Construction site intelligent safety monitoring method, system, equipment and medium

Abstract

The invention relates to a construction site intelligent safety monitoring method, system and device and a medium, and belongs to the technical field of construction safety monitoring, and the safety monitoring method comprises the steps: obtaining the image monitoring information of a construction site, and carrying out the data preprocessing; performing feature extraction according to the image monitoring information after data preprocessing to obtain construction site feature data; on the basis of a machine learning algorithm, abnormal state prediction is carried out according to the construction site feature data, and a behavior feature abnormal prediction result and an equipment state abnormal prediction result are obtained; generating a behavior rectification prompt based on the behavior characteristic abnormity prediction result, and sending the behavior rectification prompt to the mobile terminal of the field personnel with the behavior characteristic abnormity; and generating an equipment hidden danger rectification prompt based on the equipment state abnormity prediction result, and sending the equipment hidden danger rectification prompt to a maintenance terminal corresponding to the field equipment with the equipment state abnormity. According to the invention, the safety management level of the construction site is improved, and the safety accident risk caused by human or equipment factors is reduced.

Description

Technical Field

[0001] The present application relates to the technical field of construction safety monitoring, and in particular to a method, system, equipment and medium for intelligent safety monitoring of a construction site. Background Art

[0002] With the development of the construction industry, safety management of construction sites has become increasingly important. Traditional safety management methods mainly rely on manual inspections and paper records, which is not only inefficient but also easily affected by human factors, making it difficult to achieve comprehensive and real-time monitoring of the construction site. Especially in large-scale construction sites, where there are many people and complex equipment, it is often difficult to cover all potential risk points by relying on human safety supervision, which can easily lead to the omission of safety hazards.

[0003] In order to improve the safety management level of construction sites, some technical means, such as video surveillance systems, have been gradually adopted in recent years. However, conventional video surveillance systems are mainly used for post-event review, and it is difficult to achieve real-time safety risk warning and rapid response. Therefore, in the face of the complex and changing dynamic environment of construction sites, how to use modern information technology to improve the safety management level of construction sites and realize real-time monitoring and intelligent analysis of construction sites has become a problem that needs to be solved urgently. Summary of the invention

[0004] In order to improve the safety management level of construction sites, the present application provides a construction site intelligent safety monitoring method, system, equipment and medium.

[0005] In the first aspect, the present application provides a construction site intelligent safety monitoring method, which adopts the following technical solutions: A construction site intelligent safety monitoring method, the safety monitoring method comprising: Obtain image monitoring information from the construction site and perform data preprocessing; Performing feature extraction based on the image monitoring information after data preprocessing to obtain construction site feature data; wherein the construction site feature data includes behavioral features of on-site personnel and status features of on-site equipment; Based on the machine learning algorithm, abnormal state prediction is performed according to the construction site characteristic data to obtain abnormal behavior characteristic prediction results and abnormal equipment state prediction results; Based on the abnormal behavior characteristic prediction result, a behavior correction prompt is generated and sent to the mobile terminal of the on-site personnel with abnormal behavior characteristics; Based on the equipment status abnormality prediction result, an equipment hidden danger rectification prompt is generated and sent to the maintenance terminal corresponding to the field equipment with the equipment status abnormality.

[0006] By adopting the above technical solutions, based on the real-time acquisition and processing of image monitoring information, feature extraction, abnormal state prediction and corresponding feedback mechanism, effective monitoring of abnormal behaviors of construction site personnel and hidden dangers of equipment is achieved, which not only improves the safety management level of the construction site, but also promotes the improvement of work efficiency and reduces the risk of safety accidents caused by human or equipment factors.

[0007] Optionally, before the step of obtaining the image monitoring information of the construction site, the method further includes: Collecting facial information of new entrants to the construction site and performing facial recognition identity verification on the new entrants; After the identity verification is passed, the admission information of the new entrant is collected and stored; wherein the admission information includes identity information, information of the subcontracting unit, information of the team, and information of the special operation certificate; Obtain the completion status of the safety training for the new entrant; The admission information of the new entrant who has completed the training is marked as having been safely trained, and the role information of the new entrant is updated to be an on-site staff.

[0008] By adopting the above technical solution, multiple information of new entrants, such as identity information, safety training status, etc., is collected and verified one by one, ensuring that only legal and trained personnel can enter the construction site, which not only improves the safety and work efficiency of the construction site, but also provides an efficient means of personnel management. At the same time, comprehensive and accurate data collection and processing also provides a reliable basis for subsequent management and decision-making.

[0009] Optionally, after the step of generating a violation voice prompt based on the abnormal behavior feature prediction result and sending the prompt to the mobile terminal of the on-site personnel with abnormal behavior features, the method further includes: Marking the on-site personnel with abnormal behavioral characteristics as personnel to be checked; Acquire the facial information of the person to be checked, and match the admission information of the person to be checked according to the facial information; Verify the entry information of the person to be verified and obtain the verification result; If the verification result is not in compliance with the preset requirements, a personnel rectification prompt is generated and sent to the management terminal corresponding to the area where the person to be verified is located.

