A safety warning method and device
By using image acquisition and deep learning algorithms to identify work content and violations, generating warning information and uploading it to a remote control center, the system solves the problems of insufficient real-time performance and accuracy in traditional safety supervision methods, and achieves real-time and accurate safety monitoring of high-risk work scenarios.
Patent Information
- Application Number
- CN202610641624.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-05-11
- Publication Date
- 2026-07-10
AI Technical Summary
Traditional safety supervision methods lack real-time and accuracy in high-risk operation scenarios, making it difficult to achieve dynamic coverage around the clock and across all areas, and to identify the behavioral norms of workers in real time, resulting in the inability to detect potential safety hazards in a timely manner.
Image acquisition and deep learning algorithms are used to identify job content, assess safety conditions, and determine violations. Warning information is generated and violations are prompted via voice and text. Serious violations are uploaded to a remote control center for processing.
It enables real-time and precise monitoring and guidance of the work site, improves the accuracy and timeliness of safety accident judgment, and reduces the risk of accidents.
Smart Images
Figure CN122369185A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of safety production technology, specifically to a safety warning method and device. Background Technology
[0002] In high-risk work scenarios, such as construction, power maintenance, and chemical production, ensuring the safe operation of workers is a core prerequisite for the smooth operation of production activities. Traditional safety supervision methods mainly rely on manual inspections and post-event accountability mechanisms, but their limitations are becoming increasingly apparent. On the one hand, manual inspections are limited by insufficient human resources, fixed inspection frequencies, and uneven professional competence among personnel, making it difficult to achieve dynamic coverage around the clock and across all areas. For example, in large construction sites, safety officers need to monitor multiple high-risk aspects such as high-altitude operations, machinery operation, and temporary power supply simultaneously, making it easy to miss potential hazards due to blind spots or distractions. On the other hand, traditional supervision methods lack intelligent technology support, making it impossible to identify and intervene in the compliance of workers' behavior (such as whether they wear protective equipment and whether their operating procedures meet standards) in real time. Especially during live-line work in power maintenance or inspection of chemical reaction units, complex on-site environments and hidden risks are often difficult to detect in a timely manner by human observation. Once the best time to deal with potential accidents is missed, they may escalate into serious consequences such as fires, explosions, or falls. This highlights the structural deficiencies of the traditional regulatory system in terms of real-time performance, accuracy, and systematicity, and urgently requires technological innovation to achieve a transformation and upgrade from "human-based" to "technology-based" prevention. Summary of the Invention
[0003] The purpose of this application is to provide a safety warning method and device for real-time and accurate monitoring and guidance of the construction process, effectively improving the accuracy and timeliness of determining work safety violations and reducing the risk of safety accidents.
[0004] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:
[0005] A safety warning method includes:
[0006] Acquire images and preprocess them;
[0007] The pre-processed images are used to identify the work content, assess safety conditions, and determine violations. When a violation is determined, a warning message is generated.
[0008] Warning messages will be sent to those who violate the rules, informing them of their misconduct and the correct operating procedures.
[0009] Furthermore, in the image acquisition and preprocessing, the preprocessing includes grayscale conversion, noise reduction, and filtering.
[0010] Furthermore, the process of identifying the work content, assessing safety conditions, and determining violations based on the preprocessed image includes, where,
[0011] The task content recognition includes establishing an image feature library for various types of tasks, and using deep learning algorithms to extract features and classify the preprocessed images.
[0012] The safety condition assessment includes using target detection algorithms to identify whether workers are wearing complete safety protective equipment, using image analysis to determine whether warning signs in the work area are clearly visible and whether the equipment is operating normally, and quantifying each detection item according to preset safety condition standards.
[0013] The violation determination includes constructing a violation identification model, learning from a large amount of historical violation image data to identify various common violation patterns, comparing real-time acquired images with the violation model, and immediately determining the type of violation and recording detailed violation information once a matching feature is found.
[0014] Furthermore, the preprocessed image is subjected to feature extraction and classification recognition using deep learning algorithms to determine the specific type of work being performed by the operator and to match it with preset work safety regulations to determine whether it meets the corresponding operational requirements.
[0015] Furthermore, each testing item is quantitatively evaluated, and a certain number of points are deducted for each missing item. When the total score is lower than a set threshold, it is determined that the safety conditions are not met and there is a safety hazard.
[0016] Furthermore, the system immediately generates warning information to issue an instant alert to the violator. The warning information includes voice prompts and text messages, and is uploaded to a remote control center via the network.
[0017] To achieve the above objectives, this application also provides a safety warning device, the device comprising:
[0018] The image acquisition module is used to acquire images and preprocess them.
[0019] The processing module is used to identify the work content, assess the safety conditions, and determine the violations based on the pre-processed images. When a violation is determined, a warning message will be generated.
[0020] The warning module is used to send warning messages to violators, informing them of their violations and the correct operating procedures.
[0021] The image acquisition module, processing module, and warning device are all integrated into the peripheral module.
[0022] Furthermore, the device also includes:
[0023] The wireless communication module is used for data transmission with the remote control center;
[0024] A power module is used to provide power to the various modules in the device;
[0025] The inductive power detection alarm module is used to issue an alarm in a timely manner when the power level is lower than a set threshold.
