Building construction quality detection and evaluation system based on deep neural network

Abstract

The invention discloses a building construction quality detection and evaluation system based on a deep neural network, and relates to the technical field of building construction quality detection, and the system comprises a data collection module which collects building construction site data in real time; the data preprocessing module is used for preprocessing the collected data; the feature extraction module is used for extracting building construction quality related features based on a deep neural network algorithm; the quality evaluation module is used for intelligently evaluating the building construction quality through a deep neural network model; the visualization module is used for displaying the quality evaluation result to a user; and the system interface module provides a data interaction interface and realizes construction quality real-time monitoring and report pushing. According to the method, accurate construction quality detection and evaluation can be realized through multi-dimensional data acquisition and a deep learning algorithm, the construction quality management efficiency can be improved, the influence of human factors can be reduced, the potential safety hazards and quality problems in building construction can be reduced, and the method has a wide application prospect.

Description

Technical Field

[0001] The present invention relates to the technical field of building construction quality detection, and in particular to a building construction quality detection and evaluation system based on a deep neural network. Background Art

[0002] Construction quality inspection and evaluation is one of the key issues in the construction industry. Traditional quality monitoring methods mainly rely on manual inspection and manual experience judgment, which is not only inefficient, but also easily affected by human factors, resulting in inaccurate and in-time judgment of construction quality. With the development of intelligent and digital technologies, especially the application of deep learning and artificial intelligence technologies, the inspection and evaluation of construction quality has entered a new era. The current construction quality inspection technology still faces some challenges. For example, traditional monitoring methods are difficult to achieve real-time, accurate and comprehensive. In order to improve the efficiency and accuracy of construction quality inspection, there is an urgent need for an intelligent evaluation system based on multidimensional data and deep neural networks. The system can collect, process and evaluate various data in the construction process in real time and generate construction quality reports.

[0003] In recent years, deep learning technology has been widely used in computer vision, sensor data processing, time series analysis and other fields. Especially in the construction industry, it provides a more efficient and accurate method for construction quality detection and evaluation. Summary of the invention

[0004] In order to solve the above technical problems, a construction quality inspection and evaluation system based on deep neural network is provided. This technical solution solves the above problems.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] The construction quality detection and evaluation system based on deep neural network includes:

[0007] Data acquisition module: real-time acquisition of multi-dimensional data including image data, environmental data and sensor data of the construction site;

[0008] Data preprocessing module: The data preprocessing module is electrically connected to the data acquisition module, and the data preprocessing module is used to clean, denoise and standardize the collected data;

[0009] Feature extraction module: The feature extraction module is electrically connected to the data preprocessing module, and the feature extraction module is used to extract building construction quality related features from multi-dimensional data based on a deep neural network algorithm;

[0010] Quality assessment module: The quality assessment module is electrically connected to the feature extraction module, and is used to intelligently assess the quality of the building construction through a deep neural network model based on the extracted features, and generate a quality report;

[0011] Visualization module: The visualization module is electrically connected to the quality assessment module, and the visualization module is used to display the quality assessment results to the user in the form of graphics and text;

[0012] System interface module: The system interface module is electrically connected to the visualization module. The system interface module is used to provide a data interaction interface with the building management personnel and the external system of the construction party to realize real-time monitoring of construction quality and push of reports.

[0013] Preferably, the data acquisition module specifically includes:

[0014] Image data acquisition unit: uses high-definition cameras to capture image data of the construction site in real time, including visual information of the construction process, material status, and personnel operations;

[0015] Environmental data collection unit: Use temperature and humidity sensors to monitor the temperature and humidity of the construction environment in real time, and collect meteorological data around the site by connecting with the meteorological station system;

[0016] Sensor data acquisition unit: collects structural data of the construction site in real time through acceleration sensors, strain sensors and displacement sensors;

[0017] Equipment operation data collection unit: collects the operation status data of construction equipment in real time through sensors and monitoring equipment, including power, load and operation time;

[0018] Personnel behavior data collection unit: monitors the location and behavior of on-site workers through wearable devices to ensure that construction workers operate in accordance with safety regulations and quality requirements.

