Deep foundation pit construction quality safety supervision system and method

Abstract

The invention relates to the field of safety supervision, and particularly discloses a deep foundation pit construction quality safety supervision system and method. Firstly, a pithead image, collected by a camera, of a deep foundation pit construction site, deep foundation pit side wall soil layer inclination displacement measurement text data collected by an inclinometer and underground water level values, collected by an underground water level monitoring device, at a plurality of preset time points are obtained, then feature extraction and correlation analysis are conducted on the three parts through the deep learning technology, and the deep foundation pit side wall soil layer inclination displacement measurement text data and the underground water level values are obtained. And finally, a classification result is obtained through a classifier to judge whether a deep foundation pit construction quality safety early warning is sent out or not, so that potential safety hazards are found in time, early warning is carried out in advance, and smooth proceeding and safety of engineering construction are guaranteed.

Description

Technical Field

[0001] This application relates to the field of safety supervision, and more specifically, to a deep foundation pit construction quality and safety supervision system and method. Background Art

[0002] Deep foundation pit construction refers to the deep excavation operation carried out in projects such as buildings, subways, and bridges for underground space excavation or infrastructure construction. This kind of construction needs to consider various factors such as the groundwater level, soil layer stability, and the influence of surrounding buildings. Usually, a support structure is used to ensure construction safety, and real-time monitoring and early warning are required to ensure construction quality and safety.

[0003] However, due to the complex construction procedures, many potential safety hazards, and numerous uncertain factors in deep foundation pit construction, once a foundation pit safety accident occurs, it will bring huge economic losses and casualties, and may lead to project delays and adverse social impacts. The current safety monitoring technology mainly relies on on-site camera monitoring. Therefore, there is an urgent need to develop an intelligent management and control system to provide a more effective solution for the construction safety problems of deep foundation pit projects.

[0004] Therefore, a deep foundation pit construction quality and safety supervision system and method are expected. Summary of the Invention

[0005] To solve the above technical problems, this application is proposed. Embodiments of this application provide a deep foundation pit construction quality and safety supervision system and method. First, it obtains the pit mouth image of the deep foundation pit construction site collected by a camera, the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by an inclinometer, and the groundwater level values at multiple predetermined time points collected by underground water level monitoring equipment. Then, using deep learning technology, feature extraction and correlation analysis are performed on the three, and finally, a classification result is obtained through a classifier to determine whether to issue a deep foundation pit construction quality and safety early warning, so as to timely discover potential safety hazards, give an early warning, and thus ensure the smooth progress and safety of the project construction.

[0006] According to one aspect of this application, a deep foundation pit construction quality and safety supervision system is provided, which includes:

[0007] A deep foundation pit construction safety data acquisition module, configured to obtain the pit mouth image of the deep foundation pit construction site collected by a camera, the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by an inclinometer, and the groundwater level values at multiple predetermined time points collected by underground water level monitoring equipment;

[0008] A deep foundation pit construction safety data extraction module is used to extract a deep foundation pit soil layer safety correlation feature matrix and a groundwater level feature vector from the pit mouth images of the deep foundation pit construction site collected by the camera, the text data of the inclined displacement measurement of the deep foundation pit side wall soil layer collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device;

[0009] A deep foundation pit safety warning generation module is used to judge whether to issue a deep foundation pit construction quality and safety warning based on the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector.

[0010] According to another aspect of the present application, a deep foundation pit construction quality and safety supervision method is provided, which includes:

[0011] Obtain the pit mouth images of the deep foundation pit construction site collected by the camera, the text data of the inclined displacement measurement of the deep foundation pit side wall soil layer collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device;

[0012] Extract a deep foundation pit soil layer safety correlation feature matrix and a groundwater level feature vector from the pit mouth images of the deep foundation pit construction site collected by the camera, the text data of the inclined displacement measurement of the deep foundation pit side wall soil layer collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device;

[0013] Judge whether to issue a deep foundation pit construction quality and safety warning based on the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector.

[0014] Compared with the prior art, a deep foundation pit construction quality and safety supervision system and method provided by the present application first obtain the pit mouth images of the deep foundation pit construction site collected by the camera, the text data of the inclined displacement measurement of the deep foundation pit side wall soil layer collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device, then use deep learning technology to perform feature extraction and correlation analysis on the three, and finally obtain a classification result through a classifier to judge whether to issue a deep foundation pit construction quality and safety warning, so as to timely discover potential safety hazards and give early warnings, thereby ensuring the smooth progress and safety of the project construction. Description of the Drawings

[0015] 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.

[0016] Figure 1 It is a block diagram of a deep foundation pit construction quality and safety supervision system according to an embodiment of the present application.

[0017] Figure 2 It is a block diagram of a deep foundation pit construction safety data extraction module in the deep foundation pit construction quality and safety supervision system according to an embodiment of the present application.

[0018] Figure 3 It is a block diagram of a deep foundation pit safety warning generation module in the deep foundation pit construction quality and safety supervision system according to an embodiment of the present application.

[0019] Figure 4 It is a flowchart of a deep foundation pit construction quality and safety supervision method according to an embodiment of the present application.

[0020] Figure 5 It is a block diagram of an electronic device according to an embodiment of the present application. Detailed implementation manners

[0021] Next, exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. It should be understood that the present application is not limited by the exemplary embodiments described herein.

