Risk early warning model construction method and device, electronic equipment and storage medium

Abstract

The invention discloses a risk early warning model construction method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining a target image; detecting the target image, and determining a target object and a moving speed of the target object; identifying the posture of the target object to obtain a target posture; identifying the safety measure wearing condition of the target object to obtain a wearing result; and according to the moving speed, the target attitude and the wearing result of the target object, constructing an anti-high-fall safety risk early warning model. According to the technical scheme, automatic tracking and monitoring and safety risk pre-judgment of safety protection measure implementation conditions of the high-altitude operation personnel are realized.

Description

Technical Field

[0001] The present invention relates to the field of risk warning technology, and in particular to a risk warning model construction method, device, electronic equipment and storage medium. Background Art

[0002] In power transmission operation scenarios, ensuring the safety of workers working at height is crucial. However, due to the complexity of on-site operations, there are some technical challenges that limit the effective monitoring and management of worker safety.

[0003] First, human factors lead to irregular placement of cameras. Workers may not be able to place cameras in the best position to ensure comprehensive monitoring of the work area due to various reasons (such as space limitations, equipment layout, etc.). In addition, the design and technical parameters of the camera itself also affect its field of view and monitoring capabilities. For example, the viewing angle of traditional cameras may not be wide enough, resulting in some work areas being in blind spots. This means that when workers move to these blind spots, the monitoring system may not be able to accurately capture their position and activity status. At the same time, due to the complexity of structures such as power lines and towers, cameras need to have strong zoom capabilities to cover long-distance targets, but existing monitoring equipment still has limitations in long-distance target recognition. Furthermore, existing technologies are also insufficient in automatic tracking and monitoring. Although some systems can detect the general location of personnel, they are unable to continuously track the movements and postures of individual workers. Especially in complex working environments, the system's tracking capabilities may be disturbed, thereby affecting the accuracy and timeliness of safety warnings. In terms of safety warnings, existing technologies often rely on simple logical judgments or preset rules to issue warning signals. This rule-based approach is difficult to adapt to changing working environments and emergencies, so the accuracy and reliability of early warnings need to be improved. For example, if a worker does not wear a safety belt or safety rope as required, the existing system may not accurately identify this violation, or fail to respond in time when an early warning should be issued.

[0004] In summary, although video surveillance systems play an important role in monitoring workers working at heights, existing technical means are not sufficient to overcome the challenges brought about by irregular human placement, limited camera viewing angles, and the ability to identify small targets at long distances. Summary of the invention

[0005] The present invention provides a risk warning model construction method, device, electronic equipment and storage medium, which realize automatic tracking and monitoring of the implementation of safety protection measures for high-altitude workers and safety risk prediction.

[0006] According to one aspect of the present invention, a method for constructing a risk warning model is provided, the method comprising:

[0007] Get the target image;

[0008] Detecting the target image to determine the target object and the moving speed of the target object;

[0009] Recognizing the posture of the target object to obtain the target posture;

[0010] Identify the wearing status of the safety measures of the target object and obtain a wearing result;

[0011] A high-fall prevention safety risk warning model is constructed according to the moving speed, target posture and wearing result of the target object.

[0012] According to another aspect of the present invention, a risk warning model construction device is provided, the device comprising:

[0013] A target image acquisition module, used for acquiring a target image;

[0014] A detection module, used to detect the target image and determine the target object and the moving speed of the target object;

[0015] A posture recognition module, used to recognize the posture of the target object to obtain the target posture;

[0016] A safety measure wearing condition identification module is used to identify the safety measure wearing condition of the target object and obtain a wearing result;

[0017] The anti-fall safety risk warning model construction module is used to construct an anti-fall safety risk warning model according to the moving speed, target posture and wearing result of the target object.

[0018] According to another aspect of the present invention, there is provided an electronic device, the electronic device comprising:

[0019] At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute a risk warning model construction method described in any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement a risk warning model construction method described in any embodiment of the present invention when executed.

[0021] The technical solution of the embodiment of the present invention obtains a target image, detects the target image, determines the target object and the moving speed of the target object, identifies the posture of the target object to obtain the target posture, and identifies the wearing of the safety measures of the target object to obtain the wearing result, and then constructs a high-fall prevention safety risk warning model according to the moving speed of the target object, the target posture and the wearing result. This technical solution realizes the automatic tracking and monitoring of the implementation of safety protection measures for high-altitude workers and the prediction of safety risks.