[0019] Preferably, the data preprocessing module specifically includes:

[0020] Data cleaning unit: handles errors, omissions and inconsistencies in the collected data, uses mean interpolation algorithm to fill in missing parts of the data, and deletes duplicate records;

[0021] Data denoising unit: Use Kalman filter to remove high-frequency noise. For image data, use median to remove noise and blurred areas in the image, smooth the data based on sliding average, and use signal removal algorithm to remove redundant noise and irrelevant data.

[0022] Data standardization unit: Use Z-Score standardization to transform the data to a mean of 0 and a variance of 1 to make it conform to the normal distribution, and convert data in different units to ensure that the data can be compared on the same scale.

[0023] Preferably, the feature extraction module specifically includes:

[0024] Image data feature extraction unit: extracts visual features of image data of construction sites based on the convolutional neural network formula, inputs image data, extracts low-level features in the image through multiple convolutional layers, then reduces feature dimensions through pooling layers, extracts high-level features through fully connected layers, and forms image features related to construction quality;

[0025] Time series data feature extraction unit: Use a recurrent neural network to extract time series data features in the construction process. Input time series data, extract long-term and short-term dependency features in the time series through the LSTM layer, and use the fully connected layer to convert the extracted time series features into features related to construction quality.

[0026] Feature output unit: outputs the extracted features to the subsequent prediction model.

[0027] Preferably, the extracting of visual features of image data of a construction site based on a convolutional neural network formula specifically includes:

[0028] Among them, the convolutional neural network formula is:

[0029] h l+1 =σ(e l *h l +b l )

[0030] In the formula, h l is the input of the lth layer, w l is the convolution kernel, σ is the activation function, b l is the bias term, and * indicates the convolution operation.

[0031] Preferably, the quality assessment module specifically includes:

[0032] Feature receiving unit: The quality assessment module receives the features related to the construction from the feature extraction module, inputs the features into the neural network for processing, and predicts the construction quality score;

[0033] Model reasoning and quality assessment unit: After reasoning, the deep neural network outputs the predicted results of construction quality, outputs the quality score according to the type of task, and maps the output results of the neural network to the actual quality assessment results;

[0034] Quality report generation unit: Generates and outputs quality assessment reports, including quality scores, assessment indicators and recommendations.

[0035] Preferably, the quality assessment module receives the features related to the construction from the feature extraction module, inputs the features into the neural network for processing, and predicts the construction quality score specifically including:

[0036] Among them, the construction quality score prediction formula is:

[0037]

[0038] In the formula, is the predicted value of construction quality score, f j is the weight matrix of the jth layer, k j-1 is the output of the j-1th layer, b j is the bias vector of the jth layer.

[0039] Preferably, the quality report generating unit specifically includes:

[0040] Based on the output of the deep neural network, the construction quality score is obtained, and the mean square error of the model is calculated to measure the square average of the difference between the model prediction value and the true value;

[0041] Generate specific suggestions for improving construction quality based on the model prediction results and evaluation indicators;

[0042] Based on the prediction results and evaluation indicators of the neural network, determine whether there are potential problems and which parts of the construction quality are at risk;

[0043] Generate targeted suggestions for specific quality issues. For low-scoring areas, further inspection of the construction quality of the area is recommended.

[0044] For unqualified parts, it is recommended to strengthen the construction process and check the quality of construction materials;

[0045] For high-risk areas, it is recommended to add monitoring equipment to strengthen real-time data collection during the construction process.

[0046] Propose actionable improvement measures, such as technical optimization, process improvement, and increased quality inspection, generate quality improvement suggestions, and clarify improvement directions and operational steps;

[0047] Save and export the report.