[0022] Exemplary system

[0023] Figure 1 It is a block diagram of a deep foundation pit construction quality and safety supervision system according to an embodiment of the present application. As Figure 1 shown, the deep foundation pit construction quality and safety supervision system 100 according to an embodiment of the present application includes: a deep foundation pit construction safety data acquisition module 110, configured to acquire the pit mouth image of the deep foundation pit construction site collected by a camera, the deep foundation pit sidewall soil layer inclination displacement measurement text data collected by an inclinometer, and the groundwater level values at multiple predetermined time points collected by an underground water level monitoring device; a deep foundation pit construction safety data extraction module 120, configured to extract a deep foundation pit soil layer safety correlation feature matrix and a groundwater level feature vector from the pit mouth image of the deep foundation pit construction site collected by the camera, the deep foundation pit sidewall soil layer inclination displacement measurement text data collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the underground water level monitoring device; a deep foundation pit safety warning generation module 130, configured to determine whether to issue a deep foundation pit construction quality and safety warning based on the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector.

[0024] In the above-mentioned deep foundation pit construction quality and safety supervision system 100, the deep foundation pit construction safety data acquisition module 110 is used to obtain the pit mouth image of the deep foundation pit construction site collected by the camera, the deep foundation pit side wall soil layer tilt displacement measurement text data collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring equipment. It should be understood that a deep foundation pit refers to a deep tunnel excavated for the construction of high-rise buildings or underground structures in urban construction. Deep foundation pit construction needs to consider factors such as geological conditions, support structures, and groundwater levels. There are many challenges and risks in the construction process, such as landslides and groundwater leakage. Therefore, deep foundation pit projects require rigorous design planning, precise construction technology, and effective safety monitoring to ensure that the project is carried out safely and stably. However, deep foundation pit construction involves complex processes and multiple safety hazards, and there are many uncertain factors. Once a foundation pit safety accident occurs, it will lead to huge economic losses and casualties, and may cause construction delays and negative social impacts. At present, safety monitoring mainly relies on on-site camera monitoring, but this method has limitations. Therefore, in the technical solution of the present application, by acquiring the pit mouth image of the deep foundation pit construction site collected by the camera, the deep foundation pit side wall soil layer tilt displacement measurement text data collected by the inclinometer and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring equipment, combined with deep learning technology, it is judged whether to issue a deep foundation pit construction quality and safety warning. Specifically, the pit mouth image collected by the camera and the soil layer tilt displacement data collected by the inclinometer can monitor the situation of the deep foundation pit construction site in real time and help identify possible safety hazards and problems. Through the water level values collected by the groundwater level monitoring equipment, the impact of groundwater on deep foundation pit construction can be evaluated, abnormal changes in groundwater levels can be discovered in time, and possible safety risks can be warned. These data can be systematically analyzed and processed, features can be extracted and warnings can be generated, providing decision support for construction management personnel, helping to take timely measures to avoid accidents, thereby timely discovering potential problems and taking preventive measures to ensure a smooth construction process and reduce safety risks.

[0025] In the above deep foundation pit construction quality and safety supervision system 100, the deep foundation pit construction safety data extraction module 120 is used to extract the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector from the pit mouth image of the deep foundation pit construction site collected by the camera, the deep foundation pit side wall soil layer inclination displacement measurement text data collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device. It should be understood that by comprehensively analyzing the information from different data sources, the situation of the deep foundation pit construction site can be more comprehensively understood, including the condition of the pit mouth soil layer, the inclination situation, and the groundwater level change situation. Among them, the deep foundation pit side wall soil layer inclination displacement measurement text data collected by the inclinometer includes the time stamp, the measurement point position information, the inclination displacement value, and the possible environmental conditions (i.e., temperature, humidity). These data can help engineers and researchers monitor the soil layer deformation situation, evaluate the foundation pit stability, and formulate corresponding safety measures and engineering design plans. Extracting features from these data can help the system identify key safety features and patterns, so as to better judge the safety state of the deep foundation pit soil layer and the impact of groundwater on the construction, discover potential safety risks and give early warnings to ensure the construction quality and safety.

[0026] Figure 2 The block diagram of the deep foundation pit construction safety data extraction module in the deep foundation pit construction quality and safety supervision system according to the embodiment of the present application. As Figure 2As shown, in a specific embodiment of the present application, the deep foundation pit construction safety data extraction module 120 includes: a deep foundation pit opening feature extraction unit 121, configured to extract features from the opening images of the deep foundation pit construction site collected by the camera to obtain the opening feature vectors of the deep foundation pit construction site; a deep foundation pit side wall inclination semantic encoding unit 122, configured to perform semantic encoding on the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by the inclinometer to obtain the semantic understanding feature vectors of the inclined displacement of the soil layer on the side wall; a deep foundation pit safety feature association unit 123, configured to associate the opening feature vectors of the deep foundation pit construction site and the semantic understanding feature vectors of the inclined displacement of the soil layer on the side wall to obtain the safety association feature matrix of the deep foundation pit soil layer; and a groundwater level feature extraction unit 124, configured to extract features from the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device to obtain the groundwater level feature vectors. It should be understood that the opening images of the deep foundation pit construction site may contain a large amount of details and information, and feature extraction can convert these complex image data into more concise and meaningful feature vectors, facilitating subsequent processing and analysis. By extracting the feature vectors of the opening images, it is possible to help identify different types of soil layers, rock layers or other geological features, and then conduct classification and analysis to better understand the underground conditions. The feature vectors after feature extraction can be used to monitor and detect potential abnormal conditions, such as abnormal soil layer displacement or abnormal geological structure, which helps to timely warn of possible safety risks.