[0022] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0024] Figure 1 is a flow chart of a method for constructing a risk warning model provided according to Embodiment 1 of the present invention;

[0025] Figure 2 A schematic diagram of the risk warning model construction process provided in Example 1 of the present application;

[0026] Figure 3 This is a schematic diagram of the deep learning detection of safety belts and safety ropes for aerial work provided by the network framework based on the YOLO network model provided in the first implementation of this application;

[0027] Figure 4 A schematic diagram of the structure of a risk warning model building device provided in Embodiment 2 of the present invention;

[0028] Figure 5 It is a structural schematic diagram of an electronic device for implementing a risk warning model building method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0030] It should be noted that the term "target" and the like in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] Embodiment 1

[0032] Figure 1 This is a flow chart of a risk warning model construction method provided according to the first embodiment of the present invention. This embodiment is applicable to risk warnings for the safety of high-altitude workers. The method can be executed by a risk warning model construction device. The risk warning model construction device can be implemented in the form of hardware and / or software. The risk warning model construction device can be configured in a device. For example, the device can be a background server or other device with communication and computing capabilities. Figure 1 As shown, the method includes:

[0033] S110, acquiring a target image.

[0034] In this solution, the target image may refer to a video image of a person working at height, and a video may be shot using a shooting device to obtain multiple target images. The shooting device may be a video camera, a still camera, or the like.

[0035] S120: Detect the target image to determine the target object and the moving speed of the target object.

[0036] Among them, the target object may refer to personnel performing high-altitude operations.

[0037] In this embodiment, target detection technology can be used to detect the target image and determine the target object in the target image. The target image can be detected continuously to determine the moving speed of the target object. The target detection technology can include neural network, deep learning, etc.

[0038] Optionally, detecting the target image to determine the target object and the moving speed of the target object includes:

[0039] The target image is detected using a small target detection technology to determine the target object and the moving speed of the target object.

[0040] Among them, small target detection technology is an advanced image processing technology that aims to improve the monitoring system's ability to identify small targets at long distances, and is used to identify personnel working at long distances and at high altitudes.

[0041] Optionally, the target image is detected using a small target detection technology to determine the target object and the moving speed of the target object, including:

[0042] Performing gridding processing on the target image to obtain a processed target image;

[0043] Performing feature extraction on the processed target image to obtain target features;

[0044] The target features are predicted to determine the target object and the moving speed of the target object.

[0045] In this program, Figure 2 A schematic diagram of the risk warning model construction process provided in Example 1 of the present application is shown in FIG. Figure 2 As shown in the figure, small target detection technology includes image gridding, feature extraction and target prediction. These steps allow the system to accurately detect and track high-altitude workers from video surveillance, even if they occupy a small proportion of the screen. The video screen is gridded, human features are detected in each grid, and target prediction is performed to achieve the identification of high-altitude workers at a distance.

[0046] Specifically, image gridding includes: the system divides the real-time video stream, i.e., multiple target images, into multiple small areas or grids, and each grid is used as an independent analysis unit, i.e., the processed target image. By comparing the differences between adjacent frames, the system can effectively reduce the impact of background noise and lighting changes; gridding also helps to centralize computing resources, because each grid can be processed independently, thereby reducing the burden on the central processor.

[0047] Furthermore, human feature extraction includes: using computer vision algorithms, the system extracts key features that may contain human bodies from each grid, such as shape information such as the head, shoulders, and torso. These features are further refined and enhanced to make them easier to identify in the subsequent target prediction step. In order to improve accuracy, the system may use pre-trained models to assist in feature extraction, which have been trained on a large amount of labeled data.

[0048] In this solution, target prediction includes: after feature extraction, the system uses a pattern recognition algorithm to predict the target object in each grid, as well as the target position and motion trajectory of the target object. The target recognition algorithm can be a support vector machine (SVM), a neural network, etc. For moving target objects, the system combines the information of continuous frames to track its motion state and update the target's position information. If the system detects a new target object entering the monitoring range, it will mark it as a new target and start tracking it.

[0049] Furthermore, the moving speed of the target object is achieved by the following steps:

[0050] 1. Key point detection: Use target detection algorithms (such as YOLO, SSD, etc.) to detect the position information of key points of the human body in the video frame, such as the head, shoulders, elbows, wrists, knees, etc.

[0051] 2. Posture estimation: Using key point position information, the relative position and posture of each part of the human body can be estimated through geometric relationships or regression models. For example, the degree of arm bending can be estimated through the position of the shoulder and elbow.

[0052] 3. Motion trajectory tracking: Track the motion trajectory of key points in the video, for example, record the position changes of key points such as head and shoulders in consecutive frames.

[0053] 4. Speed ​​calculation: Calculate the moving speed of the target object based on the motion trajectory of the key points. Common speed calculation methods include:

[0054] Average speed:

[0055] Among them, v represents the velocity, Δx represents the change in the position of the key point, and Δt represents the time interval.

