Control method and equipment of construction safety management system and storage medium

Abstract

The invention discloses a control method and device of a construction safety management system and a storage medium, and belongs to the technical field of image data processing. The method comprises the following steps: acquiring visual feature information extracted by a visual perception model based on a construction scene, determining a risk detection result of the visual feature information, if the risk detection result is that a risk exists, constructing a cue word according to the risk detection result and the visual feature information, and obtaining a cue word through a large language model; and generating a safety detection result corresponding to the construction scene based on the prompt word. According to the method, the risk information of the construction scene is identified in a grading manner, and deep reasoning is performed through the large language model, so that the risk identification accuracy of construction safety detection is improved.

Description

Technical Field

[0001] This application relates to the field of image data processing technology, and in particular to a control method, equipment and storage medium for a construction safety management system. Background Technology

[0002] To further improve the accuracy of safety and compliance testing at construction sites, related technologies utilize visual sensors such as cameras and drones, along with computer vision models deployed on-site, to continuously collect monitoring footage of the construction site. These visual sensors are then analyzed in real-time using pre-trained visual perception models to identify target objects such as people, machines, materials, methods, and the environment, as well as their status. Based on preset risk identification rules, the system outputs detection results indicating the presence and type of risk.

[0003] However, the risk assessment in this construction safety inspection process relies on the visual model matching preset rules. This risk assessment is difficult to integrate with specific construction scenarios for comprehensive analysis and reasoning. For example, when a person is detected not wearing a safety helmet, it's difficult to determine whether that person is in a high-altitude work area and whether there is a risk of falling objects nearby. In this situation, the construction safety inspection's assessment of risks and hazards is not comprehensive enough, limiting the accuracy of safety management decisions.

[0004] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0005] The main purpose of this application is to provide a control method, equipment and storage medium for a construction safety management system, which aims to solve the technical problem that the assessment of risks and hidden dangers in construction safety detection is not comprehensive enough.

[0006] To achieve the above objectives, this application provides a control method for a construction safety management system, the method comprising the following steps: Obtain visual feature information extracted from the construction scene by the visual perception model, and determine the risk detection result of the visual feature information; If the risk detection result indicates that a risk exists, a prompt word is constructed based on the risk detection result and the visual feature information; Using a large language model, safety inspection results corresponding to the construction scenario are generated based on the prompt words.

[0007] In one embodiment, the step of acquiring visual feature information extracted from the construction scene by the visual perception model and determining the risk detection result of the visual feature information includes: Receive the visual feature information output by the visual perception model, and determine the construction scene information based on the visual feature information; Based on the construction scenario information, obtain the risk identification rules corresponding to the construction scenario; The risk identification rules are matched with the visual feature information, and the risk detection result is determined based on the matching result.

[0008] In one embodiment, the step of matching the risk identification rule with the visual feature information and determining the risk detection result based on the matching result includes: Based on the visual feature information or the visual information of the construction scene, determine the target location data of the construction scene; The target location data, the visual feature information, and the risk identification rules are correlated and analyzed to generate risk information and the confidence level corresponding to the risk information. When the confidence level is lower than the confidence level threshold, the risk information is marked as having a risk, and the risk detection result is generated based on the marking result.

[0009] In one embodiment, before the step of acquiring visual feature information extracted from the construction scene by the visual perception model and determining the risk detection result of the visual feature information, the method further includes: Receive visual information of the construction scene collected by a visual sensor, wherein the visual sensor includes a wearable device, a camera and / or a drone; The visual perception model is used to identify target objects in the visual information of the construction scene, as well as the target feature information of the target objects. The visual feature information is constructed based on the target object and the target feature information of the target object.

[0010] In one embodiment, the step of constructing a prompt word based on the risk detection result and the visual feature information if the risk detection result indicates the presence of risk includes: If, based on the risk detection results, it is determined that there is a risk in the construction scenario, a preset prompt word template is obtained; The risk detection results and the visual feature information are filled into the prompt word template to form the prompt word.

[0011] In one embodiment, the step of filling the prompt word template with the risk detection result and the visual feature information to form the prompt word includes: Identify at least one information placeholder in the prompt word template that is related to the risk object, risk behavior, and / or risk status; Based on the risk detection results, the target object, the attribute information of the target object, and the environmental information are extracted from the visual feature information. The target object, attribute information, and environmental information are matched with the type of the information placeholder and filled into the prompt word template to form the prompt word.

[0012] In one embodiment, the step of filling the prompt word template with the risk detection result and the visual feature information to form the prompt word includes: Obtain the risk identification rules corresponding to the construction scenario; The risk detection results, the risk identification rules, and the visual feature information are matched with the fill words in the prompt word template. Based on the matching results, the filler words in the prompt word template are replaced according to the risk detection results, the risk identification rules, and the visual feature information to form the prompt word.