[0027] Furthermore, by performing semantic encoding on the inclined displacement data of the soil layer on the side wall, it is possible to better understand the meaning and association behind the data, and help identify the patterns and regularities of the inclined displacement. Semantic encoding can help extract important features from the data, convert complex text data into feature vectors with semantic information, and contribute to subsequent analysis and processing. By performing semantic encoding on the inclined displacement data, it is possible to more accurately detect abnormal conditions, such as sudden changes in the inclined displacement of the soil layer, so as to early warn of possible safety risks. Moreover, the semantic understanding feature vectors can help to associate and analyze the inclined displacement data of the soil layer on the side wall with other data sources (such as opening images, groundwater level data), providing more information for comprehensive analysis. Based on the feature vectors of semantic understanding, it is possible to provide a more intuitive and easily understandable data representation form for managers, helping them to make more accurate decisions and respond to possible problems.

[0028] Furthermore, correlating the pit mouth feature vector and the side wall inclination displacement feature vector can provide a more comprehensive data perspective, which helps in comprehensively analyzing the safety status of the deep foundation pit construction site. By combining the pit mouth features and the side wall inclination displacement data, the surface and underground conditions can be considered simultaneously, thus providing a more comprehensive understanding of the soil layer changes and potential safety hazards. By analyzing the correlated feature matrix, it is easier to detect the correlation anomalies between the pit mouth features and the side wall inclination displacement, helping to warn of potential safety risks. By comprehensively considering the pit mouth features and the side wall inclination displacement data, the safety status of the deep foundation pit soil layer can be more accurately evaluated, providing a scientific basis and guidance for the construction process.

[0029] In particular, by extracting the feature vector of the groundwater level, it can help analyze the changing trend of the groundwater level, including rising, falling or remaining stable, to help detect abnormal changes in the groundwater level data, such as sudden water level fluctuations or outliers, which is helpful for timely discovering potential problems. Through feature vector extraction, repetitive patterns or periodic changes in the groundwater level data can be identified, helping to understand the operation law of the groundwater system. The obtained groundwater level feature vector can compress and simplify the original groundwater level data, reducing the data dimension for easier storage and analysis.

[0030] In a specific embodiment of the present application, the deep foundation pit mouth feature extraction unit 121 includes: passing the pit mouth image of the deep foundation pit construction site collected by the camera through a deep foundation pit construction pit mouth noise reduction model based on a convolutional neural network to obtain an enhanced pit mouth image of the deep foundation pit construction site after noise reduction; passing the enhanced pit mouth image of the deep foundation pit construction site after noise reduction through a deep foundation pit construction site pit mouth feature extractor based on a spatial attention mechanism to obtain the deep foundation pit construction site pit mouth feature vector. It should be understood that noise reduction processing helps to reduce blurring and distortion in the image, improve the clarity of details in the image, and thus improve the recognition accuracy of various elements in the deep foundation pit construction site. Through noise reduction processing, noise and interference in the image can be removed, improving the clarity and quality of the image to better suit subsequent image processing and analysis tasks, such as object detection, deep learning, etc., improving the accuracy and efficiency of automated analysis.

[0031] Furthermore, the feature extractor based on the spatial attention mechanism can effectively identify the key features in the image, helping to capture important information about the pit mouth of the deep foundation pit construction site, such as structure, shape, texture, etc. Through the spatial attention mechanism, different regions of the image can be weighted, highlighting the features of important regions, thereby enhancing the feature expression ability and making the features of the pit mouth of the deep foundation pit construction site more accurate and rich. Using the spatial attention mechanism for feature extraction can improve the performance and generalization ability of the model, making subsequent tasks such as pit mouth analysis, identification, or monitoring of the deep foundation pit construction site more accurate and reliable. Specifically, use the convolutional encoding part of the pit mouth feature extractor of the deep foundation pit construction site based on the spatial attention mechanism to perform depth convolution encoding on the enhanced image of the pit mouth of the deep foundation pit construction site after noise reduction to obtain an initial convolutional feature map; input the initial convolutional feature map into the spatial attention part of the pit mouth feature extractor of the deep foundation pit construction site based on the spatial attention mechanism to obtain a spatial attention map; pass the spatial attention map through the Softmax activation function to obtain a spatial attention feature map; calculate the element-wise multiplication of the spatial attention feature map and the initial convolutional feature map to obtain the pit mouth feature vector of the deep foundation pit construction site. More specifically, perform average pooling and maximum pooling along the channel dimension on the initial convolutional feature map respectively to obtain an average feature matrix and a maximum feature matrix; concatenate and adjust the channels of the average feature matrix and the maximum feature matrix to obtain a channel feature matrix; use the convolutional layer of the spatial attention feature map to perform convolutional encoding on the channel feature matrix to obtain a spatial attention map.