[0056] Instantaneous speed: When Δt->0;

[0057] Among them, v(t) represents the instantaneous velocity, Δx represents the change in the position of the key point, and Δt represents the time interval approaching 0.

[0058] Velocity vector:

[0059] Among them, v represents the velocity vector, Δx and Δy represent the position changes of the key points in the horizontal and vertical directions respectively, and Δt represents the time interval.

[0060] Furthermore, small target detection technology is used to identify people working at high altitudes at a distance. Small target detection technology is used to identify people working at high altitudes at a distance. This includes adjusting the focal length and angle of the camera, and optimizing the image processing algorithm so that it can still accurately identify people under low-resolution conditions. Specifically, it includes:

[0061] (1) Data collection and preprocessing:

[0062] Collect video data of people working at high altitudes at long distances, covering different weather conditions, lighting environments, and background complexity. Perform preprocessing on the video data, including denoising, contrast enhancement, brightness adjustment, etc., to improve the recognizability of the target.

[0063] (2) Algorithm selection and training:

[0064] Select an appropriate small target detection algorithm, such as a deep learning-based method (e.g., YOLO, SSD, Faster R-CNN, etc.). Use an annotated dataset to train the selected algorithm to ensure that the model can learn how to identify small target features at a distance.

[0065] (3) Feature extraction and model optimization:

[0066] In the feature extraction stage, the system needs to be able to extract key features of the human body from the video frame, such as the head, shoulders, torso, etc. The model is optimized according to the actual application scenario, such as adjusting the size and shape of the anchor box to better adapt to the characteristics of distant targets.

[0067] (4) Real-time detection and tracking:

[0068] Implement a detection system that can run on a real-time video stream, which should be able to quickly and accurately identify new target objects that enter the field of view. For moving targets, the system needs to be able to track between consecutive frames to maintain continuous monitoring of the target.

[0069] (5) System integration and testing:

[0070] Integrate the trained model into the monitoring system to ensure that the system can operate stably in various environments. Perform comprehensive tests on the system in the actual environment to evaluate its performance and stability under real conditions.

[0071] (6) User interface and feedback:

[0072] Develop a user-friendly interface that enables operators to easily configure and use the small object detection system. Provide a feedback mechanism that allows users to report problems and make suggestions for improvements in order to continuously optimize system performance.

[0073] Through the small target detection algorithm, accurate identification of people working at high altitudes at long distances is achieved, which greatly improves the monitoring capability of the monitoring system at long distances. Even in complex backgrounds and extreme lighting conditions, the key features of the workers can be clearly captured, thereby ensuring the effectiveness of work safety supervision.

[0074] S130: Recognize the posture of the target object to obtain the target posture.

[0075] In this solution, an image processing algorithm may be used to recognize the posture of the target object, thereby determining the target posture of the target object.

[0076] Optionally, recognizing the posture of the target object to obtain the target posture includes:

[0077] Construct a high-altitude human posture database;

[0078] The posture of the target object is recognized based on the high-altitude human posture database to obtain the target posture.

[0079] Specifically, Figure 2 As shown in the figure, the high-altitude human posture database is based on the key nodes of the human skeleton to build an action network and classify actions. By building a database containing a variety of high-altitude working postures, the key skeleton node information can be used to identify the specific actions of high-altitude workers.

[0080] The method based on the high-altitude human posture database provides accurate classification and recognition of the movements of high-altitude workers. This database contains rich human movement information, enabling the system to learn and understand various complex working postures, and accurately classify and recognize them on this basis. This not only improves work efficiency, but also helps to monitor the safety of workers in real time and prevent potential risks.

[0081] Optionally, build a high-altitude human posture database, including:

[0082] Get posture image;

[0083] Annotating the posture image to determine key nodes of the human skeleton;

[0084] Based on the key nodes of the human skeleton, an action network is constructed, and based on the action network, a high-altitude human posture database is constructed.

[0085] In this solution, the high-altitude human posture database is used to construct an action network and classify actions based on the key nodes of the human skeleton. It involves capturing the key skeletal node data of people in different working postures and using these data to build an action network. Based on this network, the system can classify and identify the actions of the workers. The high-altitude human posture database is a key resource that provides the monitoring system with a basis for identifying and understanding the actions of workers. The construction of this database involves a lot of data collection, processing and analysis work. The specific steps include:

[0086] (1) Data collection:

[0087] First, we need to collect posture images of aerial workers at different heights, angles, and lighting conditions. These images should contain a wide range of human actions, such as climbing, standing, and moving, to ensure that the database covers a variety of work scenarios. In order to improve the diversity and coverage of the data, we can invite actual workers to participate in the shooting, or use virtual reality technology to generate simulated actions.

