High-altitude operation risk early warning method and device and electronic equipment
Through the camera equipment, the position and posture of high-altitude workers are identified and the moving position is predicted, which solves the problem of low safety and accuracy of high-altitude operations under manual supervision, and improves the accuracy and timeliness of risk monitoring.
Patent Information
- Application Number
- CN202411052161.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-01
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2044-08-01
AI Technical Summary
Safety supervision of high-altitude operations in the prior art mainly relies on manual methods, with low efficiency, incomplete coverage, susceptible to subjective factors, and slow response speed, resulting in low safety and accuracy.
Obtain the image pictures of the target user through the camera device, identify their position and posture, predict the moving position, judge the operation risks, and send early warning information in real time to improve the accuracy and timeliness of risk monitoring.
It has achieved the accuracy and timeliness of risk monitoring for high-altitude workers, provided timely and effective safety guarantees, and reduced the defects of manual supervision.
Smart Images

Figure CN119028081A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of artificial intelligence, and more specifically, to a risk warning method, device, and electronic equipment for high-altitude operations. Background Art
[0002] In the field of power transmission, aerial work is an indispensable part, which covers a series of key tasks such as line erection, equipment maintenance, troubleshooting, etc., which is crucial to ensure the stable operation of the power grid. Aerial workers need to perform delicate operations at a height of tens or even hundreds of meters, which not only tests their professional skills, but also places extremely high demands on their safety. However, it is precisely such a working environment that makes aerial work face many potential risks, such as falling from height, electric shock injuries, and the impact of bad weather, each of which may pose a serious threat to the life safety of the workers.
[0003] Traditionally, the safety supervision of aerial work mainly relies on manual methods, but this supervision model has obvious defects. First, manual supervision is inefficient and it is difficult to achieve comprehensive coverage of vast working areas and long working processes, which makes it difficult to timely discover and deal with safety hazards. Secondly, the supervision effect is easily affected by subjective factors, such as the experience, concentration, and judgment of the supervisors, which may affect the accurate assessment and timely response to safety risks. In addition, in the face of emergencies, the response speed of manual supervision is often not fast enough to provide timely and effective safety protection for aerial workers.
[0004] Currently, no effective solution has been proposed to address the problem of low safety and accuracy in supervising workers working at heights through manual supervision in related technologies. Summary of the invention
[0005] The present application provides a risk warning method, device and electronic equipment for high-altitude operations to solve the problem of low safety and accuracy in supervising high-altitude workers through manual supervision in related technologies.
[0006] According to one aspect of the present application, a risk warning method for high-altitude operations is provided. The method includes: obtaining a target image of a target user through a camera device, and identifying the target position of the target user in the target image; inputting the target image into a posture recognition model to obtain the target posture information of the target user; predicting the target moving position of the target user according to the target position and the target posture information, and determining the operation risk warning information of the target user according to the target moving position, and sending the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk.
[0007] Optionally, obtaining the target image of the target user through the camera device includes: obtaining image pictures taken by multiple cameras to obtain multiple initial image pictures, and screening image pictures containing the target user from the multiple initial image pictures to obtain multiple candidate image pictures; obtaining the resolution level of each candidate image picture and the size ratio of the target user to the candidate image picture, wherein the resolution level and the resolution size are in direct proportion; performing weighted summation of the resolution level and size ratio of each candidate image picture according to preset weights to obtain a score for each candidate image picture; and determining the candidate image picture with the highest score as the target image picture.
[0008] Optionally, identifying the target position of the target user in the target image picture includes: inputting the target image picture into a position recognition model to obtain the target position of the target user, wherein the target position is determined by the boundary point coordinates of the target user's character frame, and the position recognition model is trained by a first sample data set, wherein the first sample data set includes multiple first sample data, each first sample data includes a historical image picture, a character frame in the historical image picture, and the boundary point coordinates of the character frame.
[0009] Optionally, the posture recognition model is trained in the following manner: obtaining multiple second sample information, wherein each second sample information includes a human body image and posture information corresponding to the human body image; training the initial recognition model according to the multiple second sample information to obtain the posture recognition model.
[0010] Optionally, predicting the target moving position of the target user based on the target position and target posture information includes: determining the moving acceleration of the target user based on the target posture information; obtaining multiple historical moving positions of the target user, and calculating the multiple historical moving positions, target positions and moving accelerations based on a Kalman filter algorithm to obtain a predicted moving trajectory of the target user; determining the target moving position of the target user based on the predicted moving trajectory.