[0032] In a specific embodiment of the present application, the deep foundation pit sidewall inclination semantic encoding unit 122 includes: passing the text data of the soil layer inclination displacement measurement of the deep foundation pit sidewall collected by the inclinometer through the sidewall soil layer inclination displacement measurement word embedding model to obtain a plurality of sidewall soil layer inclination displacement measurement data items; passing the plurality of sidewall soil layer inclination displacement measurement data items through the sidewall soil layer inclination displacement context semantic encoder to obtain the sidewall soil layer inclination displacement semantic understanding feature vector. It should be understood that the word embedding model can map the words in the text data into a continuous vector space, thereby better capturing the semantic relationships between words, which helps to more accurately represent the meaning in the text data of the soil layer inclination displacement measurement of the deep foundation pit sidewall. Through the word embedding model, the high-dimensional sparse text data can be converted into low-dimensional dense word vector representations, reducing the dimension of the data and improving the data expression ability and processing efficiency. The word embedding model can learn the semantic information of words from large-scale text data and apply this learned knowledge to the text data collected by the inclinometer, enriching the soil layer inclination displacement measurement data of the sidewall and making the data more comprehensive and rich. Through the word embedding model, the relevance between words in the text data can be discovered, helping to understand the internal connections in the text data of the soil layer inclination displacement measurement of the deep foundation pit sidewall, so as to obtain more data items related to the soil layer inclination displacement of the sidewall, which is helpful to better understand and monitor the soil layer inclination in the deep foundation pit project and improve the engineering safety and efficiency.

[0033] Furthermore, there may be a certain context relationship between the sidewall soil layer inclination displacement measurement data items. Through the context semantic encoder, the relevance between the data items and the underlying semantic information can be better understood, which helps to extract richer and more accurate features. Among them, the context semantic encoder can integrate multiple data items into a unified semantic representation, better capturing the meaning and features of the sidewall soil layer inclination displacement data, thereby improving the data expression ability and understanding depth. Through the context semantic encoder, the important features of the sidewall soil layer inclination displacement data can be effectively extracted, and multiple data items can be mapped into a low-dimensional feature space, reducing the dimension of the data, improving the compactness and interpretability of the data, and facilitating subsequent processing and applications. The obtained sidewall soil layer inclination displacement semantic understanding feature vector can better capture the semantic information of the data, helping to deeply understand the change trend and law of the sidewall soil layer inclination displacement, and providing a more accurate reference basis for engineering monitoring and decision-making. Specifically, use the embedding layer of the sidewall soil layer inclination displacement context semantic encoder to map each data item in the plurality of sidewall soil layer inclination displacement measurement data items into a word embedding vector to obtain a sequence of word embedding vectors; use the Transformer-based Bert model of the sidewall soil layer inclination displacement context semantic encoder to perform global context semantic encoding on the sequence of word embedding vectors to obtain a plurality of feature vectors; and, cascade the plurality of feature vectors to obtain the sidewall soil layer inclination displacement semantic understanding feature vector.

[0034] In a specific embodiment of the present application, the groundwater level feature extraction unit 124 includes: arranging the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device in the time dimension to obtain a groundwater level input vector; and passing the groundwater level input vector through a groundwater level time series encoder to obtain the groundwater level feature vector. It should be understood that groundwater level data usually has obvious time series characteristics. Arranging the data in the time dimension can better retain the time series relationship between data points and help capture the change law of the data over time. By arranging the groundwater level data in the time dimension, features in the time dimension, such as seasonal changes and trend changes, can be better extracted, which helps to reveal the law and trend of groundwater level changes.

[0035] Furthermore, the groundwater level usually has time series characteristics during the construction of deep foundation pits. Through the time series encoder, the time series information of the groundwater level, including change trends, periodicity, etc., can be effectively extracted, which helps to more comprehensively understand the change law of the groundwater level. The groundwater level time series encoder can convert complex time series data into representative feature vectors, extract the key features of the groundwater level change, and help to conduct a more in-depth analysis and understanding of the groundwater level. Among them, the time series encoder can perform dimensionality reduction processing on the original groundwater level data, extract the most representative features, reduce the dimension of the data, help to simplify the data structure, and improve the efficiency of data processing and analysis. Specifically, use the fully connected layer of the groundwater level time series encoder to perform fully connected encoding on the groundwater level input vector to extract the high-dimensional hidden features of the feature values at each position in the groundwater level input vector; and use the one-dimensional convolutional layer of the groundwater level time series encoder to perform one-dimensional encoding on the groundwater level input vector to extract the high-dimensional hidden correlation features of the correlation between the feature values at each position in the groundwater level input vector.

[0036] In the above deep foundation pit construction quality and safety supervision system 100, the deep foundation pit safety warning generation module 130 is used to determine whether to issue a deep foundation pit construction quality and safety warning based on the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector. It should be understood that by combining the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector, comprehensive analysis can be carried out, considering both the soil layer mechanical properties and the influence of groundwater, to more comprehensively evaluate the safety status of the deep foundation pit. Considering that the construction safety of deep foundation pits is affected by various factors, including soil layer stability, groundwater level changes, etc., considering multi-dimensional information comprehensively can more accurately judge the construction quality and safety status. By establishing a warning model based on the soil layer safety correlation features and groundwater level features, real-time monitoring and warning of the deep foundation pit construction quality and safety can be realized, potential safety risks can be discovered in time, and corresponding measures can be taken to avoid accidents. Using a data-driven method and combining the characteristic information of the soil layer and the groundwater level, a more accurate model can be established to judge the deep foundation pit construction quality and safety status, improving the accuracy and reliability of the warning. Based on the analysis results of the soil layer and groundwater level features, scientific basis can be provided for relevant decisions, guiding the formulation of safety management and control measures for deep foundation pit construction.