[0088] (2) Data annotation:

[0089] The collected posture images need to be annotated manually or semi-automatically to mark the key skeletal nodes of the human body, such as the head, shoulders, elbows, wrists, torso, knees and ankles. The position of each joint point should be ensured to be accurate during the annotation process, because this information will be used for subsequent action network construction and classification.

[0090] (3) Action network construction:

[0091] Using the labeled data, we develop an action network based on the key nodes of the human skeleton. This network can capture the spatial relationship and movement patterns between the various parts of the human body. The design of the action network should take into account the naturalness and coherence of human movement so that it can accurately describe and predict changes in human posture.

[0092] (4) Action classification:

[0093] Based on the action network, a classification algorithm is further developed to classify the actions in the database into different categories, such as normal working posture, dangerous actions (such as not wearing a seat belt), resting state, etc. The classification algorithm can be a supervised learning model, which is trained to learn how to judge the action type based on the input posture data. In order to improve the accuracy of classification, deep learning techniques such as convolutional neural networks (CNN) or recurrent neural networks (RNN) can be used. These models can be trained end-to-end on a large amount of data.

[0094] (5) Database application:

[0095] After completing the above steps, the high-altitude human posture database can be integrated into the monitoring system for real-time or offline posture analysis and motion recognition. The system can use the information in the database to assist the small target detection algorithm and the high-fall safety risk warning model, thereby improving the accuracy and timeliness of monitoring.

[0096] Furthermore, the method based on the high-altitude human posture database is used to achieve accurate classification and recognition of the operator's actions, so that the system can accurately classify and recognize the operator's actions based on the captured target image, and maintain high accuracy even under complex or changing background conditions. The method based on the high-altitude human posture database is a technology that uses pre-collected and labeled human motion data to achieve accurate classification and recognition of operator actions. Specifically, it includes:

[0097] (1) Data collection and annotation:

[0098] Collect image or video datasets containing various actions of people working at heights, covering different heights, angles, and lighting conditions. Annotate each action in the dataset in detail, including the location of key bone nodes of the human body and the type of action.

[0099] (2) Database construction:

[0100] Use the labeled data to build a high-altitude human posture database, which should contain rich action types and detailed human posture information. The database can be stored in a structured way, such as XML or JSON format, to facilitate subsequent queries and applications.

[0101] (3) Algorithm design and training:

[0102] Design machine learning or deep learning algorithms suitable for human posture recognition, such as convolutional neural networks (CNN), recurrent neural networks (RNN), or long short-term memory networks (LSTM). Use data from the high-altitude human posture database to train the designed algorithm so that the model can learn how to recognize and classify different actions.

[0103] (4) Real-time action recognition:

[0104] Implement an action recognition system that can run on real-time video streams. The system should be able to quickly and accurately identify human actions in the current monitoring screen. For moving targets, the system has the ability to track between consecutive frames to maintain continuous monitoring of the target.

[0105] (5) System integration and testing:

[0106] Integrate the trained model into the monitoring system to ensure that the system can operate stably in various environments. Perform comprehensive tests on the system in the actual environment to evaluate its performance and stability under real conditions.

[0107] (6) User interface and feedback:

[0108] Develop a user-friendly interface that enables operators to easily configure and use the method based on the high-altitude human posture database. Provide a feedback mechanism to allow users to report problems and make suggestions for improvements in order to continuously optimize system performance.

[0109] (7) Early warning and response:

[0110] If dangerous actions or behaviors that do not meet safety regulations are detected, the system should immediately trigger a safety warning, ensuring that the warning signal can be quickly conveyed to operators and managers through on-site broadcasting or other communication means.

[0111] The method based on the high-altitude human posture database provides accurate classification and recognition of the movements of high-altitude workers. This database contains rich human movement information, enabling the system to learn and understand various complex working postures, and accurately classify and recognize them on this basis. This not only improves work efficiency, but also helps to monitor the safety of workers in real time and prevent potential risks.

[0112] S140: Identify the wearing status of the safety measures of the target object and obtain a wearing result.

[0113] Among them, the wearing of safety measures may refer to the situation in which workers working at heights wear safety ropes or safety belts.

[0114] Specifically, the wearing of safety measures by workers working at heights can be detected in real time by designing a network model.

[0115] Optionally, identifying the wearing status of the safety measures of the target object to obtain the wearing result includes:

[0116] The wearing status of the safety measures of the target object is identified based on the YOLO network model to obtain the wearing result.