[0010] By adopting the above technical solution, the system can monitor and record the behavioral characteristics of on-site personnel in real time, and verify whether the entry information of personnel with abnormal behavior is compliant, so as to provide timely personnel rectification prompts, thereby improving the standardization and safety of on-site personnel management.

[0011] Optionally, the step of performing feature extraction based on the image monitoring information after data preprocessing to obtain construction site feature data includes: Extracting a foreground image from the preprocessed image monitoring information to obtain a segmented foreground image; Performing contour detection on the segmented foreground image based on an edge detection algorithm to obtain image information including contours of on-site personnel and on-site equipment; Extracting key point information of the human body from the image information containing the contour of the on-site personnel to obtain the behavioral characteristics of the on-site personnel; wherein the behavioral characteristics of the on-site personnel include the personnel position, posture, behavioral action and helmet wearing status; The key point information of the equipment is extracted from the image information containing the contour of the on-site equipment to obtain the state characteristics of the on-site equipment; wherein the state characteristics of the on-site equipment include the equipment position, motion trajectory, vibration amplitude and equipment surface state.

[0012] By adopting the above technical solutions, high-quality image information is obtained based on foreground image extraction and edge detection. By utilizing personnel posture estimation, target detection, and equipment status recognition, the key points and status features of personnel and equipment are accurately extracted, thus achieving comprehensive monitoring and management of the construction site and improving the safety and operational efficiency of the construction site.

[0013] Optionally, based on a machine learning algorithm, the step of predicting an abnormal state according to the construction site characteristic data includes: The construction site feature data is input into a pre-trained abnormal state prediction model, and based on the model output results, the behavior feature abnormality prediction results and the equipment state abnormality prediction results are obtained; wherein the abnormal state prediction model is trained based on multiple normal or abnormal historical sample image data and pre-labeled abnormal labels.

[0014] By adopting the above technical solution and using deep learning technology to analyze and predict construction site characteristic data, automatic abnormal monitoring of personnel behavior and equipment status is achieved, greatly improving the efficiency and reliability of construction site safety management and equipment maintenance.

[0015] Optionally, the method further includes a training step of the abnormal state prediction model, wherein the training step includes: Obtain sample data sets and perform data preprocessing; The sample data set includes a plurality of normal or abnormal historical sample image data and pre-labeled abnormal labels, and the historical sample image data includes personnel behavior sample images and equipment status sample images; Perform feature extraction on the preprocessed sample data set to obtain sample feature data; Divide the sample data set after feature extraction into training set, validation set and test set; Input the training set into a pre-built machine learning model for training, optimize the model parameters and calculate the loss function of the machine learning model until the loss function meets a preset condition or the number of model iterations reaches a preset number, thereby obtaining the trained abnormal state prediction model; Validating the abnormal state prediction model based on the validation set, evaluating the performance of the abnormal state prediction model and adjusting the hyperparameters of the model; The prediction ability of the adjusted abnormal state prediction model is tested based on the test set.

[0016] By adopting the above technical solution, the abnormal state prediction model based on support vector machine (SVM) has greatly improved the model accuracy through scientific sample data preprocessing, effective feature extraction and reasonable model training and verification process, realized accurate monitoring of real-time image data of the construction site and effective anomaly identification, and enhanced the safety of the construction site and the reliability of equipment management.

[0017] In the second aspect, the present application provides a construction site intelligent safety monitoring system, which adopts the following technical solutions: An intelligent safety monitoring system for a construction site, the safety monitoring system comprising: Image monitoring module, used to obtain image monitoring information of the construction site; A preprocessing module, used for performing data preprocessing on the image monitoring information; A feature extraction module, used to extract features based on the image monitoring information after data preprocessing to obtain construction site feature data; wherein the construction site feature data includes the behavior features of the on-site personnel and the status features of the on-site equipment; An abnormal state prediction module is used to predict abnormal states based on the construction site characteristic data based on a machine learning algorithm, and obtain abnormal behavior characteristic prediction results and abnormal equipment state prediction results; A behavior correction prompt module, used to generate a behavior correction prompt based on the abnormal behavior characteristic prediction result and send it to the mobile terminal of the on-site personnel with abnormal behavior characteristics; The equipment hidden danger rectification prompt module is used to generate an equipment hidden danger rectification prompt based on the equipment status abnormality prediction result and send it to the maintenance terminal corresponding to the field equipment with the equipment status abnormality.

[0018] Optionally, the safety monitoring system further includes: A face verification module is used to collect face information of new entrants to the construction site and perform face recognition identity verification on the new entrants; An admission information storage module is used to collect and store the admission information of the new entrant after the identity verification is passed; wherein the admission information includes identity information, subcontracting unit information, team information and special operation certificate information; An acquisition module is used to acquire the completion status of the safety training of the new entrant; A marking module, used to mark the admission information of the new entrants who have completed the training as having been safely trained; The role information updating module is used to update the role information of the new entrant to an on-site staff.