[0011] Optionally, determining the operation risk warning information of the target user based on the target moving position includes: judging whether the target moving position is in a preset area, wherein the preset area is a dangerous area; when the target moving position is in the preset area, determining the operation risk warning information as high-risk warning information; when the target moving position is not in the preset area, judging whether the predicted moving trajectory passes through the preset area; when the predicted moving trajectory passes through the preset area, determining the operation risk warning information as medium-risk warning information; when the predicted moving trajectory does not pass through the preset area, determining the operation risk warning information as low-risk warning information.
[0012] Optionally, after obtaining the target image of the target user through the camera device, the method also includes: inputting the target image into the equipment recognition model, determining whether the target user carries the target equipment through the equipment recognition model, and obtaining a recognition result, wherein the equipment recognition model is obtained by training multiple third sample data, each third sample data includes at least one equipment, and labeling information corresponding to each equipment; when the recognition result indicates that the target user does not carry the target wearable device, the operation risk warning information is determined as high-risk warning information; when the recognition result indicates that the target user carries the target wearable device, the operation risk warning information is determined as low-risk warning information.
[0013] According to another aspect of the present application, a risk warning device for high-altitude operations is provided. The device includes: a first recognition unit, which is used to obtain a target image of a target user through a camera device and identify the target position of the target user in the target image; a second recognition unit, which is used to input the target image into a posture recognition model to obtain the target posture information of the target user; a first determination unit, which is used to predict the target moving position of the target user based on the target position and the target posture information, and determine the operation risk warning information of the target user based on the target moving position, and send the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk.
[0014] According to another aspect of the present invention, a computer program product is provided, including a computer program, which, when executed by a processor, implements a risk warning method for high-altitude operations in the present application.
[0015] According to another aspect of the present invention, there is also provided an electronic device comprising one or more processors and a memory; the memory stores computer-readable instructions, and the processor is used to run the computer-readable instructions, wherein when the computer-readable instructions are run, a risk warning method for high-altitude operations is executed.
[0016] Through this application, the following steps are adopted: obtaining the target image of the target user through a camera device, and identifying the target position of the target user in the target image; inputting the target image into the posture recognition model to obtain the target posture information of the target user; predicting the target moving position of the target user according to the target position and the target posture information, and determining the operation risk warning information of the target user according to the target moving position, and sending the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk. The problem of low safety and accuracy of supervising high-altitude workers through manual supervision in the related technology is solved. By identifying the target position of the worker in the image picture and identifying the target posture of the worker, the moving position of the worker is predicted according to the target position and the target posture, so as to determine the position change of the worker in the future period of time, and judge whether there is a risk in the high-altitude operation according to the predicted moving position, thereby achieving the effect of improving the accuracy and timeliness of risk monitoring. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:
[0018] Figure 1 is a flow chart of a risk warning method for high-altitude operations provided in an embodiment of the present application;
[0019] Figure 2 is a flowchart of a method for selecting a picture according to an embodiment of the present application;
[0020] Figure 3 is a schematic diagram of a risk warning device for high-altitude operations provided according to an embodiment of the present application;
[0021] Figure 4 It is a schematic diagram of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION
[0022] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.
[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.
[0024] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. 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 comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0025] It should be noted that the risk warning method, device and electronic equipment for high-altitude operations determined in the present disclosure can be used in the field of artificial intelligence, and can also be used in any field except the field of artificial intelligence. The application field of the risk warning method, device and electronic equipment for high-altitude operations determined in the present disclosure is not limited.
[0026] It should be noted that the collected information, user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) used in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with the relevant laws, regulations and standards of the relevant regions, necessary confidentiality measures are taken, and public order and good customs are not violated. Corresponding operation portals are provided for users to choose to authorize or refuse to use. For example, an interface is set up between this system and relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.
[0027] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:
[0028] YOLOv5: YOLOv5 (You Only Look Once) is a deep learning model for object detection.
[0029] OpenPose network architecture: a framework for real-time human key point detection and full body pose estimation based on deep learning.
[0030] Heatmap: A heat map used to predict the location of key points on the human body. Each key point has a corresponding peak area on the heatmap. By finding the peak point on the heatmap, the specific location of the key point can be determined.
[0031] Kalman filter: is an efficient recursive filter (autoregressive filter) that can estimate the state of a dynamic system from a series of measurements containing statistical noise.
[0032] According to an embodiment of the present application, a risk warning method for high-altitude operations is provided.
[0033] Figure 1 1 is a flow chart of a risk warning method for high-altitude operations provided in accordance with an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:
[0034] Step S101 : acquiring a target image of a target user through a camera device, and identifying a target position of the target user in the target image.