[0048] Preferably, the step of saving and outputting the report specifically includes:

[0049] Select HTML format to output the quality assessment report, and automatically generate the quality assessment report according to user needs;

[0050] The report includes construction quality score, display quality score, lists the calculated values ​​of various indicators, and generates specific suggestions for construction quality improvement based on the prediction results and evaluation indicators;

[0051] Save the report as a file for cloud storage.

[0052] Preferably, the system interface module specifically includes:

[0053] Construction quality report push unit: pushes the construction quality report, warning information and improvement suggestions generated by the quality assessment module to the building management personnel and the construction party in real time, and automatically pushes the construction quality report at set time intervals;

[0054] Data synchronization unit of external system: synchronize data with external building management system, construction management system and equipment monitoring system;

[0055] Data security and privacy protection unit: Use HTTPS encryption technology to ensure that data is not stolen during transmission, strengthen identity authentication and permission control of external systems, and ensure that only authorized users can access specific data interfaces.

[0056] Compared with the prior art, the present invention has the following beneficial effects:

[0057] The present invention proposes a construction quality inspection and evaluation system based on deep neural networks, which can not only realize accurate construction quality inspection and evaluation through multi-dimensional data collection and deep learning algorithms, but also reduce the influence of human factors while improving the efficiency of construction quality management, and reduce safety hazards and quality problems in construction. The system has significant advantages such as strong real-time performance, high data processing capability, intelligent quality assessment, automated report generation and push, and accurate risk prediction. It provides strong support for the digital and intelligent transformation of construction quality management and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a system framework diagram of the present invention. DETAILED DESCRIPTION

[0059] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0060] Reference Figure 1 As shown in the figure, the construction quality detection and evaluation system based on deep neural network includes:

[0061] Data acquisition module:

[0062] Image data acquisition unit: uses high-definition cameras to capture image data of the construction site in real time, including visual information of the construction process, material status, and personnel operations;

[0063] Environmental data collection unit: Use temperature and humidity sensors to monitor the temperature and humidity of the construction environment in real time, and collect meteorological data around the site by connecting with the meteorological station system;

[0064] Sensor data acquisition unit: collects structural data of the construction site in real time through acceleration sensors, strain sensors and displacement sensors;

[0065] Equipment operation data collection unit: collects the operation status data of construction equipment in real time through sensors and monitoring equipment, including power, load and operation time;

[0066] Personnel behavior data collection unit: monitors the location and behavior of on-site workers through wearable devices to ensure that construction workers operate in accordance with safety regulations and quality requirements.

[0067] Data preprocessing module:

[0068] Data cleaning unit: handles errors, omissions and inconsistencies in the collected data, uses mean interpolation algorithm to fill in missing parts of the data, and deletes duplicate records;

[0069] Data denoising unit: Use Kalman filter to remove high-frequency noise. For image data, use median to remove noise and blurred areas in the image, smooth the data based on sliding average, and use signal removal algorithm to remove redundant noise and irrelevant data.

[0070] Data standardization unit: Use Z-Score standardization to transform the data to a mean of 0 and a variance of 1 to make it conform to the normal distribution, and convert data in different units to ensure that the data can be compared on the same scale.

[0071] Feature extraction module:

[0072] Image data feature extraction unit: extracts visual features of image data of construction sites based on the convolutional neural network formula, inputs image data, extracts low-level features in the image through multiple convolutional layers, then reduces feature dimensions through pooling layers, extracts high-level features through fully connected layers, and forms image features related to construction quality;

[0073] Among them, the convolutional neural network formula is:

[0074] h l+1 =σ(w l *h l +b l )

[0075] In the formula, h l is the input of the lth layer, w l is the convolution kernel, σ is the activation function, b l is the bias term, * indicates the convolution operation;

[0076] Time series data feature extraction unit: Use a recurrent neural network to extract time series data features in the construction process. Input time series data, extract long-term and short-term dependency features in the time series through the LSTM layer, and use the fully connected layer to convert the extracted time series features into features related to construction quality.

[0077] Feature output unit: outputs the extracted features to the subsequent prediction model.