[0026] The display module is used to display various safety information, work instructions, and warning text.
[0027] The wireless communication module, power supply module, inductive power detection and alarm module, and display module are also integrated into the peripheral module.
[0028] Furthermore, the peripheral module is a name tag.
[0029] Compared with the prior art, the beneficial effects of the present invention are:
[0030] This application collects and intelligently analyzes image and video data from the work site to quickly and accurately determine whether workers' operations are in violation of regulations. Based on the severity of the violation, it selects appropriate warning methods. For minor violations, an on-site voice warning is issued, informing workers of the correct operating procedures. For serious violations, while issuing an on-site warning, the captured video of the violation is uploaded to a remote control center via the network. Management personnel at the control center can receive and view the violation in real time for further analysis and processing, such as educating and penalizing violators, and adjusting and optimizing safety management measures at the work site to prevent similar violations from recurring. This allows for real-time and precise monitoring and guidance of the construction process, ensuring operational safety.
[0031] This application provides a safety warning device for implementing the above-mentioned safety warning method. The device is designed in the form of a name tag for easy wearing and use by construction workers. Attached Figure Description
[0032] Figure 1 This is a flowchart illustrating a safety warning method according to this application.
[0033] Figure 2 This is a structural block diagram of a safety warning device according to this application.
[0034] Figure 3 This is a schematic diagram of one embodiment of a safety warning device according to this application. Detailed Implementation
[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
[0036] Please see Figure 1 This application provides a safety warning method, the method comprising:
[0037] In step S101, an image is acquired and preprocessed.
[0038] In step S102, the work content is identified, safety conditions are assessed, and violations are determined based on the preprocessed image. When a violation is determined, a warning message is generated.
[0039] In step S103, a warning message is sent to the person who violated the rules, informing them of the violation and the correct operating procedure.
[0040] In practice, the image acquisition and preprocessing described in step S101 includes grayscale conversion, noise reduction, and filtering, thereby improving the quality and clarity of the image and laying the foundation for subsequent feature extraction and recognition.
[0041] In practice, more specifically, step S102 involves identifying the work content, assessing safety conditions, and determining violations based on the preprocessed image.
[0042] The task content recognition includes establishing an image feature library for various tasks, and using deep learning algorithms to extract features and classify the preprocessed images. For example, for welding tasks, the appearance features of the welding machine, the shape and color of the welding sparks, etc. are identified. For high-altitude operations, feature elements such as scaffolding and safety belts are detected to determine the specific type of task being performed by the operator and match it with the preset work safety specifications to determine whether it meets the corresponding operational requirements.
[0043] The safety condition assessment includes using target detection algorithms to identify whether workers are wearing complete safety protective equipment, using image analysis to determine whether warning signs in the work area are clearly visible and whether the equipment is operating normally, and quantifying each detection item according to preset safety condition standards. In this embodiment, the target detection algorithm can use algorithms such as R-CNN series and Transformer-Based series.
[0044] In some implementations, each test item is quantitatively evaluated according to preset safety condition standards. For example, full marks are awarded for wearing all safety protective equipment, and a certain number of points are deducted for each missing item. When the total score is lower than a set threshold, it is determined that the safety conditions are not met and there is a safety hazard.
[0045] The violation determination process includes constructing a violation recognition model. By learning from a large amount of historical violation image data, it identifies various common violation patterns, such as personnel crossing safety fences or performing repairs on equipment that is still powered on. Real-time acquired images are compared with the violation model. Once a matching feature is found, the type of violation is immediately determined, and detailed information about the violation is recorded, including time, location, violator, and specific violation event.
[0046] In practice, at step S103, upon determining a violation, the system immediately generates a warning message, including voice prompts and text information. This is delivered instantly to the violator via on-site safety warning devices, informing them of the violation and the correct operating procedure. For serious violations, while issuing an on-site warning, the system simultaneously uploads the captured video footage to a remote control center. Management personnel at the control center can receive and view the violation in real time for further analysis and processing, such as educating and penalizing the violator, and adjusting and optimizing on-site safety management measures to prevent similar violations from recurring and ensure operational safety.
[0047] It should be noted that in some embodiments, whether a violation is serious can be reflected by a quantitative evaluation score, such as a quantitative score of less than 60 being judged as a serious violation.
[0048] On the other hand, this application also provides a safety warning device, please refer to [link / reference]. Figure 2 The device includes:
[0049] Image acquisition module 3 is used to acquire images and preprocess them;
[0050] In this embodiment, the image acquisition module 3 uses a high-definition camera.
[0051] Processing module 6 is used to identify the work content, assess the safety conditions, and determine the violations based on the preprocessed image. When a violation is determined, a warning message will be generated.
[0052] In this embodiment, the processing module 6 can be an AI chip. The AI chip is equipped with at least image feature libraries for various tasks, target detection algorithms, violation recognition models, deep learning algorithms, and warning broadcast text and voice libraries, thereby realizing the process of content recognition, security condition assessment, and violation judgment.