[0013] In one embodiment, the step of generating the safety inspection result corresponding to the construction scenario based on the prompt words using a large language model further includes: Based on the suggested keywords, construct a search query; The search query is matched with the construction safety knowledge base to retrieve at least one safety regulation clause or historical handling case related to the risk detection result, thereby forming enhanced contextual information; The enhanced context information is combined with the prompt word to form an enhanced prompt word; Using the large language model and based on the enhanced prompt words, safety inspection results corresponding to the construction scenario are generated.

[0014] In addition, to achieve the above objectives, this application also provides a control device for a construction safety management system, the device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the control method for the construction safety management system as described above.

[0015] In addition, to achieve the above objectives, this application also provides a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of the control method of the construction safety management system as described above.

[0016] One or more technical solutions proposed in this application have at least the following technical effects: This application uses a visual perception model to extract real-time visual feature information of the construction scene and determine preliminary risk detection results. When a risk is detected, the risk detection results are combined with the visual feature information to construct a prompt word. Based on this prompt word, a large language model is used for deep understanding and reasoning to generate a safety detection result. This improves the accuracy of the system's risk identification and allows users to obtain information based on the safety detection result that is completely corresponding to the specific construction scene and includes risk cause analysis, potential consequence assessment, and specific response suggestions in natural language. This enhances the auxiliary decision-making capability and the accuracy and efficiency of management response in construction safety detection. Attached Figure Description

[0017] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a flowchart illustrating the first embodiment of the control method for the construction safety management system of this application; Figure 2 This is a flowchart illustrating the second embodiment of the control method for the construction safety management system of this application; Figure 3 This is a flowchart illustrating the third embodiment of the control method for the construction safety management system of this application; Figure 4 This is a schematic diagram of the structure of the control equipment of the construction safety management system of the hardware operating environment involved in the embodiments of this application.

[0020] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0021] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0022] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0023] The main solution of this application embodiment is: to obtain visual feature information extracted by the visual perception model based on the construction scene, and to determine the risk detection result of the visual feature information. If the risk detection result indicates that there is a risk, a prompt word is constructed based on the risk detection result and the visual feature information. Then, a safety detection result corresponding to the construction scene is generated based on the prompt word through a large language model.

[0024] Existing technologies utilize visual sensors such as cameras and drones deployed on-site, along with computer vision models, to continuously collect monitoring footage from the construction site. These sensors then analyze the footage in real-time using pre-trained visual perception models to identify target objects such as people, machinery, materials, methods, and the environment, along with their states. Based on preset risk identification rules, the system outputs detection results indicating the presence and type of risk. However, the risk assessment in this construction safety detection process relies heavily on the visual model's matching of preset rules. This makes it difficult to comprehensively analyze and reason about specific construction scenarios. For example, when a worker is detected not wearing a safety helmet, it's challenging to determine whether the worker is in a high-altitude work area and whether there is a risk of falling objects nearby. In such cases, the construction safety detection system's risk assessment is insufficiently comprehensive, limiting the accuracy of safety management decisions.

[0025] This application uses a visual perception model to extract real-time visual feature information of the construction scene and determine preliminary risk detection results. When a risk is detected, the risk detection results are combined with the visual feature information to construct a prompt word. Based on this prompt word, a large language model is used for deep understanding and reasoning to generate a safety detection result. This improves the accuracy of the system's risk identification and allows users to obtain information based on the safety detection result that is completely corresponding to the specific construction scene and includes risk cause analysis, potential consequence assessment, and specific response suggestions in natural language. This enhances the auxiliary decision-making capability and the accuracy and efficiency of management response in construction safety detection.

[0026] To better understand the above technical solutions, exemplary embodiments of this application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this application and to fully convey the scope of this application to those skilled in the art.

[0027] It should be noted that the executing entity in this embodiment can be a construction safety management system, or a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or control device of the construction safety management system capable of the above functions. This embodiment does not specifically limit it. The following uses a construction safety management system as an example to describe this embodiment and the following embodiments.

[0028] Based on this, the embodiments of this application provide a control method for a construction safety management system, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the control method for the construction safety management system of this application.

[0029] In this embodiment, the control method of the construction safety management system includes steps S10 to S40: Step S10: Obtain visual feature information extracted from the construction scene by the visual perception model, and determine the risk detection results of the visual feature information; In this embodiment, the construction safety management system can be deployed on a cloud server or an edge server at the construction site. It comprises various components such as a visual perception model, a risk identification module, and a large language model. It can receive construction scene environmental information transmitted by sensor devices used for collecting construction scene information, such as cameras, smart safety helmets, smart work clothes, and / or drones. This construction scene environmental information includes visual information such as construction images and videos for analysis and recognition by the visual perception model. It may also include data such as sound and radar data to support the construction safety management system's multimodal recognition of risks at the construction site.