[0035] Specifically, the operator may be an operator performing high-altitude operations. First, a camera, a camcorder or other imaging device may be used to shoot the scene to obtain a real-time image including the target user. For example, a high-resolution camera may be used to capture a video image of the high-altitude operation scene to obtain a target image including the target user.
[0036] Furthermore, after obtaining the target image, the content in the image can be analyzed by a pre-trained machine learning model or deep learning algorithm to identify the target user. After identifying the target user in the image, it is also necessary to determine the target position of the target user in the target image. When determining the target position of the target user, it can be achieved by outlining a human body frame of the target user in the target image, where the coordinates of the human body frame are usually represented by four numerical values, corresponding to the horizontal and vertical coordinates (x1, y1) of the upper left corner of the frame and the horizontal and vertical coordinates (x2, y2) of the lower right corner. The above two coordinates can define the specific position range of the target user in the image. Through the coordinates of the target position, not only can the location information of the target user be accurately determined, but also accurate location information can be provided for subsequent analysis (such as behavior analysis, path tracking, etc.).
[0037] Step S102: input the target image into a posture recognition model to obtain target posture information of the target user.
[0038] Specifically, after obtaining a target image containing the target user, the target image can be input into a posture recognition model, thereby analyzing the target user's body posture, joint position and other information in the target image through the posture recognition model to obtain the target user's posture information.
[0039] Among them, the posture recognition model is usually built based on deep learning technology, and the video image is input into the deep learning model to obtain the posture recognition result. The posture recognition model learns how to recognize the posture feature information of the human body in the image by training a large number of image training sets that are annotated with relevant human posture information such as the position of human joints, and recognizes the posture information based on the feature information. Among them, the posture recognition model uses multiple convolutional layers to extract image features, uses pooling layers to reduce the size of feature maps and retain important information, and uses fully connected layers for classification and regression, so as to obtain the position, category, and posture information of the human target.
[0040] For example, the posture recognition model can use the OpenPose network architecture and combine human key point detection and posture estimation algorithms to achieve accurate posture recognition. Here, the image needs to be input into the OpenPose network, the convolution layer and pooling layer are used to extract image features, and the heatmap regression is used to predict the position of the human key points. The key point connection algorithm is used to connect the detected human key points into skeletons, and posture estimation is performed to complete the posture estimation operation process.
[0041] Step S103, predicting the target moving position of the target user based on the target position and target posture information, and determining the target user's operation risk warning information based on the target moving position, and sending the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk.
[0042] Specifically, when the target position and target posture information of the target user are obtained, the possible future moving position of the target user can be inferred based on the target position and target posture information, thereby determining whether there is a risk in the high-altitude operation.
[0043] Among them, when predicting the mobile position, the prediction algorithm can be used to predict the target user's next movement path and possible target movement position based on the user's current position and posture. After predicting the target movement position of the target user, the operational risks that the target user may face can be evaluated based on this location information, and corresponding warning information can be generated.
[0044] Furthermore, when evaluating the operational risks that the target user may face, the predicted movement position can be compared with pre-defined risk areas or safety specifications. These risk areas may include dangerous operating areas, prohibited entry areas, high-risk operating areas, etc. Through comparison, the risks that the target user may encounter in future actions can be evaluated, and then the operational risk warning information can be determined, so that the target user can avoid risks in time according to the operational risk warning information, thereby achieving the technical effect of improving the safety of high-altitude operations. Among them, the operational risk warning information can be information, light indications, alarm bells, etc., which are not limited here.
[0045] The risk warning method for high-altitude operations provided in the embodiment of the present application obtains the target image of the target user through a camera device, and identifies the target position of the target user in the target image; inputs the target image into a posture recognition model to obtain the target posture information of the target user; predicts the target moving position of the target user according to the target position and the target posture information, and determines the operation risk warning information of the target user according to the target moving position, and sends the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk. The problem of low safety and accuracy in supervising high-altitude workers through manual supervision in the related art is solved. By identifying the target position of the operator in the image picture and identifying the target posture of the operator, the moving position of the operator is predicted according to the target position and the target posture, so as to determine the position change of the operator in the future period of time, and judge whether there is a risk in the high-altitude operation according to the predicted moving position, thereby achieving the effect of improving the accuracy and timeliness of risk monitoring.
[0046] Optionally, Figure 2 is a flowchart of a method for selecting a picture according to an embodiment of the present application. Figure 2 As shown, in the risk warning method for high-altitude operations provided in the embodiment of the present application, obtaining a target image of a target user through a camera device includes:
[0047] Step S201, acquiring images captured by a plurality of camera devices to obtain a plurality of initial images, and screening images containing a target user from the plurality of initial images to obtain a plurality of candidate images.