[0037] Figure 3 It is a block diagram of the deep foundation pit safety warning generation module in the deep foundation pit construction quality and safety supervision system according to the embodiment of the present application. As Figure 3 shown, in a specific embodiment of the present application, the deep foundation pit safety warning generation module 130 includes: a deep foundation pit safety data feature fusion unit 131, which is used to perform bit-by-bit fusion entropy optimization on the deep foundation pit soil layer safety correlation feature vector and the groundwater level feature vector to obtain a construction quality and safety judgment feature vector; a construction quality and safety judgment generation unit 132, which is used to pass the construction quality and safety judgment feature vector through a classifier to obtain a classification result, and the classification result is used to determine whether to issue a deep foundation pit construction quality and safety warning. It should be understood that the construction quality and safety of deep foundation pits are affected by various factors such as soil layer stability and groundwater level changes. Fusing the features of both can more comprehensively consider the safety status of deep foundation pits and avoid the limitations of single-factor evaluation. By fusing the soil layer safety correlation features and the groundwater level features, the representation ability of the feature vector can be enhanced, better describing the complex situation of the deep foundation pit construction environment and improving the judgment accuracy. Considering the features of the soil layer and the groundwater level comprehensively can realize the comprehensive evaluation of the deep foundation pit construction quality and safety, better discover potential risks, take preventive measures in time, and ensure construction safety.

[0038] Specifically, performing bit-by-bit fusion entropy optimization on the deep foundation pit soil layer safety correlation feature vector and the groundwater level feature vector to obtain a construction quality safety judgment feature vector, including: First, calculating the autocovariance matrix of the deep foundation pit soil layer safety correlation feature vector, and performing dominant dimension compression on the autocovariance matrix of the deep foundation pit soil layer safety correlation feature vector to obtain a set of deep foundation pit soil layer safety correlation feature basic factor coding vectors, which is expressed by the formula:

[0039]

[0040] where V represents the deep foundation pit soil layer safety correlation feature vector, T represents the transpose of the vector, M z represents the autocovariance matrix, U represents the set of deep foundation pit soil layer safety correlation feature basic factor coding vectors, v1, v2, v m respectively represent the first, second, and m-th deep foundation pit soil layer safety correlation feature basic factor coding vectors, Λ represents the deep foundation pit soil layer safety correlation feature diagonal matrix after dominant dimension compression, and λ1, λ m respectively represent the first and m-th eigenvalues on the diagonal of the deep foundation pit soil layer safety correlation feature diagonal matrix.

[0041] That is, by constructing the autocovariance matrix, the statistical dependence strength between each feature dimension under different time series can be quantified. Especially for data with spatio-temporal continuity characteristics such as soil layer displacement, its autocorrelation function can effectively reveal the displacement change law between adjacent measuring points or time periods. Specifically, when performing PCA dimensionality reduction on the autocovariance matrix, by finding the principal components in the direction of the maximum variance of the data, not only the decoupled reconstruction of the high-dimensional feature space is realized, but also the secondary variation information generated by sensor noise or environmental interference can be filtered out, enabling the system to focus on the key patterns indicating the precursor of soil layer instability. This processing method not only retains the core criteria affecting the safety of the support structure (such as the sudden change feature of the displacement rate), but also maps the multi-dimensional time series data to the low-dimensional manifold space, significantly improving the decision-making efficiency of the subsequent classifier and the ability to control the false alarm rate.

[0042] Second, inputting the set of deep foundation pit soil layer safety correlation feature basic factor coding vectors into a sequence encoder based on the forward LSTM model to obtain a set of deep foundation pit soil layer safety correlation feature basic factor context correlation coding vectors, which is expressed by the formula:

[0043] F = LSTM([v1, v2,..., v m ) = [s1, s2,..., s m

[0044] where LSTM represents the forward LSTM model, F represents the set of deep foundation pit soil layer safety correlation feature basic factor context correlation coding vectors, and s1, s2, s​m Denote the context - associated coding vectors of the first, second, …, m - th safety - related characteristic basic factors of the deep foundation pit soil layers.

[0045] That is, through LSTM sequence modeling, the model can analyze the arrangement order of the characteristic basic factors from a sequential perspective, capture their non - linear context dependencies, strengthen the associated reasoning ability for the precursors of soil layer instability (such as the characteristic combination law in the displacement acceleration stage), and also transform discrete features into context - associated coding vectors of safety - related characteristic basic factors of deep foundation pit soil layers with temporal semantics, enabling subsequent classifiers to identify complex risk patterns based on the dynamically evolving characteristic context.