[0117] In this scheme, if Figure 2 As shown in the figure, the network framework based on the YOLO network model is used for deep learning detection of safety belts and safety ropes for aerial work. This framework is trained to identify and detect whether the operator is wearing a safety belt or safety rope, and can accurately perform this task in various environments and lighting conditions. The YOLO (You Only Look Once) network model is a popular deep learning framework that is designed to handle real-time target detection problems in images. Figure 3This is a schematic diagram of the deep learning detection of high-altitude work safety belts and safety ropes based on the network framework of the YOLO network model provided by the implementation of this application, such as Figure 3 As shown in the figure, in the detection application of high-altitude work safety belts and safety ropes, the YOLO network model includes the following key steps:

[0118] (1) Dataset preparation:

[0119] Collect a dataset of images of people working at heights wearing safety belts and safety ropes. This dataset should include different lighting conditions, background complexity, and changes in the posture of the person. Annotate the data to ensure that the safety belts and safety ropes in each image are accurately marked, and indicate whether they are worn correctly.

[0120] (2) Model selection and adjustment:

[0121] Choose a YOLO version suitable for target detection tasks, such as YOLOv3, YOLOv4, or YOLOv5. Adjust the network architecture according to the characteristics of high-altitude operations, such as adding convolutional layers to handle the detection of small targets at a distance, or adjusting the size and shape of the anchor box to better fit the shape of the safety belt and safety rope.

[0122] (3) Training and validation:

[0123] The selected YOLO network model is trained using the prepared dataset. During the training process, the model will learn how to identify and locate the seat belts and safety ropes in the image. In order to improve the accuracy of detection, data augmentation techniques such as rotation, scaling, shearing, etc. can be used to expand the training set and improve the generalization ability of the model.

[0124] (4) Testing and evaluation:

[0125] Evaluate the performance of the model on an independent test set, using metrics such as precision, recall, and F1 score to measure the accuracy and reliability of the model. If the performance does not meet the expected standards, you can further adjust the network parameters or increase the amount of training data for optimization.

[0126] (5) Integration and deployment:

[0127] Integrate the trained YOLO model into the monitoring system, so that it can analyze video streams or static images in real time and detect whether the workers are wearing safety belts and safety ropes. Ensure that the system can operate stably in various environments and issue early warning signals for detected problems in a timely manner.

[0128] Furthermore, the safety belt / rope deep learning method is used to realize the automatic detection of safety measures, and develop and train deep learning methods specifically for detecting safety belts and safety ropes. This includes collecting a large number of annotated safety belt and safety rope image samples, and designing and adjusting the network to maximize the detection accuracy. The safety belt / rope deep learning method is a method that uses deep learning technology to automatically detect whether high-altitude workers are wearing safety belts and safety ropes. Specifically including:

[0129] (1) Data preparation and annotation:

[0130] Collect video or image datasets of people working at heights wearing safety belts and safety ropes. These data should cover various lighting conditions, background complexity, and changes in people's postures. Annotate the dataset in detail to ensure that the safety belts and safety ropes in each picture are accurately marked and indicate whether they are worn correctly.

[0131] (2) Model selection and training:

[0132] Select a suitable deep learning model as a basis, such as a convolutional neural network (CNN) or a recurrent neural network (RNN), to process image or video data. Use the labeled dataset to train the selected model so that the model can learn how to recognize and classify the wearing of seat belts and safety ropes.

[0133] (3) Feature extraction and optimization:

[0134] During the model training process, key features are extracted, such as the position, shape, color of the safety belt, and the length and direction of the safety rope. The model is optimized according to the actual application scenario, such as adjusting the network structure to improve the recognition ability of small targets at a long distance.

[0135] (4) Real-time detection and analysis:

[0136] Implement a detection system that can run on a real-time video stream and can quickly and accurately identify new seat belts and safety ropes that enter the field of view. For moving targets, the system needs to be able to track between consecutive frames to maintain continuous monitoring of the target.

[0137] (5) System integration and testing:

[0138] Integrate the trained model into the monitoring system to ensure that the system can operate stably in various environments. Perform comprehensive tests on the system in the actual environment to evaluate its performance and stability under real conditions.

[0139] (6) User interface and feedback:

[0140] Develop a user-friendly interface that enables operators to easily configure and use the belt / rope deep learning approach. Provide a feedback mechanism that allows users to report issues and suggest improvements in order to continuously optimize system performance.

[0141] (7) Early warning and response:

[0142] If insufficient safety measures such as not wearing a safety belt or safety rope are detected, the system should immediately trigger a safety warning. Ensure that the warning signal can be quickly conveyed to operators and managers through on-site broadcasting or other communication means.