[0048] It should be noted that in order to ensure that the model can accurately identify the target position of the target user in the target image, the image content in the target image needs to be accurate and clear. Therefore, the initial image can be collected from multiple camera devices that can capture the target user, and the initial image is analyzed and screened to obtain the target image.
[0049] Specifically, first determine whether the initial image contains the image of the target user, and use image recognition technology to search and filter out images containing the target user in these initial images, thereby filtering out initial images in which the target user is too small or does not contain the target user, to ensure that the portrait of the target user in the obtained candidate image can be clearly viewed.
[0050] Step S202, obtaining the resolution level of each candidate image and the size ratio of the target user to the candidate image, wherein the resolution level and the resolution size are in direct proportion.
[0051] It should be noted that after screening and obtaining multiple candidate image pictures, it is also necessary to obtain the resolution level and target user ratio of the candidate image pictures, so as to select the target image picture from the multiple candidate image pictures according to the resolution level and the target user ratio. Among them, the resolution level can be determined according to the resolution level table, and the resolution level table can contain multiple resolution intervals. Different resolution intervals correspond to different resolution levels. The higher the resolution, the higher the level.
[0052] Specifically, the resolution of each candidate image can be obtained, and the resolution level of each candidate image can be determined according to a resolution level table, and the approximate outline size of the target user can be obtained, and the size ratio of the target user in the candidate image can be determined. For example, the target user occupies 5% of the area in the candidate image, thereby obtaining the size ratio of the target user in the candidate image.
[0053] Step S203 , performing weighted summation on the resolution level and size ratio of each candidate image picture according to a preset weight, to obtain a score for each candidate image picture.
[0054] Furthermore, after obtaining the resolution level and size ratio of each candidate image picture, the resolution level and size ratio of each candidate image picture can be weighted and summed according to preset weights to obtain a score for each candidate image picture, and then the target image picture is selected from the candidate image pictures according to the score.
[0055] Step S204: determine the candidate image picture with the highest score as the target image picture.
[0056] Specifically, after obtaining the score of each candidate image, the image with the highest score can be determined as the target image, and then the target image with the largest and clearest portrait of the target user can be obtained.
[0057] This embodiment improves the accuracy of identifying the target user's posture based on image information by screening images taken by multiple camera devices.
[0058] Optionally, in the risk warning method for high-altitude operations provided in an embodiment of the present application, identifying the target position of the target user in the target image picture includes: inputting the target image picture into a position recognition model to obtain the target position of the target user, wherein the target position is determined by the boundary point coordinates of the target user's character frame, and the position recognition model is obtained by training a first sample data set, wherein the first sample data set includes multiple first sample data, and each first sample data includes a historical image picture, a character frame in the historical image picture, and the boundary point coordinates of the character frame.
[0059] Specifically, when determining the target position of the target user, the target image can be input into a pre-trained position recognition model. The task of the position recognition model is to recognize the specific position of the target user from the input image, wherein the target position refers to the specific position of the target user in the image, and this position is determined by the boundary point coordinates of the target user's character frame (usually a rectangular frame used to mark and frame the character). The boundary point coordinates define the specific position and size of the character frame in the image, thereby indirectly indicating the position of the target user.
[0060] It should be noted that the location recognition model is trained by learning a large amount of historical data (i.e., the first sample data set). The location recognition model can automatically extract features from the image and predict or recognize the location of the target user in the image based on these features. The first sample data set consists of historical image pictures, each of which is marked with at least one character frame, and the boundary point coordinates of each character frame are given. These boundary point coordinates accurately define the position and size of the character frame in the picture, i.e., the position of the target user, so that the model can be trained using the first sample data set to obtain a location recognition model.
[0061] It should be noted that the location recognition model can use the YOLOv5 network architecture, input the image into the location recognition model, use multiple convolutional layers to extract image features, use the pooling layer to reduce the size of the feature map and keep important information, use the fully connected layer for classification and regression, and obtain the location information of the human target.
[0062] For example, the target position can be represented by the upper left corner coordinates and lower right corner coordinates of the character box, and the model output result can be: target position: upper left corner coordinates of the bounding box (x1=200, y1=300), lower right corner coordinates of the bounding box (x2=250, y2=350).
[0063] This embodiment achieves the technical effect of accurately identifying the target location of the target user by training the location recognition model.
[0064] In order to improve the accuracy of posture recognition, optionally, in the risk warning method for high-altitude operations provided in an embodiment of the present application, the posture recognition model is trained in the following manner: obtaining multiple second sample information, wherein each second sample information includes a human body image and posture information corresponding to the human body image; training the initial recognition model according to the multiple second sample information to obtain a posture recognition model.