[0019] In a third aspect, the present application provides a computer device, which adopts the following technical solution: A computer device comprises a memory, a processor and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method as described in the first aspect.

[0020] In a fourth aspect, the present application provides a computer-readable storage medium, which adopts the following technical solution: A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any one of the methods in the first aspect.

[0021] To summarize, the present application includes at least one of the following beneficial technical effects: based on the real-time acquisition and processing of image monitoring information, feature extraction, abnormal state prediction and corresponding rectification feedback mechanism, it realizes effective monitoring of abnormal behaviors of construction site personnel and hidden dangers of equipment, and has the advantages of high real-time, strong accuracy and wide coverage, which effectively reduces the incidence of safety accidents and promotes the continuous development of the construction industry towards intelligence and modernization. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 It is a first flow chart of a construction site intelligent safety monitoring method according to one of the embodiments of the present application.

[0023] Figure 2 This is a second flow chart of a construction site intelligent safety monitoring method according to one of the embodiments of the present application.

[0024] Figure 3 It is a third flow chart of a construction site intelligent safety monitoring method according to one of the embodiments of the present application.

[0025] Figure 4 It is a fourth flow chart of a construction site intelligent safety monitoring method according to one of the embodiments of the present application.

[0026] Figure 5It is a fifth flow chart of a construction site intelligent safety monitoring method according to one of the embodiments of the present application. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of this application more clear, the following Figure 1-5 It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] The embodiment of the present application discloses a construction site intelligent safety monitoring method.

[0029] Reference Figure 1 , a construction site intelligent safety monitoring method, the safety monitoring method comprising: Step S101, obtaining image monitoring information of the construction site and performing data preprocessing; In some embodiments, on-site image information can be collected through a wearable intelligent image acquisition device installed on the safety officer, and image monitoring information can also be obtained in real time through high-definition camera equipment at the construction site.

[0030] Among them, image monitoring information may contain data from multiple perspectives, multiple resolutions, and different time periods; data preprocessing may include steps such as image denoising, unification of image resolution, and filtering of irrelevant information in the image to ensure the consistency and high quality of the input data.

[0031] For example, after acquiring an image of the construction site, the image can be converted into a grayscale image through a preprocessing step, and smoothed filtered to remove noise interference, and then the image size can be adjusted to a fixed resolution for subsequent processing.

[0032] Step S102, extracting features based on the image monitoring information after data preprocessing to obtain construction site feature data; Among them, the construction site characteristic data includes the behavior characteristics of the on-site personnel and the status characteristics of the on-site equipment; Specifically, feature extraction refers to identifying meaningful information from preprocessed images. Computer vision technology can be used to extract feature data of the construction site from preprocessed images. These feature data may include personnel's position, posture, behavior, helmet wearing status, etc., and equipment status characteristics such as equipment's position, motion trajectory, vibration amplitude, surface crack conditions, image information, etc.

[0033] It is understandable that for construction sites, the goal of feature extraction is to extract behavioral characteristics of on-site personnel (such as whether they wear safety helmets, whether they observe safe distances, whether they are close to dangerous locations, whether there are any violations, etc.) and status characteristics of on-site equipment (such as whether the equipment is placed abnormally, whether the equipment's movement trajectory is abnormal, whether there are cracks or damage on the equipment surface, whether the equipment is leaking liquid or oil, etc.).

[0034] For example, the edge detection algorithm can be used to extract the outline of a person, detect the wearing status of the person's helmet, the person's operating behavior and other characteristics; by analyzing the device position, device vibration or movement status and device surface image features, it can be identified whether there is an abnormal equipment status.

[0035] Step S103, based on a machine learning algorithm, abnormal state prediction is performed according to the construction site characteristic data to obtain a behavior characteristic abnormality prediction result and an equipment state abnormality prediction result; In some embodiments, a trained machine learning model (such as a support vector machine (SVM) model) can be used to input the data obtained by feature extraction into the model for classification and evaluation, thereby predicting whether there is an abnormal state in human behavior or equipment status.

[0036] Step S104, based on the abnormal behavior characteristic prediction result, generating a behavior correction prompt and sending it to the mobile terminal of the on-site personnel with abnormal behavior characteristics; Among them, when abnormal behavioral characteristics are predicted, the system will automatically generate a voice prompt containing information such as the type of abnormality, and send it to the mobile terminal of the corresponding personnel (such as mobile phones, walkie-talkies, smart watches, etc.) through wireless communication technology, immediately reminding personnel to correct safety violations and reduce the chance of accidents.