[0046] Next, calculate the bit - by - bit fusion entropy between each corresponding context - associated coding vector of the safety - related characteristic basic factors of the deep foundation pit soil layers and the coding vector of the safety - related characteristic basic factors of the deep foundation pit soil layers in the set of context - associated coding vectors of the safety - related characteristic basic factors of the deep foundation pit soil layers to obtain a set of bit - by - bit fusion entropies, which is expressed by the formula:

[0047]

[0048]

[0049] where w represents the bit - by - bit comparison function, v i represents the i - th coding vector of the safety - related characteristic basic factors of the deep foundation pit soil layers, represents the feature value at the j - th position of the i - th coding vector of the safety - related characteristic basic factors of the deep foundation pit soil layers, s i represents the i - th context - associated coding vector of the safety - related characteristic basic factors of the deep foundation pit soil layers, represents the feature value at the j - th position of the i - th context - associated coding vector of the safety - related characteristic basic factors of the deep foundation pit soil layers, ε represents a predetermined threshold, r i represents the i - th bit - by - bit displacement matching feature vector, represents the feature value at the j - th position of the i - th bit - by - bit displacement matching feature vector, L represents the length of the i - th bit - by - bit displacement matching feature vector, e i represents the i - th bit - by - bit fusion entropy.

[0050] That is, by fusing the entropy quantization bit by bit to reveal the change pattern of binary bits, the sensitive fluctuations of features at the information entropy level can be disclosed. Compared with traditional measurement methods such as the Euclidean distance, the bit-by-bit fusion entropy not only enhances the sensitivity to the latent risk signals in feature representation, but also establishes a mapping between the physical meaning of binary bits and the soil layer mechanical behavior, enabling the subsequent optimization strategy to adjust the feature weights based on the quantization gradient of information gain. The generated bit-by-bit fusion entropy provides a quantization basis for the fusion decision of multi-source heterogeneous data with different sensitivities, thereby improving the early warning accuracy and false alarm suppression ability for progressive disasters in deep foundation pits.

[0051] Then, weight the set of the bit-by-bit fusion entropy based on the Softmax function to obtain a set of bit-by-bit fusion entropy adjustment factors, which is expressed by the formula:

[0052] a i =softmax(e i )

[0053] where softmax represents the normalized exponential function, and a i represents the i-th bit-by-bit fusion entropy adjustment factor.

[0054] That is, through the exponential property of the Softmax function, the entropy value difference can be amplified into a weight coefficient in the form of a probability distribution, enabling the basic factor corresponding to the high entropy value to obtain a significantly higher optimization weight, while the low entropy value component is suppressed. This competitive allocation mechanism not only realizes the focused enhancement of key risk-sensitive features, but also avoids problems such as model overfitting or gradient oscillation caused by weight imbalance through the normalization property of the probability distribution. It enables the subsequent feature fusion process to dynamically allocate computing resources based on the quantization ranking of risk contribution degrees.

[0055] Next, based on the set of the bit-by-bit fusion entropy adjustment factors, fuse the set of encoded vectors of the deep foundation pit soil layer safety-related feature basic factors to obtain an optimized deep foundation pit soil layer safety-related feature vector, which is expressed by the formula:

[0056]

[0057] where v f represents the optimized deep foundation pit soil layer safety-related feature vector.

[0058] That is, by amplifying the weight of the optimization factor corresponding to the high entropy value, the retention and enhancement of key risk-sensitive features are strengthened, while redundant or noise features with low entropy values are suppressed. The generated optimized deep foundation pit soil layer safety-related feature vector not only retains the core information structure after PCA and LSTM processing, but also realizes the adaptive reconstruction of the feature space through weight redistribution, thereby improving the sensitivity and generalization ability of the early warning model to progressive risk evolution.

[0059] Finally, the optimized deep foundation pit soil layer safety correlation feature vector and the groundwater level feature vector are weighted and fused to obtain the construction quality safety judgment feature vector. That is, through weighted fusion, a dynamic correlation framework of the optimized deep foundation pit soil layer safety correlation feature vector and the groundwater level feature vector is established, significantly improving the early warning accuracy and decision-making reliability of the system for the multi-factor linkage risk in complex construction environments.

[0060] Furthermore, the classifier can identify the patterns and rules in the feature vector, thereby classifying and predicting the construction quality safety of the deep foundation pit more accurately. The classifier can realize real-time monitoring and analysis of the data during the construction process, timely discover potential safety risks, give early warnings, and help avoid accidents. By classifying the feature vector through the classifier, the accuracy and reliability of the judgment can be improved, the possibility of misjudgment can be reduced, and the effectiveness of the early warning can be ensured.

[0061] In summary, in the embodiment of the present application, first, the pit mouth image of the deep foundation pit construction site collected by the camera, the text data of the inclined displacement measurement of the deep foundation pit side wall soil layer collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device are obtained. Then, using deep learning technology, feature extraction and correlation analysis are performed on the three. Finally, a classification result is obtained through the classifier to determine whether to issue a warning for the construction quality safety of the deep foundation pit, thereby timely discovering potential safety hazards, giving early warnings, and ensuring the smooth progress and safety of the project construction.

[0062] As described above, the deep foundation pit construction quality safety supervision system 100 according to the embodiment of the present application can be implemented in various terminal devices, such as a server deployed with the deep foundation pit construction quality safety supervision algorithm. In one example, the deep foundation pit construction quality safety supervision system 100 can be integrated into the terminal device as a software module and / or a hardware module. For example, the deep foundation pit construction quality safety supervision system 100 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 deep foundation pit construction quality safety supervision system 100 can also be one of the many hardware modules of the terminal device.

[0063] Alternatively, in another example, the deep foundation pit construction quality safety supervision system 100 and the terminal device can also be separate devices, and the deep foundation pit construction quality safety supervision system 100 can be connected to the terminal device through a wired and / or wireless network and transmit interaction information in accordance with a predefined data format.