[0065] Specifically, in order to ensure accurate acquisition of the target user's posture information, the posture recognition model needs to be trained. First, a large amount of second sample information needs to be collected as training data. Each second sample information contains two parts: a human image and the human posture information corresponding to the image. The human image can be a real human photo or video frame from different scenes, different angles, and different postures. The posture information is a detailed description of the human posture in the image, usually including the position of the key points of the human body (such as the head, shoulders, elbows, knees, etc.) and the relative relationship between these key points, as well as the name of the posture.
[0066] Furthermore, after collecting a plurality of second sample information, the collected plurality of second sample information can be used to train the initial recognition model to obtain a posture recognition model that can accurately recognize human postures. During the training process, the model will learn how to extract useful features from human images and establish a mapping relationship between these features and human posture information.
[0067] It should be noted that in order to achieve accurate posture recognition, OpenPose can be selected as the network architecture of the model. OpenPose is a posture estimation method based on deep learning. It can combine human key point detection and posture estimation algorithms to extract the position of human key points from images or videos in real time, and infer the posture of the human body based on this. The video image is input into the OpenPose network. First, the convolution layer and pooling layer are used to extract features of the image. The convolution layer slides on the image through the convolution kernel and calculates the convolution operation to extract local features in the image. The pooling layer downsamples the feature map output by the convolution layer to reduce the amount of calculation and extract higher-level features. The position of the key points of the human body is further predicted by heatmap regression. Then, the key point connection algorithm is used to connect these key points into bones to form the skeleton structure of the human body, so as to perform posture estimation based on the skeleton structure and infer the overall posture of the human body.
[0068] In order to ensure the accuracy of risk prediction, optionally, in the risk warning method for high-altitude operations provided in an embodiment of the present application, predicting the target moving position of the target user based on the target position and target posture information includes: determining the moving acceleration of the target user based on the target posture information; obtaining multiple historical moving positions of the target user, and calculating the multiple historical moving positions, target positions and moving accelerations according to the Kalman filtering algorithm to obtain the predicted moving trajectory of the target user; determining the target moving position of the target user based on the predicted moving trajectory.
[0069] Specifically, after obtaining the target posture information, the target posture information can be analyzed to obtain the target user's movement acceleration. In this step, the target posture information needs to be analyzed to obtain posture information, such as speed, direction, tilt angle and other parameters. These parameters can directly or indirectly reflect the user's motion state. By analyzing these posture information, the current movement acceleration of the target user can be calculated, and then the predicted movement trajectory of the target user can be predicted based on the movement acceleration.
[0070] For example, posture information can be recognized through a posture parameter recognition model, that is, the skeleton information, key point information, current posture and other related contents in the target posture information are input into the posture parameter recognition model, and the posture information is obtained through the posture parameter recognition model.
[0071] Furthermore, in order to more accurately predict the future movement trajectory of the target user, it is also necessary to obtain multiple historical movement positions of the target user in the past period of time, so that the multiple historical movement positions, target positions and movement accelerations can be calculated according to the Kalman filter algorithm to obtain the predicted movement trajectory of the target user. The Kalman filter algorithm predicts the next position based on multiple historical movement positions and movement accelerations, and then updates the multiple historical movement positions and movement accelerations in combination with the new position, so as to obtain the predicted movement trajectory of the target user by continuously iterating the above prediction process.
[0072] Furthermore, after completing the above iterative calculation, a predicted movement trajectory of the target user based on historical data, current status and future trend prediction can be obtained. This predicted movement trajectory not only includes every location point that the user may pass through, but also reflects the changes in the speed and direction of the user's movement. Therefore, based on the predicted movement trajectory, the target movement position of the target user at a certain point in the future can be further determined, and then based on the target movement position, it can be determined whether there is an operation risk in the high-altitude operation performed by the target user.
[0073] Optionally, in the risk warning method for high-altitude operations provided in an embodiment of the present application, determining the operation risk warning information of the target user based on the target moving position includes: judging whether the target moving position is in a preset area, wherein the preset area is a dangerous area; when the target moving position is in the preset area, determining the operation risk warning information as high-risk warning information; when the target moving position is not in the preset area, judging whether the predicted moving trajectory passes through the preset area; when the predicted moving trajectory passes through the preset area, determining the operation risk warning information as medium-risk warning information; when the predicted moving trajectory does not pass through the preset area, determining the operation risk warning information as low-risk warning information.
[0074] Specifically, after obtaining the target moving position of the target user, it can be determined whether the target moving position is in a preset area, wherein the preset area may be a preset dangerous area. When the user is in a dangerous area, it can be determined that the user's current operation risk is high.