[0037] For example, when the system detects that a site worker is not wearing a safety helmet correctly, it immediately generates a voice prompt: "Please note that you are not wearing a safety helmet as required. Please correct it immediately." and sends this prompt to the worker's mobile phone or walkie-talkie or other mobile terminal.

[0038] Step S105: Based on the equipment status abnormality prediction result, an equipment hidden danger rectification prompt is generated and sent to the maintenance terminal corresponding to the field equipment with the equipment status abnormality.

[0039] Among them, when it is predicted that the equipment status is abnormal, the system will generate equipment hidden danger rectification prompt information, including equipment number, abnormality description, etc., and send it to the maintenance terminal (such as tablet computer, mobile phone, computer terminal, etc.) of the maintenance personnel responsible for the equipment through the wireless network.

[0040] In some embodiments, abnormalities in equipment status include, but are not limited to, abnormal equipment position (for example, the equipment is vacant or appears in a position where it should not appear), abnormal equipment movement trajectory (for example, the movement trajectory of a tower crane arm with excessive swing amplitude), cracks or damage on the equipment surface (for example, cracks appear on the surface of an excavator bucket), abnormal equipment vibration (for example, the vibration of the compressor is significantly too large), and equipment leakage or oil leakage (for example, oil leakage in hydraulic equipment).

[0041] For example, when a device is identified to have an oil leak or other problems, the system will generate a prompt to correct the hidden dangers of the equipment: "The xx device numbered XX has an oil leak, please repair it immediately." This prompt is then sent to the computer or mobile terminal of the maintenance department via a wireless network.

[0042] In the above implementation, based on the real-time acquisition and processing of image monitoring information, feature extraction, abnormal state prediction and corresponding feedback mechanism, effective monitoring of abnormal behaviors of construction site personnel and hidden dangers of equipment is achieved, which not only improves the safety management level of the construction site, but also promotes the improvement of work efficiency and reduces the risk of safety accidents caused by human or equipment factors.

[0043] Reference Figure 2 As a further implementation of the safety monitoring method, before the step of obtaining the image monitoring information of the construction site, it also includes: Step S201, collecting facial information of new entrants to the construction site and performing facial recognition and identity verification on the new entrants; Specifically, a facial recognition camera is installed at the entrance of the construction site. When new employees enter the construction site, the system automatically collects their facial images. This process can complete identity verification by capturing facial features through the camera and comparing them with the identity information pre-entered into the facial database.

[0044] It can be understood that this step can ensure that every new employee entering the construction site is accurately identified, preventing unauthorized persons from entering the construction site, and improving the safety and management efficiency of the construction site by confirming the identities of each person as a legal entrant.

[0045] Step S202, collecting and storing the admission information of the new entrant after the identity verification is passed; Among them, the admission information includes identity information, subcontractor information, team information and special operation certificate information; Specifically, after identity authentication is passed, the system will further collect detailed information of new entrants so that managers can clearly understand the identity and responsibility allocation of each staff member, which is convenient for subsequent personnel management and scheduling.

[0046] For example, after identity verification is passed, the detailed information of the new entrant is entered into the system, including his / her ID number, the subcontracting unit and specific team to which he / she belongs. In addition, information on a specific type of work certificate held will also be registered.

[0047] Step S203, obtaining the completion status of the safety training for new entrants; The system automatically queries and determines whether new entrants have completed the necessary safety training. This can be achieved by connecting to the internal training management system, or by the management staff manually entering the completion status information.

[0048] Step S204: Mark the admission information of the new entrant who has completed the training as having been safely trained, and update the role information of the new entrant to on-site personnel.

[0049] Specifically, once it is confirmed that the new entrant has completed safety training, the system will mark the entry information of the person as "safety training completed" and change his role from "new entrant" to "on-site personnel". Through the update of roles, the system can distinguish which are qualified personnel who have been trained, and thus allow these personnel to enter the construction site to work.

[0050] In the above implementation, by collecting and verifying multiple information of new entrants one by one, such as identity information, safety training status, etc., it is ensured that only legal and trained personnel can enter the construction site, which not only improves the safety and work efficiency of the construction site, but also provides an efficient means of personnel management. At the same time, comprehensive and accurate data collection and processing also provides a reliable basis for subsequent management and decision-making.

[0051] Reference Figure 3 As a further implementation method of the intelligent safety monitoring method for construction sites, after the step of generating a violation voice prompt based on the abnormal behavior feature prediction result and sending it to the mobile terminal of the on-site personnel with abnormal behavior features, it also includes: Step S301, marking on-site personnel with abnormal behavior characteristics as personnel to be checked; Specifically, when generating a violation alert, the system will mark the on-site personnel with abnormal behavior as a person to be verified, ensuring that all abnormal behaviors are tracked and recorded to avoid missing any potential risks.

[0052] Step S302, obtaining facial information of the person to be checked, and matching the admission information of the person to be checked according to the facial information; Among them, the system obtains the facial information of the person to be verified, and matches it with the records in the entry information database through facial recognition technology to confirm their identity and other relevant information, ensuring the effectiveness and reliability of the subsequent verification process.