[0143] Through such program refinement, the safety belt / rope deep learning method will be able to effectively improve the monitoring system's ability to automatically detect the safety measures of high-altitude workers, and can effectively identify whether the safety equipment is worn correctly and its fixation, thereby improving the accuracy and timeliness of safety supervision. This is crucial to protecting the lives of workers because it can provide early warnings at critical moments to prevent accidents.

[0144] S150: construct a high-fall prevention safety risk warning model according to the moving speed, target posture and wearing result of the target object.

[0145] Among them, the high-fall prevention safety risk warning model integrates the personnel's movement speed, the wearing of safety measures and the personnel's movement posture information, which can realize the rapid analysis and dynamic warning of potential risks.

[0146] Optionally, build a high-fall prevention safety risk warning model, including:

[0147] Determine the early warning mechanism based on the risk assessment results of the height fall prevention safety risk early warning model.

[0148] In this scheme, if Figure 2 As shown in the figure, the high-fall prevention safety risk warning model integrates information such as personnel movement speed, safety measures worn, and personnel movement posture. Construct and integrate different data inputs (such as personnel movement speed, safety belt wearing, and personnel movement posture) into a unified high-fall prevention safety risk warning model. This model can assess the risk level faced by operators and trigger warnings when potential dangerous situations are detected. The high-fall prevention safety risk warning model is a comprehensive system that predicts and prevents high-altitude fall risks in high-altitude operations by analyzing a variety of information. Specifically including:

[0149] (1) Data collection and integration:

[0150] Collect data on workers' movement speed, posture changes, and safety belt wearing during aerial work. This data can come from video surveillance, sensors, or other monitoring equipment. Integrate the collected data to ensure that data from different sources and types can be processed and analyzed uniformly.

[0151] (2) Feature extraction:

[0152] The integrated data is processed to extract key features, such as the person's position, movement speed, acceleration, posture angle, etc. For the wearing of seat belts, image recognition technology can be used to detect whether the seat belts are worn completely and correctly fastened.

[0153] (3) Model construction:

[0154] Use machine learning algorithms (such as support vector machines, decision trees, neural networks, etc.) to build an early warning model. This model should be able to predict potential high-fall risks based on the input feature data. The training of the model requires a large amount of historical data as a training set to help the algorithm learn how to identify normal operating behaviors and risky behaviors that may lead to high falls.

[0155] (4) Risk assessment:

[0156] Based on the risk assessment results output by the model, corresponding warning thresholds are formulated. For example, if a worker moves faster than normal or does not wear a safety belt completely, the system will mark them as high-risk. The warning model can also consider environmental factors such as wind speed and weather conditions to improve the accuracy of the warning.

[0157] (5) Early warning mechanism:

[0158] When the system detects high-risk behavior, it should immediately trigger an early warning signal to notify on-site management personnel or the security monitoring system. The early warning signal can be sent through a variety of methods such as sound and light alarms, SMS notifications, and mobile application push notifications to ensure that it is conveyed to relevant personnel in a timely manner.

[0159] (6) System testing and optimization:

[0160] The early warning system is tested in a real environment to verify its effectiveness and reliability.

[0161] The model is adjusted and optimized based on test results and user feedback to improve warning accuracy and reduce false alarm rates.

[0162] Furthermore, dynamic warnings can be used to alert operators through on-site broadcasting. When the system identifies a high-fall risk through the warning model, it will issue a call through the on-site broadcasting system to promptly remind operators to pay attention to safety and take necessary preventive measures. When the dynamic warning system detects potential safety risks, it can alert operators through on-site broadcasting. Specifically, it includes:

[0163] (1) Warning triggering conditions:

[0164] Determine the conditions for triggering the warning, which may include abnormal speed of personnel movement, abnormal posture recognition, improper wearing of safety belts / ropes, etc. The warning system should be able to automatically trigger the warning mechanism when the preset risk threshold is reached.

[0165] (2) Speech Generation and Synthesis:

[0166] Use speech synthesis technology (such as text-to-speech TTS) to generate warning voice messages. This requires pre-recording or synthesizing a series of standardized voice prompts to adapt to different warning contents. According to different warning levels and types, select the corresponding voice prompts to ensure the accuracy and timeliness of the information.

[0167] (3) Broadcasting equipment deployment:

[0168] Ensure that there are enough broadcasting equipment on site, such as megaphones, horns, or mini speakers integrated into hard hats, to cover the entire work area. Test the volume and clarity of the broadcasting equipment to ensure that the warning information can be clearly conveyed in various noise environments.