[0075] Furthermore, when the target moving position is in the preset area, the operation risk warning information is determined as high-risk warning information, and the target user is notified in time to suspend the operation and adjust the moving trajectory, and feedback is given at the target user end in the form of a red light. When the target moving position is not in the preset area, it is necessary to determine whether the target user will pass through the preset area based on the predicted moving trajectory, and when the predicted moving trajectory passes through the preset area, the operation risk warning information is determined as medium-risk warning information, and the target user is notified in time to pay attention to the need to change the moving trajectory, and feedback is given at the target user end in the form of a yellow light. When the predicted moving trajectory does not pass through the preset area, it indicates that the target user will not pass through the preset area during the operation. At this time, the operation risk warning information can be determined as low-risk warning information, and feedback is given at the target user end in the form of a green light, so that the target user can continue to work, thereby achieving the technical effect of ensuring the safety of the operation and the timeliness of the risk prompt.
[0076] In order to improve the accuracy of risk identification and improve the operating safety of the target user, optionally, in the risk warning method for high-altitude operations provided in an embodiment of the present application, after obtaining a target image of the target user through a camera device, the method also includes: inputting the target image into an equipment recognition model, determining whether the target user carries the target equipment through the equipment recognition model, and obtaining a recognition result, wherein the equipment recognition model is obtained by training multiple third sample data, each third sample data includes at least one equipment, and labeling information corresponding to each equipment; when the recognition result indicates that the target user does not carry the target wearable device, the operation risk warning information is determined as high-risk warning information; when the recognition result indicates that the target user carries the target wearable device, the operation risk warning information is determined as low-risk warning information.
[0077] It should be noted that when judging the risks of high-altitude operations, it is also necessary to determine whether the target user is wearing safety wearable equipment correctly, such as safety belts / ropes, safety helmets, and insulating clothing, so as to determine whether the operation is risky based on the equipment status of the safety wearable equipment.
[0078] Specifically, the equipment recognition model can be used to determine whether the target user carries the target equipment. The equipment recognition model can be based on the YOLOv5 network architecture, perform feature extraction and classification on the safety protection equipment, input the target image into the YOLOv5 network, use the convolution layer and the pooling layer to extract the image features, and combine the attention mechanism and the feature pyramid network to perform multi-scale feature fusion, use the fully connected layer to perform detection and recognition, and obtain the recognition result of the safety protection equipment, so as to determine whether the target user carries the target equipment according to the recognition result.
[0079] Furthermore, when the recognition result indicates that the target user is not carrying the target wearable device, the operation risk warning information is determined as high-risk warning information. That is, when the target user is not carrying the target wearable device, the operation needs to be stopped immediately. Therefore, the operation risk warning information needs to be determined as high-risk warning information; when the recognition result indicates that the target user is carrying the target wearable device, it indicates that the wearable device is worn normally. At this time, the operation risk warning information is determined as low-risk warning information, and the operator can perform high-altitude operations normally, thereby achieving the technical effect of accurately identifying the risks of high-altitude operations.
[0080] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0081] The embodiment of the present application also provides a risk warning device for high-altitude operations. It should be noted that the risk warning device for high-altitude operations in the embodiment of the present application can be used to execute the risk warning method for high-altitude operations provided in the embodiment of the present application. The risk warning device for high-altitude operations provided in the embodiment of the present application is introduced below.
[0082] Figure 3 Schematic diagram of a risk warning device for high-altitude operations provided according to an embodiment of the present application. Figure 3 As shown, the device includes: a first identification unit 31, a second identification unit 32, and a first determination unit 33.
[0083] The first recognition unit 31 is used to obtain a target image of a target user through a camera device, and recognize a target position of the target user in the target image.
[0084] The second recognition unit 32 is used to input the target image into the posture recognition model to obtain the target posture information of the target user.
[0085] The first determination unit 33 is used to predict the target moving position of the target user based on the target position and target posture information, and determine the operation risk warning information of the target user based on the target moving position, and send the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk.
[0086] The risk warning device for high-altitude operation provided by the embodiment of the present application obtains the target image of the target user through the camera device through the first recognition unit 31, and recognizes the target position of the target user in the target image; the second recognition unit 32 inputs the target image into the posture recognition model to obtain the target posture information of the target user; the first determination unit 33 predicts the target moving position of the target user according to the target position and the target posture information, and determines the operation risk warning information of the target user according to the target moving position, and sends the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk. The problem of low safety and accuracy of supervising high-altitude workers through manual supervision in the related technology is solved. By identifying the target position of the worker in the image picture and the target posture of the worker, the moving position of the worker is predicted according to the target position and the target posture, so as to determine the position change of the worker in the future period of time, and judge whether there is a risk in the high-altitude operation according to the predicted moving position, thereby achieving the effect of improving the accuracy and timeliness of risk monitoring.