[0053] Step S303, verifying the entry information of the person to be verified and obtaining the verification result; In some embodiments, a comprehensive check is performed on the entry information of the personnel to be checked, including but not limited to checking identity information, information of the subcontracting unit, information of the team, special operation certificate information and completion status of safety training.

[0054] Among them, the focus is on verifying the special operation certificate information and the completion status of safety training to ensure that personnel are legally qualified to engage in specific operations, and to ensure that personnel have received safety education and training before entering the site; through strict verification, it is ensured that every on-site personnel has legal and compliant work qualifications, thereby effectively reducing management and production risks.

[0055] Step S304: If the verification result is that the preset requirements are not met, a personnel rectification prompt is generated and sent to the management terminal corresponding to the area where the person to be verified is located.

[0056] Among them, the preset requirements can be pre-set and adjusted according to actual conditions; in the embodiment of the present application, the preset requirements include, for example, verifying whether the personnel have the corresponding special operation certificate and whether the certificate information is compliant, and verifying whether the safety training completion status is completed.

[0057] Specifically, if the verification result shows that the requirements are not met, the system will generate relevant rectification prompts and send the prompt information to the management terminal in the area where the person to be verified is located, so that the management personnel can take further corrective measures in a timely manner.

[0058] In the above implementation, the system can monitor and record the behavioral characteristics of on-site personnel in real time, and verify whether the entry information of personnel with abnormal behavior is compliant, so as to provide timely personnel rectification prompts, thereby improving the standardization and safety of on-site personnel management.

[0059] Reference Figure 4 As an implementation method of step S102, the step of extracting features based on the image monitoring information after data preprocessing to obtain construction site feature data includes: Step S401, extracting a foreground image from the preprocessed image monitoring information to obtain a segmented foreground image; In some embodiments, foreground image extraction is to segment the region of interest (foreground image) from the image through an image processing algorithm, usually using a background subtraction method, which is mainly based on extracting foreground objects that are different from the background, such as people and equipment, by comparing the differences between consecutive frames in a static environment, thereby laying the foundation for subsequent feature extraction.

[0060] Step S402, performing contour detection on the segmented foreground image based on an edge detection algorithm to obtain image information including contours of on-site personnel and on-site equipment; Among them, the Canny edge detection algorithm can be used to detect the edges in the foreground image by identifying the parts of the image with large changes in pixel values, and then the contours of people and equipment are extracted through connectivity analysis, which helps to accurately locate the shape and position of the foreground object and provide high-quality input data for the subsequent key point extraction.

[0061] Step S403, extracting human body key point information from the image information containing the contour of the on-site personnel to obtain the behavioral characteristics of the on-site personnel; wherein the behavioral characteristics of the on-site personnel include the personnel's position, posture, behavioral action, and helmet wearing status; In some embodiments, the human posture and behavior features in the image can be extracted based on a human posture estimation algorithm such as OpenPose; at the same time, the position of the person and the helmet wearing status of the person's head area can be detected based on a target detection algorithm (such as YOLOv4); Step S404, extracting key point information of the equipment from the image information containing the contour of the on-site equipment to obtain the state characteristics of the on-site equipment; wherein the state characteristics of the on-site equipment include the equipment position, motion trajectory, vibration amplitude and equipment surface state.

[0062] In some embodiments, the position of the device in the image can be located through a target detection algorithm, and then its coordinates can be calculated to obtain the position information of the device; the motion trajectory of the device can be obtained by tracking the device for multiple frames; the subtle movements of the device can be captured by a high-frame rate camera and the optical flow analysis can be performed on the video frames to obtain the vibration amplitude of the device. For example, the jitter of the tower crane boom can be monitored to obtain its vibration frequency and amplitude; image texture analysis and surface defect detection algorithms can be used to identify and analyze various abnormalities on the surface of the equipment, such as cracks, liquid leakage, oil leakage, etc., to provide timely reminders to rectify equipment hidden dangers, extend equipment life and ensure safe operation.

[0063] In the above implementation, high-quality image information is obtained based on foreground image extraction and edge detection; by utilizing personnel posture estimation, target detection, and equipment status recognition, the key points and status features of personnel and equipment are accurately extracted, thereby achieving comprehensive monitoring and management of the construction site and improving the safety and operational efficiency of the construction site.

[0064] As an implementation of step S103, based on a machine learning algorithm, the step of predicting an abnormal state according to the construction site characteristic data includes: Input the construction site feature data into the pre-trained abnormal state prediction model, and obtain the behavior feature abnormality prediction results and the equipment state abnormality prediction results according to the model output results; Among them, the abnormal state prediction model is trained based on multiple normal or abnormal historical sample image data and pre-labeled abnormal labels.

[0065] Specifically, two main outputs are obtained through model calculation: abnormal behavior characteristics and abnormal equipment status. By distinguishing different types of abnormalities, it is helpful to take targeted corrective measures.