[0064] Exemplary method

[0065] Figure 4The flowchart of the deep foundation pit construction quality and safety supervision method according to an embodiment of the present application. As Figure 4 shown, the deep foundation pit construction quality and safety supervision method according to an embodiment of the present application includes: S110, obtaining the pit mouth image of the deep foundation pit construction site collected by a camera, the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by an inclinometer, and the groundwater level values at multiple predetermined time points collected by a groundwater level monitoring device; S120, extracting the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector from the pit mouth image of the deep foundation pit construction site collected by the camera, the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device; S130, based on the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector, determining whether to issue a deep foundation pit construction quality and safety warning.

[0066] Here, those skilled in the art can understand that the specific operations of each step in the above deep foundation pit construction quality and safety supervision method have been introduced in detail in the description of the deep foundation pit construction quality and safety supervision system above Figures 1 to 3 and therefore, the repeated description thereof will be omitted.

[0067] Exemplary electronic device

[0068] Next, reference Figure 5 is made to describe the electronic device according to an embodiment of the present application.

[0069] As Figure 5 shown, the electronic device 10 includes an input device 11, an input interface 12, a central processing unit 13, a memory 14, an output interface 15, an output device 16, and a bus 17. Among them, the input interface 12, the central processing unit 13, the memory 14, and the output interface 15 are connected to each other through the bus 17, and the input device 11 and the output device 16 are respectively connected to the bus 17 through the input interface 12 and the output interface 15, and further connected to other components of the electronic device 10.

[0070] Specifically, the input device 11 receives external input information and transmits the input information to the central processing unit 13 through the input interface 12; the central processing unit 13 processes the input information based on the computer-executable instructions stored in the memory 14 to generate output information, temporarily or permanently stores the output information in the memory 14, and then transmits the output information to the output device 16 through the output interface 15; the output device 16 outputs the output information to the outside of the electronic device 10 for user use.

[0071] In one embodiment, Figure 5The illustrated electronic device 10 can be implemented as a network device, which may include: a memory configured to store programs; and a processor configured to run the programs stored in the memory to execute any one of the deep foundation pit construction quality and safety supervision methods described in the above embodiments.

[0072] According to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program tangibly embodied on a machine-readable medium, and the computer program includes program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network, and / or installed from a removable storage medium.

[0073] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations. In the hardware implementation, the division between the functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be executed by several physical components in cooperation. Some or all of the physical components may be implemented as software executed by a processor, such as a central processing unit, a digital signal processor, or a microprocessor, or implemented as hardware, or implemented as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which may include a computer storage medium (or non-transitory medium) and a communication medium (or transitory medium). As is well known to those of ordinary skill in the art, the term computer storage medium includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information, such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer. In addition, as is well known to those of ordinary skill in the art, the communication medium generally includes computer-readable instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and may include any information delivery medium.

[0074] It is understandable that the above embodiments are merely exemplary embodiments adopted to illustrate the principle of the present application. However, the present application is not limited thereto. For those of ordinary skill in the art, various modifications and improvements can be made without departing from the spirit and essence of the present application, and these modifications and improvements are also regarded as the protection scope of the present application.

Claims

1. A deep foundation pit construction quality and safety supervision system, characterized in that Including: A deep foundation pit construction safety data acquisition module, configured to acquire the pit mouth image of the deep foundation pit construction site collected by a camera, the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by an inclinometer, and the groundwater level values at multiple predetermined time points collected by an underground water level monitoring device; A deep foundation pit construction safety data extraction module, configured to extract a deep foundation pit soil layer safety correlation feature matrix and a groundwater level feature vector from the pit mouth image of the deep foundation pit construction site collected by the camera, the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the underground water level monitoring device; A deep foundation pit safety warning generation module, configured to determine whether to issue a quality and safety warning for deep foundation pit construction based on the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector.

2. The deep foundation pit construction quality and safety supervision system according to claim 1, characterized in that, The deep foundation pit construction safety data extraction module includes: A deep foundation pit pit mouth feature extraction unit, configured to extract features from the pit mouth image of the deep foundation pit construction site collected by the camera to obtain a deep foundation pit construction site pit mouth feature vector; A deep foundation pit side wall inclination semantic encoding unit, configured to perform semantic encoding on the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by the inclinometer to obtain a side wall soil layer inclined displacement semantic understanding feature vector; A deep foundation pit safety feature correlation unit, configured to correlate the deep foundation pit construction site pit mouth feature vector and the side wall soil layer inclined displacement semantic understanding feature vector to obtain the deep foundation pit soil layer safety correlation feature matrix; A groundwater level feature extraction unit, configured to extract features from the groundwater level values at multiple predetermined time points collected by the underground water level monitoring device to obtain the groundwater level feature vector.

3. The deep foundation pit construction quality and safety supervision system according to claim 2, characterized in that The deep foundation pit pit mouth feature extraction unit includes: Passing the pit mouth image of the deep foundation pit construction site collected by the camera through a deep foundation pit construction pit mouth noise reduction model based on a convolutional neural network to obtain a denoised and enhanced pit mouth image of the deep foundation pit construction site; Passing the denoised and enhanced pit mouth image of the deep foundation pit construction site through a deep foundation pit construction site pit mouth feature extractor based on a spatial attention mechanism to obtain the deep foundation pit construction site pit mouth feature vector.