[0087] Optionally, in the risk warning device for high-altitude operations provided in an embodiment of the present application, the first identification unit 31 includes: a first acquisition module, used to acquire image pictures taken by multiple camera devices, obtain multiple initial image pictures, and screen image pictures containing target users from the multiple initial image pictures to obtain multiple candidate image pictures; a second acquisition module, used to acquire the resolution level of each candidate image picture and the size ratio of the target user to the candidate image picture, wherein the resolution level and the resolution size are in direct proportion; a first calculation module, used to perform weighted summation of the resolution level and size ratio of each candidate image picture according to preset weights to obtain a score for each candidate image picture; a first determination module, used to determine the candidate image picture with the highest score as the target image picture.
[0088] Optionally, in the risk warning device for high-altitude operations provided in an embodiment of the present application, the first recognition unit 31 includes: a recognition module, used to input the target image picture into a position recognition model to obtain the target position of the target user, wherein the target position is determined by the boundary point coordinates of the character frame of the target user, and the position recognition model is obtained by training a first sample data set, wherein the first sample data set includes multiple first sample data, each first sample data includes a historical image picture, a character frame in the historical image picture, and the boundary point coordinates of the character frame.
[0089] Optionally, in the risk warning device for high-altitude operations provided in an embodiment of the present application, the posture recognition model is trained in the following manner: an acquisition unit is used to acquire multiple second sample information, wherein each second sample information includes a human body image and posture information corresponding to the human body image; a training unit is used to train the initial recognition model according to multiple second sample information to obtain a posture recognition model.
[0090] Optionally, in the risk warning device for high-altitude operations provided in an embodiment of the present application, the first determination unit 33 includes: a second determination module, used to determine the movement acceleration of the target user based on the target posture information; a second calculation module, used to obtain multiple historical moving positions of the target user, and calculate the multiple historical moving positions, target positions and movement accelerations according to the Kalman filter algorithm to obtain the predicted movement trajectory of the target user; a third determination module, used to determine the target movement position of the target user based on the predicted movement trajectory.
[0091] Optionally, in the risk warning device for high-altitude operations provided in an embodiment of the present application, the first determination unit 33 includes: a first judgment module, used to determine whether the target moving position is in a preset area, wherein the preset area is a dangerous area; a fourth determination module, used to determine the operation risk warning information as high-risk warning information when the target moving position is in the preset area; a second judgment module, used to determine whether the predicted moving trajectory passes through the preset area when the target moving position is not in the preset area; a fifth determination module, used to determine the operation risk warning information as medium-risk warning information when the predicted moving trajectory passes through the preset area; and a sixth determination module, used to determine the operation risk warning information as low-risk warning information when the predicted moving trajectory does not pass through the preset area.
[0092] Optionally, in the risk warning device for high-altitude operations provided in an embodiment of the present application, after obtaining a target image of the target user through a camera device, the device also includes: a third recognition unit, used to input the target image into an equipment recognition model, and determine whether the target user carries the target equipment through the equipment recognition model to obtain a recognition result, wherein the equipment recognition model is obtained by training multiple third sample data, each third sample data includes at least one equipment, and labeling information corresponding to each equipment; a second determination unit, used to determine the operation risk warning information as high-risk warning information when the recognition result indicates that the target user does not carry the target wearable device; a third determination unit, used to determine the operation risk warning information as low-risk warning information when the recognition result indicates that the target user carries the target wearable device.
[0093] The risk warning device for high-altitude operations includes a processor and a memory. The first identification unit 31, the second identification unit 32, the first determination unit 33, etc. are all stored in the memory as program units, and the processor executes the program units stored in the memory to realize corresponding functions.
[0094] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the problem of low safety and accuracy in the related art of supervising high-altitude workers by manual supervision is solved by adjusting kernel parameters.
[0095] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.
[0096] An embodiment of the present invention provides a computer-readable storage medium on which a program is stored. When the program is executed by a processor, the risk warning method for high-altitude operations is implemented.
[0097] An embodiment of the present invention provides a processor, which is used to run a program, wherein the risk warning method for high-altitude operations is executed when the program is running.
[0098] Figure 4 is a schematic diagram of an electronic device provided according to an embodiment of the present application, such as Figure 4 As shown, an embodiment of the present invention provides an electronic device, the electronic device 40 includes a processor, a memory, and a program stored in the memory and executable on the processor, and the processor executes the program to implement the steps of the risk warning method for high-altitude operations. The device in this article can be a server, a PC, a PAD, a mobile phone, etc.