[0066] In the above implementation, deep learning technology is used to analyze and predict construction site characteristic data, realizing automated abnormal monitoring of personnel behavior and equipment status, greatly improving the efficiency and reliability of construction site safety management and equipment maintenance.

[0067] Reference Figure 5 As an implementation of the abnormal state prediction model, the training steps of the abnormal state prediction model include: Step S501, obtaining a sample data set and performing data preprocessing; The sample data set includes a plurality of normal or abnormal historical sample image data and pre-labeled abnormal labels, and the historical sample image data includes personnel behavior sample images and equipment status sample images; Specifically, a large amount of historical sample image data and their corresponding anomaly labels are collected, and these data are preprocessed to improve model performance; In some embodiments, the historical sample image data includes a large amount of historical image data of construction sites, which includes not only historical images of normal feature states, but also historical images of abnormal feature states, and the corresponding abnormal labels are pre-marked; the data preprocessing part includes image processing steps such as denoising and normalization to improve the quality of the data and reduce the computational complexity during model training.

[0068] Step S502, extracting features from the preprocessed sample data set to obtain sample feature data; The feature extraction step of the sample feature data may refer to the detailed process of extracting the construction site feature data from step S401 to step S404. In order to improve the accuracy of the model output, the feature extraction methods of the two should be kept as consistent as possible.

[0069] Step S503, dividing the sample data set after feature extraction into a training set, a validation set and a test set; Specifically, the division of the data set helps to evaluate the generalization ability of the model. The training set is used for model training, the validation set is used for hyperparameter tuning, and the test set is used for final performance evaluation. Through reasonable data set division, the model can be effectively prevented from overfitting and its performance in practical applications can be improved.

[0070] Step S504, inputting the training set into a pre-built machine learning model for training, optimizing the model parameters and calculating the loss function of the machine learning model, until the loss function meets a preset condition or the number of model iterations reaches a preset number, thereby obtaining a trained abnormal state prediction model; In one of the embodiments of the present application, the pre-built machine learning model may adopt a support vector machine (SVM) model. The support vector machine has good classification ability and strong generalization performance, is suitable for small sample and high-dimensional data, and can obtain relatively stable results for detecting abnormal states in the embodiments of the present application. In this embodiment, the training set is input into the SVM model, and the optimal hyperplane is found through the optimization algorithm, so that the classification effect of the samples in the training set is the best.

[0071] For example, the SVM model can be trained using the RBF kernel function (Radial Basis Function), and the best hyperparameter combination (such as C and Gamma) can be found through grid search.

[0072] Specifically, the pre-built abnormal state prediction model mainly includes an input layer, a kernel function processing layer, a hyperplane decision layer and an output layer; wherein the input layer is used to receive sample data after preprocessing and feature extraction; the kernel function processing layer is used to map the input low-dimensional space data to a high-dimensional space through a kernel function (such as an RBF kernel function) so as to find a hyperplane in the high-dimensional space that can better separate normal and abnormal samples; the hyperplane decision layer is the core part of the SVM model, which mainly determines the best decision hyperplane through an optimization algorithm, that is, maximizes the interval between samples, through which normal and abnormal samples can be effectively classified; the output layer is used to classify the input data according to the trained hyperplane, and determine the abnormal state and abnormal type of each input data.

[0073] Step S505, verifying the abnormal state prediction model based on the verification set, evaluating the performance of the abnormal state prediction model and adjusting the hyperparameters of the model; The validation set is input into the trained SVM model to evaluate the performance of the model on an independent dataset (such as accuracy, recall, etc.), and the hyperparameters of the model are adjusted based on the results. For example, if it is found that the current hyperparameter combination causes the model to perform poorly on the validation set, the values ​​of C and Gamma can be adjusted, and the model can be retrained and validated until the performance of the model on the validation set reaches a satisfactory level.

[0074] Step S506, testing the prediction capability of the adjusted abnormal state prediction model based on the test set.

[0075] Specifically, the test set is input into the verified and adjusted SVM model to comprehensively evaluate its actual application effect on unseen data. By verifying the real application value of the model, it is ensured that it meets actual needs.

[0076] In the above implementation, the abnormal state prediction model based on support vector machine (SVM) greatly improves the model accuracy through scientific sample data preprocessing, effective feature extraction and reasonable model training and verification process, realizes accurate monitoring of real-time image data of the construction site and effective anomaly identification, and enhances the safety of the construction site and the reliability of equipment management.

[0077] In practical applications, the intelligent safety monitoring method of this application can effectively improve the safety management level of the construction site. First, the system can monitor the construction site in real time, promptly discover and warn of potential safety hazards, and prevent them before they happen; second, through accurate analysis of the behavior of on-site personnel and the status of equipment, the system can provide specific rectification suggestions to help on-site personnel correct unsafe behaviors and eliminate equipment hazards in a timely manner, thereby ensuring the smooth progress of the construction process; finally, the introduction of this system will help establish a digital and intelligent construction management system, significantly reduce safety accidents caused by human negligence, improve construction efficiency, and promote the construction industry to develop in an intelligent and modern direction.