[0078] Quality Assessment Module:

[0079] Feature receiving unit: The quality assessment module receives the features related to the construction from the feature extraction module, inputs the features into the neural network for processing, and predicts the construction quality score;

[0080] Among them, the construction quality score prediction formula is:

[0081]

[0082] In the formula, is the predicted value of construction quality score, f j is the weight matrix of the jth layer, k j-1 is the output of the j-1th layer, b j is the bias vector of the jth layer;

[0083] Model reasoning and quality assessment unit: After reasoning, the deep neural network outputs the predicted results of construction quality, outputs the quality score according to the type of task, and maps the output results of the neural network to the actual quality assessment results;

[0084] Quality report generation unit: Generates and outputs quality assessment reports, including quality scores, evaluation indicators and suggestions. According to the output results of the deep neural network, the construction quality score is obtained, and the mean square error of the model is calculated to measure the square average of the difference between the model prediction value and the true value;

[0085] Generate specific suggestions for improving construction quality based on the model prediction results and evaluation indicators;

[0086] Based on the prediction results and evaluation indicators of the neural network, determine whether there are potential problems and which parts of the construction quality are at risk;

[0087] Generate targeted suggestions for specific quality issues. For low-scoring areas, further inspection of the construction quality of the area is recommended.

[0088] For unqualified parts, it is recommended to strengthen the construction process and check the quality of construction materials;

[0089] For high-risk areas, it is recommended to add monitoring equipment to strengthen real-time data collection during the construction process.

[0090] Propose actionable improvement measures, such as technical optimization, process improvement, and increased quality inspection, generate quality improvement suggestions, and clarify improvement directions and operational steps;

[0091] Save and output the report. Select HTML format to output the quality assessment report. Automatically generate the quality assessment report according to user needs.

[0092] The report includes construction quality score, display quality score, lists the calculated values ​​of various indicators, and generates specific suggestions for construction quality improvement based on the prediction results and evaluation indicators;

[0093] Save the report as a file for cloud storage.

[0094] Visualization module:

[0095] The visualization module is electrically connected to the quality assessment module, and the visualization module is used to display the quality assessment results to the user in the form of graphics and text;

[0096] System interface module:

[0097] Construction quality report push unit: pushes the construction quality report, warning information and improvement suggestions generated by the quality assessment module to the building management personnel and the construction party in real time, and automatically pushes the construction quality report at set time intervals;

[0098] Data synchronization unit of external system: synchronize data with external building management system, construction management system and equipment monitoring system;

[0099] Data security and privacy protection unit: Use HTTPS encryption technology to ensure that data is not stolen during transmission, strengthen identity authentication and permission control of external systems, and ensure that only authorized users can access specific data interfaces.

[0100] In summary, the advantages of the present invention are:

[0101] The system collects images, environment, sensor data and other information of the construction site in real time through the data acquisition module, and fully understands the dynamics of the construction site, making up for the limitations of traditional manual inspection methods. Real-time monitoring of multiple dimensions such as environmental data, equipment operation data, and personnel behavior provides a more complete information basis for construction quality assessment;

[0102] The data preprocessing module can efficiently clean, denoise and standardize data from various sources, ensuring that the data received by the subsequent feature extraction module is accurate and unified, avoiding the impact of noise and inconsistency on quality assessment. It uses advanced Kalman filter and Z-Score standardization technology to effectively improve data quality and provide high-quality data input for the deep neural network model.