4. The deep foundation pit construction quality and safety supervision system according to claim 3, characterized in that, The deep foundation pit construction pit mouth noise reduction model based on a convolutional neural network is a depthwise separable convolutional neural network.

5. The deep foundation pit construction quality and safety supervision system according to claim 4, characterized in that, The deep foundation pit side wall inclination semantic encoding unit includes: Passing the text data of the inclined displacement measurement of the soil layer on the side wall of the deep foundation pit collected by the inclinometer through a side wall soil layer inclined displacement measurement word embedding model to obtain multiple side wall soil layer inclined displacement measurement data items; Passing the multiple side wall soil layer inclined displacement measurement data items through a side wall soil layer inclined displacement context semantic encoder to obtain the side wall soil layer inclined displacement semantic understanding feature vector.

6. The deep foundation pit construction quality and safety supervision system according to claim 5, characterized in that, The groundwater level feature extraction unit includes: Arranging the groundwater level values at multiple predetermined time points collected by the underground water level monitoring device in the time dimension to obtain a groundwater level input vector; Passing the groundwater level input vector through a groundwater level time series encoder to obtain the groundwater level feature vector.

7. The deep foundation pit construction quality and safety supervision system according to claim 6, characterized in that, The deep foundation pit safety warning generation module includes: A deep foundation pit safety data feature fusion unit, configured to perform bit-by-bit fusion entropy optimization on the deep foundation pit soil layer safety correlation feature vector and the groundwater level feature vector to obtain a construction quality safety judgment feature vector; A construction quality safety judgment generation unit, configured to pass the construction quality safety judgment feature vector through a classifier to obtain a classification result, and the classification result is used to judge whether to issue a deep foundation pit construction quality safety warning.

8. The deep foundation pit construction quality and safety supervision system according to claim 7, characterized in that, The deep foundation pit safety data feature fusion unit is configured to: Calculate the autocovariance matrix of the deep foundation pit soil layer safety correlation feature vector, and perform dominant dimension compression on the autocovariance matrix of the deep foundation pit soil layer safety correlation feature vector to obtain a set of deep foundation pit soil layer safety correlation feature basic factor coding vectors; Input the set of deep foundation pit soil layer safety correlation feature basic factor coding vectors into a sequence encoder based on a forward LSTM model to obtain a set of deep foundation pit soil layer safety correlation feature basic factor context correlation coding vectors; Calculate the bit-by-bit fusion entropy between each pair of corresponding deep foundation pit soil layer safety correlation feature basic factor context correlation coding vectors and deep foundation pit soil layer safety correlation feature basic factor coding vectors in the set of deep foundation pit soil layer safety correlation feature basic factor context correlation coding vectors and the set of deep foundation pit soil layer safety correlation feature basic factor coding vectors to obtain a set of bit-by-bit fusion entropy; Perform weight processing based on the Softmax function on the set of bit-by-bit fusion entropy to obtain a set of bit-by-bit fusion entropy adjustment factors; Based on the set of bit-by-bit fusion entropy adjustment factors, fuse the set of deep foundation pit soil layer safety correlation feature basic factor coding vectors to obtain an optimized deep foundation pit soil layer safety correlation feature vector; Perform weighted fusion on the optimized deep foundation pit soil layer safety correlation feature vector and the groundwater level feature vector to obtain the construction quality safety judgment feature vector.

9. A method for supervision of the construction quality and safety of deep foundation pits, characterized in that, It includes: Obtain the pit mouth image of the deep foundation pit construction site collected by a camera, the deep foundation pit side wall soil layer inclination displacement measurement text data collected by an inclinometer, and the groundwater level values at multiple predetermined time points collected by an underground water level monitoring device; Extract a deep foundation pit soil layer safety correlation feature matrix and a groundwater level feature vector from the pit mouth image of the deep foundation pit construction site collected by the camera, the deep foundation pit side wall soil layer inclination displacement measurement text data collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the underground water level monitoring device; Based on the deep foundation pit soil layer safety correlation feature matrix and the groundwater level feature vector, judge whether to issue a deep foundation pit construction quality safety warning.

10. The deep foundation pit construction quality and safety supervision method according to claim 9, characterized in that, Extracting a deep foundation pit soil layer safety correlation feature matrix and a groundwater level feature vector from the pit mouth image of the deep foundation pit construction site collected by the camera, the deep foundation pit side wall soil layer inclination displacement measurement text data collected by the inclinometer, and the groundwater level values at multiple predetermined time points collected by the underground water level monitoring device includes: Extract features from the pit mouth images of the deep foundation pit construction site collected by the camera to obtain the pit mouth feature vector of the deep foundation pit construction site; Perform semantic encoding on the text data of the inclined displacement measurement of the side wall soil layer of the deep foundation pit collected by the inclinometer to obtain the semantic understanding feature vector of the inclined displacement of the side wall soil layer; Correlate the pit mouth feature vector of the deep foundation pit construction site and the semantic understanding feature vector of the inclined displacement of the side wall soil layer to obtain the safety correlation feature matrix of the deep foundation pit soil layer; Extract features from the groundwater level values at multiple predetermined time points collected by the groundwater level monitoring device to obtain the groundwater level feature vector.

Images