[0099] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the steps of the above-mentioned risk warning method for high-altitude operations.
[0100] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.
[0101] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0102] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0103] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0104] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0105] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0106] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0107] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0108] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.
Claims
1. A risk warning method for high-altitude operations, characterized in that: include: Acquire a target image of a target user through a camera device, and identify a target position of the target user in the target image; Inputting the target image into a posture recognition model to obtain target posture information of the target user; The target moving position of the target user is predicted based on the target position and the target posture information, and the operation risk warning information of the target user is determined based on the target moving position, and the operation risk warning information is sent to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk.
2. The method according to claim 1, characterized in that The target images of the target user obtained through the camera device include: Acquire images captured by multiple camera devices to obtain multiple initial image pictures, and select image pictures containing the target user from the multiple initial image pictures to obtain multiple candidate image pictures; Obtaining the resolution level of each candidate image and the size ratio of the target user to the candidate image, wherein the resolution level and the resolution size are in direct proportion; Performing a weighted summation of the resolution level and the size ratio of each candidate image picture according to a preset weight to obtain a score for each candidate image picture; The candidate image picture with the highest score is determined as the target image picture.
3. The method according to claim 1, characterized in that Identifying the target position of the target user in the target image includes: The target image picture is input into the position recognition model to obtain the target position of the target user, wherein the target position is determined by the boundary point coordinates of the character frame of the target user, and the position recognition model is trained by a first sample data set, wherein the first sample data set includes multiple first sample data, each first sample data includes a historical image picture, a character frame in the historical image picture, and the boundary point coordinates of the character frame.
4. The method according to claim 1, characterized in that: The posture recognition model is trained in the following way: Acquire multiple second sample information, wherein each second sample information includes a human body image and posture information corresponding to the human body image; The initial recognition model is trained according to the plurality of second sample information to obtain the posture recognition model.
5. The method according to claim 1, characterized in that Predicting the target moving position of the target user according to the target position and the target posture information includes: Determine the movement acceleration of the target user according to the target posture information; Acquire multiple historical moving positions of the target user, and calculate the multiple historical moving positions, the target position and the moving acceleration according to a Kalman filter algorithm to obtain a predicted moving trajectory of the target user; The target moving position of the target user is determined according to the predicted moving trajectory.
6. The method according to claim 5, characterized in that Determining the operation risk warning information of the target user according to the target moving position includes: Determining whether the target moving position is in a preset area, wherein the preset area is a dangerous area; When the target moving position is within the preset area, determining the operation risk warning information as high-risk warning information; In the case where the target moving position is not in the preset area, determining whether the predicted moving trajectory passes through the preset area; In the case where the predicted movement trajectory passes through the preset area, determining the operation risk warning information as medium risk warning information; When the predicted movement trajectory does not pass through the preset area, the operation risk warning information is determined as low-risk warning information.
7. The method according to claim 1, characterized in that After obtaining the target image of the target user through the camera device, the method further includes: Inputting the target image into an equipment recognition model, determining whether the target user carries the target equipment through the equipment recognition model, and obtaining a recognition result, wherein the equipment recognition model is obtained by training a plurality of third sample data, each of which includes at least one equipment and labeling information corresponding to each equipment; When the recognition result indicates that the target user does not carry the target wearable device, determining the operation risk warning information as high-risk warning information; When the recognition result indicates that the target user carries the target wearable device, the operation risk warning information is determined as low-risk warning information.
8. A risk warning device for high-altitude operations, characterized in that: include: A first recognition unit, configured to obtain a target image of a target user through a camera device, and recognize a target position of the target user in the target image; A second recognition unit is used to input the target image into a posture recognition model to obtain target posture information of the target user; A first determination unit is used to predict the target moving position of the target user based on the target position and the target posture information, and determine the operation risk warning information of the target user based on the target moving position, and send the operation risk warning information to the target user, wherein the operation risk warning information is used to inform the target user whether there is an operation risk.
9. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the risk warning method for high-altitude operations described in any one of claims 1 to 7 is implemented.
10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the risk warning method for high-altitude operations as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Electric power climbing operation safety early warning system and early warning method
CN114038158A
Underground pedestrian border crossing risk detection method and device, electronic equipment and storage medium
CN114140832A
Power grid safety early warning method and device, computer equipment and storage medium
CN115620208A
Intelligent monitoring method, device and equipment for safe operation of data center and medium
CN116129350A
Real-time crowd measurement and management systems and methods thereof
US20220254161A1
Cited By
Early warning method and early warning system for lower part of portal crane
CN120544121A