[0078] The embodiment of the present application also discloses an intelligent safety monitoring system for a construction site.

[0079] An intelligent safety monitoring system for a construction site, the safety monitoring system comprising: Image monitoring module, used to obtain image monitoring information of the construction site; A preprocessing module, used for performing data preprocessing on image monitoring information; A feature extraction module is used to extract features based on the image monitoring information after data preprocessing to obtain construction site feature data; wherein the construction site feature data includes the behavior features of the on-site personnel and the status features of the on-site equipment; The abnormal state prediction module is used to predict the abnormal state based on the construction site characteristic data based on the machine learning algorithm, and obtain the behavior characteristic abnormality prediction results and the equipment state abnormality prediction results; A behavior correction prompt module is used to generate behavior correction prompts based on the prediction results of abnormal behavior characteristics and send them to the mobile terminals of on-site personnel with abnormal behavior characteristics; The equipment hidden danger rectification prompt module is used to generate equipment hidden danger rectification prompts based on the equipment status abnormality prediction results and send them to the maintenance terminal corresponding to the on-site equipment with abnormal equipment status.

[0080] In the above implementation, comprehensive intelligent monitoring of the construction site is achieved through the collaborative work of multiple modules. The system can automatically acquire and process image information of the construction site, extract personnel behavior characteristics and equipment status characteristics, and use machine learning algorithms to predict potential safety hazards and issue rectification prompts in a timely manner. Compared with traditional manual monitoring methods, the system has the advantages of high real-time performance, strong accuracy and wide coverage, which effectively reduces the incidence of safety accidents.

[0081] As a further implementation of the safety monitoring system, the safety monitoring system further includes: The face verification module is used to collect the face information of new entrants to the construction site and perform face recognition identity verification on the new entrants; The admission information storage module is used to collect and store the admission information of new entrants after identity verification; the admission information includes identity information, subcontractor information, team information and special operation certificate information; The acquisition module is used to obtain the safety training completion status of new entrants; A marking module is used to mark the admission information of new entrants who have completed training as having been safely trained; The role information update module is used to update the role information of new entrants to on-site personnel.

[0082] In the above implementation, by ensuring that only legal and trained personnel can enter the construction site, not only the safety and work efficiency of the construction site are improved, but also an efficient personnel management method is provided, providing a reliable basis for subsequent management and decision-making.

[0083] The construction site intelligent safety monitoring system of the embodiment of the present application can implement any of the above-mentioned construction site intelligent safety monitoring methods, and the specific working process of each module in the construction site intelligent safety monitoring system can refer to the corresponding process in the above-mentioned method embodiment.

[0084] In the several embodiments provided in this application, it should be understood that the provided methods and systems can be implemented in other ways. For example, the system embodiments described above are only illustrative; for example, the division of a certain module is only a logical function division, and there may be other division methods in actual implementation, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0085] The embodiment of the present application also discloses a computer device.

[0086] The computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the above-mentioned intelligent safety monitoring method for a construction site is implemented.

[0087] The embodiment of the present application also discloses a computer-readable storage medium.

[0088] A computer-readable storage medium stores a computer program that can be loaded by a processor and execute any of the above-mentioned construction site intelligent safety monitoring methods.

[0089] Among them, computer-readable storage media can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus or device; the program code contained on the computer-readable medium can be transmitted using any appropriate medium, including but not limited to wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0090] It should be noted that in the above embodiments, the description of each embodiment has different emphases, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0091] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. A construction site intelligent safety monitoring method, characterized in that: The safety monitoring method comprises: Obtain image monitoring information from the construction site and perform data preprocessing; Performing feature extraction based on the image monitoring information after data preprocessing to obtain construction site feature data; wherein the construction site feature data includes behavioral features of on-site personnel and status features of on-site equipment; Based on the machine learning algorithm, abnormal state prediction is performed according to the construction site characteristic data to obtain abnormal behavior characteristic prediction results and abnormal equipment state prediction results; Based on the abnormal behavior characteristic prediction result, a behavior correction prompt is generated and sent to the mobile terminal of the on-site personnel with abnormal behavior characteristics; Based on the equipment status abnormality prediction result, an equipment hidden danger rectification prompt is generated and sent to the maintenance terminal corresponding to the field equipment with the equipment status abnormality.

2. A construction site intelligent safety monitoring method according to claim 1, characterized in that: Before the step of obtaining the image monitoring information of the construction site, the method further includes: Collecting facial information of new entrants to the construction site and performing facial recognition identity verification on the new entrants; After the identity verification is passed, the admission information of the new entrant is collected and stored; wherein the admission information includes identity information, information of the subcontracting unit, information of the team, and information of the special operation certificate; Obtain the completion status of the safety training for the new entrant; The admission information of the new entrant who has completed the training is marked as having been safely trained, and the role information of the new entrant is updated to be an on-site staff.