[0053] Warning module 7 is used to send warning messages to violators, informing them of their violations and the correct operating procedures;
[0054] In this embodiment, the warning module 7 can be a high-volume player to ensure that workers can hear the reminders in a timely manner even in noisy working environments.
[0055] The image acquisition module 3, processing module 6, and warning device 7 are all integrated on the peripheral module 1.
[0056] In addition, the device also includes: a wireless communication module 2 for transmitting data with a remote control center;
[0057] Specifically, the wireless communication module 2 has GSM 5G communication capabilities and Wi-Fi connectivity, enabling it to transmit data with the remote control center in real time, thus achieving remote monitoring and dispatching.
[0058] Power module 4 is used to provide power to the various modules in the device;
[0059] The inductive power detection alarm module 5 is used to issue an alarm in a timely manner when the power is lower than a set threshold. The inductive power detection alarm module 5 is implemented by a circuit connection of a voltage sensor and a buzzer.
[0060] Display module 8 is used to display various safety information, work instructions and warning text. In this embodiment, display module 8 is an LCD screen.
[0061] The wireless communication module 2, power supply module 4, inductive power detection and alarm module 5, and display module 8 are also integrated on the peripheral module 1.
[0062] Please see Figure 3 One implementation of the device is that the peripheral module 1 is a name tag with a pocket clip or buckle on the back for easy wearing. The display module 8 is located on the front of the name tag, and all other modules are built into the name tag. It should be noted that the peripheral module 1 is not limited to the name tag and can also be other devices, such as a safety helmet.
[0063] In this specification, terms such as "one embodiment," "another embodiment," "embodiment," and "preferred embodiment" refer to specific features, structures, or characteristics described in connection with that embodiment, which are included in at least one embodiment described in the general description of this application. The appearance of the same term in multiple places in the specification does not necessarily refer to the same embodiment. Furthermore, when a specific feature, structure, or characteristic is described in connection with any embodiment, the intention is to suggest that implementing such a feature, structure, or characteristic in conjunction with other embodiments also falls within the scope of this invention.
[0064] Although the invention has been described herein with reference to several illustrative embodiments, it should be understood that many other modifications and implementations can be devised by those skilled in the art, which will fall within the scope and spirit of the principles disclosed herein. More specifically, various modifications and improvements can be made to the components or layout of the subject matter arrangement within the scope of the disclosure, drawings, and claims. Besides modifications and improvements to the components or layout, other uses will be apparent to those skilled in the art.
Claims
1. A safety warning method, characterized in that, include: Acquire images and preprocess them; The pre-processed images are used to identify the work content, assess safety conditions, and determine violations. When a violation is determined, a warning message is generated. Warning messages will be sent to those who violate the rules, informing them of their misconduct and the correct operating procedures.
2. The safety warning method according to claim 1, characterized in that: The image acquisition and preprocessing process includes grayscale conversion, noise reduction, and filtering.
3. The safety warning method according to claim 1, characterized in that: The process involves identifying the work content, assessing safety conditions, and determining violations based on the preprocessed images. The task content recognition includes establishing an image feature library for various types of tasks, and using deep learning algorithms to extract features and classify the preprocessed images. The safety condition assessment includes using target detection algorithms to identify whether workers are wearing complete safety protective equipment, using image analysis to determine whether warning signs in the work area are clearly visible and whether the equipment is operating normally, and quantifying each detection item according to preset safety condition standards. The violation determination includes constructing a violation identification model, learning from a large amount of historical violation image data to identify various common violation patterns, comparing real-time acquired images with the violation model, and immediately determining the type of violation and recording detailed violation information once a matching feature is found.
4. A safety warning method according to claim 3, characterized in that: The process involves using deep learning algorithms to extract features and classify the preprocessed images, thereby determining the specific type of work being performed by the operator and matching it with preset work safety regulations to determine whether it meets the corresponding operational requirements.
5. A safety warning method according to claim 3, characterized in that: Each test item is quantitatively evaluated, and a certain number of points are deducted for each missing item. When the total score is lower than the set threshold, it is determined that the safety conditions are not met and there is a safety hazard.
6. A safety warning method according to claim 3, characterized in that: The system immediately generates a warning message to issue an instant alert to the violator. The warning message includes voice prompts and text information, and is uploaded to a remote control center via the network.
7. A safety warning device, characterized in that, The device includes: The image acquisition module is used to acquire images and preprocess them. The processing module is used to identify the work content, assess the safety conditions, and determine the violations based on the pre-processed images. When a violation is determined, a warning message will be generated. The warning module is used to send warning messages to violators, informing them of their violations and the correct operating procedures. The image acquisition module, processing module, and warning device are all integrated into the peripheral module.
8. A safety warning device according to claim 7, characterized in that, The device further includes: The wireless communication module is used for data transmission with the remote control center; A power module is used to provide power to the various modules in the device; The inductive power detection alarm module is used to issue an alarm in a timely manner when the power level is lower than a set threshold. The display module is used to display various safety information, work instructions, and warning text. The wireless communication module, power supply module, inductive power detection and alarm module, and display module are also integrated into the peripheral module.
9. A safety warning device according to claim 7, characterized in that, The peripheral module is a name tag.