[0169] (4) Sending of early warning signals:

[0170] When the early warning system detects a potential safety risk and meets the triggering conditions, the early warning voice will be played immediately through the broadcasting equipment. The early warning voice should be concise and clear, directly point out the existing risks and provide corresponding safety suggestions or emergency measures.

[0171] (5) Manual intervention options:

[0172] Provides the option of manual intervention, allowing site managers to manually activate the early warning system when necessary. In high-risk situations, both manual and automatic methods can be used to issue early warnings to enhance the effectiveness of information delivery.

[0173] (6) Training and drills:

[0174] Train operators to familiarize them with the working principle of the warning system and the meaning of warning signals. Conduct drills regularly to ensure that all personnel can react quickly and take correct safety measures in real situations.

[0175] (7) Feedback and Optimization:

[0176] Collect feedback from operators on the early warning system to understand its effectiveness and problems in actual operation. Adjust and optimize the system based on the feedback results to improve the reliability and effectiveness of the early warning system.

[0177] Furthermore, the high-fall prevention safety risk warning model is used to achieve comprehensive assessment and early warning of risks. Construct a comprehensive high-fall prevention safety risk warning model, which not only combines multiple data inputs, but also takes into account external environmental factors and historical risk data to generate more accurate risk assessment and early warning signals. Use information such as personnel movement speed and safety belt wearing conditions to achieve comprehensive assessment and early warning of risks. The high-fall prevention safety risk warning model is a comprehensive system that predicts and prevents high-altitude fall risks in high-altitude operations by analyzing multiple information. Specifically including:

[0178] (1) Data collection and integration:

[0179] Collect various data of operators during high-altitude operations, including movement speed, posture changes, safety belt wearing conditions, etc. Integrate these data to ensure that data from different sources and types can be processed and analyzed uniformly.

[0180] (2) Feature extraction:

[0181] The integrated data is processed to extract key features, such as the person's position, movement speed, acceleration, posture angle, etc. For the wearing of seat belts, image recognition technology can be used to detect whether the seat belts are worn completely and correctly fastened.

[0182] (3) Risk assessment model construction:

[0183] Build a risk assessment model using machine learning algorithms (such as support vector machines, decision trees, neural networks, etc.). This model should be able to predict the potential risk of falling from height based on the input feature data.

[0184] (4) Warning threshold setting:

[0185] Based on the risk assessment results output by the model, corresponding warning thresholds are formulated. For example, if a worker moves faster than normal or does not wear a safety belt completely, the system will mark them as high-risk. The warning model can also consider environmental factors such as wind speed and weather conditions to improve the accuracy of the warning.

[0186] (5) Implementation of early warning mechanism:

[0187] When the system detects high-risk behavior, it should immediately trigger an early warning signal to notify on-site management personnel or the security monitoring system. The early warning signal can be sent through a variety of methods such as sound and light alarms, SMS notifications, and mobile application push notifications to ensure that it is communicated to relevant personnel in a timely manner.

[0188] (6) Testing and optimization:

[0189] Test the early warning system in a real environment to verify its effectiveness and reliability. Adjust and optimize the model based on the test results and user feedback to improve the accuracy of early warnings and reduce false alarms.

[0190] (7) Training and drills:

[0191] Train operators to familiarize them with the working principle of the warning system and the meaning of warning signals. Conduct drills regularly to ensure that all personnel can react quickly and take correct safety measures in real situations.

[0192] The built anti-fall safety risk warning model realizes comprehensive assessment and warning of risks. The model integrates multiple data inputs and advanced analysis technologies to monitor the working environment in real time and predict potential risks. When the system detects any abnormal situation, it will automatically trigger a warning signal to notify on-site managers to take measures to prevent potential accidents. This comprehensive assessment and warning capability significantly improves the safety of high-altitude operations and reduces the possibility of accidents.

[0193] The technical solution of the embodiment of the present invention obtains a target image, detects the target image, determines the target object and the moving speed of the target object, identifies the posture of the target object to obtain the target posture, and identifies the wearing of the safety measures of the target object to obtain the wearing result, and then constructs a high-fall prevention safety risk warning model according to the moving speed of the target object, the target posture and the wearing result. By executing this technical solution, automatic tracking and monitoring of the implementation of safety protection measures for high-altitude workers and safety risk prediction are realized.

[0194] Embodiment 2

[0195] Figure 4 This is a schematic diagram of the structure of a risk warning model construction device provided in the second embodiment of the present invention. Figure 4 As shown, the device comprises:

[0196] The target image acquisition module 410 is used to acquire the target image;

[0197] A detection module 420 is used to detect the target image and determine the target object and the moving speed of the target object;

[0198] A posture recognition module 430 is used to recognize the posture of the target object to obtain a target posture;

[0199] A safety measure wearing condition identification module 440 is used to identify the safety measure wearing condition of the target object and obtain a wearing result;

[0200] The anti-fall safety risk warning model construction module 450 is used to construct an anti-fall safety risk warning model according to the moving speed, target posture and wearing result of the target object.