[0103] The feature extraction module uses deep learning algorithms based on convolutional neural networks and recurrent neural networks to extract key features related to building construction quality from image data and time series data, ensuring the accuracy and scientificity of quality assessment;

[0104] The quality assessment module can intelligently infer the extracted features through deep neural networks, output accurate construction quality scores, and evaluate and optimize the accuracy of the model by combining mean square error calculation;

[0105] The system can automatically generate construction quality assessment reports, including quality scores, assessment indicators, improvement suggestions, etc., avoiding the inefficiency and subjectivity of manual report writing. The generated reports can not only be displayed in HTML format, but also saved as files and uploaded to the cloud for storage, which is convenient for viewing and management at any time. The system also has a scheduled push function, which can push construction quality reports, early warning information and improvement suggestions to building management personnel and construction parties in real time, ensuring that all parties can understand the quality status of the construction site in a timely manner;

[0106] Through intelligent analysis of deep learning models, the system can accurately identify potential quality problems in the construction process and predict possible quality risks in advance. Based on the quality assessment results, the system automatically generates targeted quality improvement suggestions, such as strengthening construction technology, improving material quality, and adding monitoring equipment, etc., to help construction units improve construction quality and reduce risks in a targeted manner.

[0107] By adopting HTTPS encryption technology and identity authentication mechanism, the system can ensure the security of data during transmission and prevent data leakage or tampering. The system interface module strengthens the authority control of external systems to ensure that only authorized users can access system data and avoid unnecessary security risks.

[0108] The system interface module supports data synchronization with external systems such as the building management system, construction management system, and equipment monitoring system, ensuring that all parties can share the latest data on the construction site and achieve collaborative work. This efficient data interaction method can improve the ability of all parties to monitor construction quality and conduct quality assessment and decision-making in real time.

[0109] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A construction quality detection and evaluation system based on deep neural network, characterized by: include: Data acquisition module: real-time acquisition of multi-dimensional data including image data, environmental data and sensor data of the construction site; Data preprocessing module: The data preprocessing module is electrically connected to the data acquisition module, and the data preprocessing module is used to clean, denoise and standardize the collected data; Feature extraction module: The feature extraction module is electrically connected to the data preprocessing module, and the feature extraction module is used to extract building construction quality related features from multi-dimensional data based on a deep neural network algorithm; Quality assessment module: The quality assessment module is electrically connected to the feature extraction module, and is used to intelligently assess the quality of the building construction through a deep neural network model based on the extracted features, and generate a quality report; Visualization module: The visualization module is electrically connected to the quality assessment module, and the visualization module is used to display the quality assessment results to the user in the form of graphics and text; System interface module: The system interface module is electrically connected to the visualization module. The system interface module is used to provide a data interaction interface with the building management personnel and the external system of the construction party to realize real-time monitoring of construction quality and push of reports.

2. The construction quality detection and evaluation system based on deep neural network according to claim 1 is characterized in that: The data acquisition module specifically includes: Image data acquisition unit: uses high-definition cameras to capture image data of the construction site in real time, including visual information of the construction process, material status, and personnel operations; Environmental data collection unit: Use temperature and humidity sensors to monitor the temperature and humidity of the construction environment in real time, and collect meteorological data around the site by connecting with the meteorological station system; Sensor data acquisition unit: collects structural data of the construction site in real time through acceleration sensors, strain sensors and displacement sensors; Equipment operation data collection unit: collects the operation status data of construction equipment in real time through sensors and monitoring equipment, including power, load and operation time; Personnel behavior data collection unit: monitors the location and behavior of on-site workers through wearable devices to ensure that construction workers operate in accordance with safety regulations and quality requirements.

3. The construction quality detection and evaluation system based on deep neural network according to claim 2 is characterized in that: The data preprocessing module specifically includes: Data cleaning unit: handles errors, omissions and inconsistencies in the collected data, uses mean interpolation algorithm to fill in missing parts of the data, and deletes duplicate records; Data denoising unit: Use Kalman filter to remove high-frequency noise. For image data, use median to remove noise and blurred areas in the image, smooth the data based on sliding average, and use signal removal algorithm to remove redundant noise and irrelevant data. Data standardization unit: Use Z-Score standardization to transform the data to a mean of 0 and a variance of 1 to make it conform to the normal distribution, and convert data in different units to ensure that the data can be compared on the same scale.

