise variational entropy between each corresponding concrete pouring parameter characteristic basis factor context-associated coding vector and concrete pouring parameter characteristic basis factor coding vector in the set of concrete pouring parameter characteristic basis factor context-associated coding vectors and the set of concrete pouring parameter characteristic basis factor coding vectors to obtain a set of bitwise variational entropies; Perform a weighting process on the set of bitwise variational entropies to obtain a set of bitwise variational entropy optimization factors; Based on the set of bitwise variational entropy optimization factors, fuse the set of concrete pouring parameter characteristic basis factor coding vectors to obtain an optimized concrete pouring parameter characteristic vector.

10. An intelligent sensor monitoring method for the construction process, characterized in that, Including: Obtain the pouring temperature values, pouring pressure values, and pouring humidity values at multiple predetermined time points including the current time point; Extract a concrete pouring parameter input matrix from the pouring temperature values, pouring pressure values, and pouring humidity values at the multiple predetermined time points; Pass the concrete pouring parameter input matrix through a first convolutional neural network model with spatial attention to obtain a concrete pouring parameter characteristic vector, and perform bitwise variational entropy optimization of global regression on the concrete pouring parameter characteristic vector to obtain an optimized concrete pouring parameter characteristic vector; Pass the optimized concrete pouring parameter characteristic vector through a classifier to obtain a classification result, and the classification result is used to indicate whether there is an abnormality in the concrete pouring process.

Images