[0201] Optionally, the detection module 420 includes:

[0202] The detection unit is used to detect the target image using a small target detection technology to determine the target object and the moving speed of the target object.

[0203] Optionally, a detection unit is used to:

[0204] Performing gridding processing on the target image to obtain a processed target image;

[0205] Performing feature extraction on the processed target image to obtain target features;

[0206] The target features are predicted to determine the target object and the moving speed of the target object.

[0207] Optionally, the gesture recognition module 430 includes:

[0208] A high-altitude human body posture database construction unit, used for constructing a high-altitude human body posture database;

[0209] The posture recognition unit is used to recognize the posture of the target object based on the high-altitude human posture database to obtain the target posture.

[0210] Optional, high-altitude human posture database construction unit, specifically used for:

[0211] Get posture image;

[0212] Annotating the posture image to determine key nodes of the human skeleton;

[0213] Based on the key nodes of the human skeleton, an action network is constructed, and based on the action network, a high-altitude human posture database is constructed.

[0214] Optionally, the safety measure wearing condition identification module 440 is specifically used for:

[0215] The wearing status of the safety measures of the target object is identified based on the YOLO network model to obtain the wearing result.

[0216] Optionally, the anti-fall safety risk warning model building module 450 is specifically used for:

[0217] Determine the early warning mechanism based on the risk assessment results of the height fall prevention safety risk early warning model.

[0218] A risk warning model construction device provided in an embodiment of the present invention can execute a risk warning model construction method provided in any embodiment of the present invention, and has functional modules and beneficial effects corresponding to the execution method.

[0219] Embodiment 3

[0220] Figure 5 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0221] like Figure 5 As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0222] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0223] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 11 executes the various methods and processes described above, such as a risk warning model construction method.

[0224] In some embodiments, a risk warning model construction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the risk warning model construction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute a risk warning model construction method in any other appropriate manner (e.g., by means of firmware).

[0225] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0226] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0227] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0228] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0229] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0230] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0231] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0232] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for constructing a risk warning model, characterized in that: include: Get the target image; Detecting the target image to determine the target object and the moving speed of the target object; Recognizing the posture of the target object to obtain the target posture; Identify the wearing status of the safety measures of the target object and obtain a wearing result; A high-fall prevention safety risk warning model is constructed according to the moving speed, target posture and wearing result of the target object.

2. The method according to claim 1, characterized in that Detecting the target image to determine the target object and the moving speed of the target object includes: The target image is detected using a small target detection technology to determine the target object and the moving speed of the target object.

3. The method according to claim 2, characterized in that The target image is detected using a small target detection technology to determine the target object and the moving speed of the target object, including: Performing gridding processing on the target image to obtain a processed target image; Performing feature extraction on the processed target image to obtain target features; The target features are predicted to determine the target object and the moving speed of the target object.

4. The method according to claim 1, characterized in that: Recognizing the posture of the target object to obtain the target posture includes: Construct a high-altitude human posture database; The posture of the target object is recognized based on the high-altitude human posture database to obtain the target posture.

5. The method according to claim 4, characterized in that Construct a high-altitude human posture database, including: Get posture image; Annotating the posture image to determine key nodes of the human skeleton; Based on the key nodes of the human skeleton, an action network is constructed, and based on the action network, a high-altitude human posture database is constructed.

6. The method according to claim 1, characterized in that Identify the wearing status of the safety measures of the target object and obtain the wearing result, including: The wearing status of the safety measures of the target object is identified based on the YOLO network model to obtain the wearing result.

7. The method according to claim 1, characterized in that Construct a high-fall prevention safety risk warning model, including: Determine the early warning mechanism based on the risk assessment results of the height fall prevention safety risk early warning model.

8. A risk warning model construction device, characterized in that: include: A target image acquisition module, used for acquiring a target image; A detection module, used to detect the target image and determine the target object and the moving speed of the target object; A posture recognition module, used to recognize the posture of the target object to obtain the target posture; A safety measure wearing condition identification module is used to identify the safety measure wearing condition of the target object and obtain a wearing result; The anti-fall safety risk warning model construction module is used to construct an anti-fall safety risk warning model according to the moving speed, target posture and wearing result of the target object.

9. An electronic device, characterized in that: The electronic device comprises: At least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute a risk warning model construction method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement a risk warning model construction method according to any one of claims 1 to 7 when executed.

Images