3. A construction site intelligent safety monitoring method according to claim 2, characterized in that: After the step of generating a violation voice prompt based on the abnormal behavior feature prediction result and sending it to the mobile terminal of the on-site personnel with abnormal behavior features, the method further includes: Marking the on-site personnel with abnormal behavioral characteristics as personnel to be checked; Acquire the facial information of the person to be checked, and match the admission information of the person to be checked according to the facial information; Verify the entry information of the person to be verified and obtain the verification result; If the verification result is not in compliance with the preset requirements, a personnel rectification prompt is generated and sent to the management terminal corresponding to the area where the person to be verified is located.

4. A construction site intelligent safety monitoring method according to claim 1, characterized in that: The step of extracting features based on the image monitoring information after data preprocessing to obtain construction site feature data includes: Extracting a foreground image from the preprocessed image monitoring information to obtain a segmented foreground image; Performing contour detection on the segmented foreground image based on an edge detection algorithm to obtain image information including contours of on-site personnel and on-site equipment; Extracting key point information of the human body from the image information containing the contour of the on-site personnel to obtain the behavioral characteristics of the on-site personnel; wherein the behavioral characteristics of the on-site personnel include the personnel position, posture, behavioral action and helmet wearing status; The key point information of the equipment is extracted from the image information containing the contour of the on-site equipment to obtain the state characteristics of the on-site equipment; wherein the state characteristics of the on-site equipment include the equipment position, motion trajectory, vibration amplitude and equipment surface state.

5. A construction site intelligent safety monitoring method according to any one of claims 1 to 4, characterized in that: Based on the machine learning algorithm, the steps of predicting abnormal conditions according to the construction site characteristic data include: The construction site feature data is input into a pre-trained abnormal state prediction model, and based on the model output results, the behavior feature abnormality prediction results and the equipment state abnormality prediction results are obtained; wherein the abnormal state prediction model is trained based on multiple normal or abnormal historical sample image data and pre-labeled abnormal labels.

6. A construction site intelligent safety monitoring method according to claim 5, characterized in that: The method further includes a training step of the abnormal state prediction model, wherein the training step includes: Acquire a sample data set and perform data preprocessing; wherein the sample data set includes a plurality of normal or abnormal historical sample image data and pre-labeled abnormal labels, and the historical sample image data includes personnel behavior sample images and equipment status sample images; Perform feature extraction on the preprocessed sample data set to obtain sample feature data; Divide the sample data set after feature extraction into training set, validation set and test set; Input the training set into a pre-built machine learning model for training, optimize the model parameters and calculate the loss function of the machine learning model until the loss function meets a preset condition or the number of model iterations reaches a preset number, thereby obtaining the trained abnormal state prediction model; Validating the abnormal state prediction model based on the validation set, evaluating the performance of the abnormal state prediction model and adjusting the hyperparameters of the model; The prediction ability of the adjusted abnormal state prediction model is tested based on the test set.

7. An intelligent safety monitoring system for a construction site, characterized in that: The safety monitoring system comprises: Image monitoring module, used to obtain image monitoring information of the construction site; A preprocessing module, used for performing data preprocessing on the image monitoring information; A feature extraction module, used to extract features based on the image monitoring information after data preprocessing to obtain construction site feature data; wherein the construction site feature data includes the behavior features of the on-site personnel and the status features of the on-site equipment; An abnormal state prediction module is used to predict abnormal states based on the construction site characteristic data based on a machine learning algorithm, and obtain abnormal behavior characteristic prediction results and abnormal equipment state prediction results; A behavior correction prompt module, used to generate a behavior correction prompt based on the abnormal behavior characteristic prediction result and send it to the mobile terminal of the on-site personnel with abnormal behavior characteristics; The equipment hidden danger rectification prompt module is used to generate an equipment hidden danger rectification prompt based on the equipment status abnormality prediction result and send it to the maintenance terminal corresponding to the field equipment with the equipment status abnormality.

8. The intelligent safety monitoring system for a construction site according to claim 7, characterized in that: The safety monitoring system also includes: A face verification module is used to collect face information of new entrants to the construction site and perform face recognition identity verification on the new entrants; An admission information storage module is used to collect and store the admission information of the new entrant after the identity verification is passed; wherein the admission information includes identity information, subcontracting unit information, team information and special operation certificate information; An acquisition module is used to acquire the completion status of the safety training of the new entrant; A marking module, used to mark the admission information of the new entrants who have completed the training as having been safely trained; The role information updating module is used to update the role information of the new entrant to an on-site staff.

9. A computer device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method according to any one of claims 1 to 6 is implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and execute the method according to any one of claims 1 to 6.

Images