4. The construction quality detection and evaluation system based on deep neural network according to claim 3 is characterized in that: The feature extraction module specifically includes: Image data feature extraction unit: extracts visual features of image data of construction sites based on the convolutional neural network formula, inputs image data, extracts low-level features in the image through multiple convolutional layers, then reduces feature dimensions through pooling layers, extracts high-level features through fully connected layers, and forms image features related to construction quality; Time series data feature extraction unit: Use a recurrent neural network to extract time series data features in the construction process. Input time series data, extract long-term and short-term dependency features in the time series through the LSTM layer, and use the fully connected layer to convert the extracted time series features into features related to construction quality. Feature output unit: outputs the extracted features to the subsequent prediction model.

5. The construction quality detection and evaluation system based on deep neural network according to claim 4 is characterized in that: The method of extracting visual features of image data of a construction site based on a convolutional neural network formula specifically includes: Among them, the convolutional neural network formula is: h l+1 =σ(w l *h l +b l ) In the formula, h l is the input of the lth layer, w l is the convolution kernel, σ is the activation function, b l is the bias term, and * indicates the convolution operation.

6. The construction quality detection and evaluation system based on deep neural network according to claim 5 is characterized in that: The quality assessment module specifically includes: Feature receiving unit: The quality assessment module receives the features related to the construction from the feature extraction module, inputs the features into the neural network for processing, and predicts the construction quality score; Model reasoning and quality assessment unit: After reasoning, the deep neural network outputs the predicted results of construction quality, outputs the quality score according to the type of task, and maps the output results of the neural network to the actual quality assessment results; Quality report generation unit: Generates and outputs quality assessment reports, including quality scores, assessment indicators and recommendations.

7. The construction quality detection and evaluation system based on deep neural network according to claim 6 is characterized in that: The quality assessment module receives the construction-related features from the feature extraction module, inputs the features into the neural network for processing, and predicts the construction quality score. include: Among them, the construction quality score prediction formula is: In the formula, is the predicted value of construction quality score, f j is the weight matrix of the jth layer, k j-1 is the output of the j-1th layer, b j is the bias vector of the jth layer.

8. The construction quality detection and evaluation system based on deep neural network according to claim 7 is characterized in that: The quality report generating unit specifically includes: Based on the output of the deep neural network, the construction quality score is obtained, and the mean square error of the model is calculated to measure the square average of the difference between the model prediction value and the true value; Generate specific suggestions for improving construction quality based on the model prediction results and evaluation indicators; Based on the prediction results and evaluation indicators of the neural network, determine whether there are potential problems and which parts of the construction quality are at risk; Generate targeted suggestions for specific quality issues. For low-scoring areas, further inspection of the construction quality of the area is recommended. For unqualified parts, it is recommended to strengthen the construction process and check the quality of construction materials; For high-risk areas, it is recommended to add monitoring equipment to strengthen real-time data collection during the construction process. Propose actionable improvement measures, such as technical optimization, process improvement, and increased quality inspection, generate quality improvement suggestions, and clarify improvement directions and operational steps; Save and export the report.

9. The construction quality detection and evaluation system based on deep neural network according to claim 8 is characterized in that: The saving and outputting of the report specifically includes: Select HTML format to output the quality assessment report, and automatically generate the quality assessment report according to user needs; The report includes construction quality score, display quality score, lists the calculated values ​​of various indicators, and generates specific suggestions for construction quality improvement based on the prediction results and evaluation indicators; Save the report as a file for cloud storage.

10. The construction quality detection and evaluation system based on deep neural network according to claim 9 is characterized in that: The system interface module specifically includes: Construction quality report push unit: pushes the construction quality report, warning information and improvement suggestions generated by the quality assessment module to the building management personnel and the construction party in real time, and automatically pushes the construction quality report at set time intervals; Data synchronization unit of external system: synchronize data with external building management system, construction management system and equipment monitoring system; Data security and privacy protection unit: Use HTTPS encryption technology to ensure that data is not stolen during transmission, strengthen identity authentication and permission control of external systems, and ensure that only authorized users can access specific data interfaces.

Images