[0030] Specifically, the construction safety management system includes a pre-set risk rule base containing corresponding risk identification rules. These risk identification rules can be obtained through neural network training, pre-set or input by the administrator, or determined based on domain rule information corresponding to the construction scenario. The construction safety management system can use rule engines or deep learning networks to match these risk identification rules with visual feature information to determine the risk detection result.

[0031] Optionally, the construction safety management system can acquire corresponding risk identification rules based on different construction scenarios. The system receives visual feature information output by the visual perception model, determines the construction scenario information based on this information, acquires the corresponding risk identification rules, matches these rules with the visual feature information, and determines the risk detection result based on the matching result. Specifically, the risk identification rules can be obtained by identifying the construction scenario type corresponding to the construction scenario information.

[0032] For example, a construction safety management system can parse high-level semantic tags in visual feature information, such as foundation pit excavation, steel structure hoisting, and interior decoration, to determine the current macro-construction scenario. Based on the identified scenario type, it dynamically filters and loads a subset of strongly related risk identification rules from the global rule base, ensuring the targeting and efficiency of the matching. For example, for high-altitude operations, rules related to fall prevention and falling object prevention are loaded; for hot work scenarios, rules related to fire-fighting equipment and flammable material isolation are loaded. During the matching phase, visual feature information, such as whether personnel are wearing safety belts and whether a warning zone is set up below, is logically verified against each loaded sub-rule. Each rule is expressed in the form of "IF [condition] THEN [risk conclusion]". The system evaluates whether the visual features meet the logical expression of the condition and, based on the number and weight of the met conditions, comprehensively determines the risk detection result, which can be refined to the specific safety clauses violated.

[0033] Furthermore, the construction safety management system determines the target location data of the construction scene based on visual feature information or visual information of the construction scene. The target location data refers to the actual construction location of the scene. For example, based on the data source of the visual information of the construction scene received by the visual perception model, the location information corresponding to the sensing device of the data source is determined as the target location data. Alternatively, based on the markers in the visual feature information, it is matched with pre-stored engineering information in the system, such as construction plans and Building Information Modeling (BIM), to determine the actual location of the corresponding construction scene as the target location data. The construction safety management system performs correlation analysis on the target location data, visual feature information, and risk identification rules to generate risk information and the corresponding confidence level. When the confidence level is lower than the confidence level threshold, the risk information is marked as having a risk, and the risk detection result is generated based on the marking result.

[0034] For example, the construction safety management system deeply integrates spatial location information and confidence assessment in rule matching to achieve more accurate risk assessment. By using the target pixel coordinates from visual feature information, combined with camera calibration parameters or a pre-built on-site digital twin model, the system calculates the target's position data in the real-world coordinate system. The risk identification rules include spatial constraints, such as safety distances, restricted areas, and equipment working radii. The system performs real-time calculations and correlation analysis between the target's position data and spatial constraints to determine whether personnel have entered hazardous areas or whether equipment spacing is insufficient. For each triggered potential risk, the construction safety detection system calculates a comprehensive confidence score. This confidence score integrates the visual recognition model's detection confidence of the relevant target, the estimation accuracy of spatial positioning, and / or the completeness of rule condition matching.

[0035] Optionally, if the calculated overall confidence level is lower than the threshold, it is determined that there is uncertainty in this risk assessment, which may be due to obstruction, lighting, or viewing angle. The construction safety management system can remove this risk assessment or specially mark such low-confidence risks, but still include them in the risk detection results, aiming to prompt subsequent analysis to focus on and identify them.

[0036] Step S20: If the risk detection result indicates that there is a risk, construct prompt words based on the risk detection result and visual feature information; Step S30: Using a large language model, generate safety inspection results corresponding to the construction scenario based on prompt words.

[0037] In this embodiment, after acquiring risk detection results, the construction safety detection system can read information such as risk type, risk object, risk attribute, and / or risk level contained in the risk detection results. Based on this, it can determine whether there is a safety risk in the construction scenario. After determining that a risk exists, the construction safety detection system will construct prompt words for a large language model based on the risk information and visual feature information in the risk detection results, and input the prompt words into the large language model. The large language model can perform analysis and deep reasoning based on the prompt words to generate safety detection results for the construction scenario. The safety detection results input into the large language model are usually text content, which may contain descriptive information such as risk type, risk level, risk cause, and / or risk solution or risk avoidance plan, so as to facilitate understanding by management personnel.

[0038] Optionally, the construction safety monitoring system also stores prompt word templates. If the risk detection results indicate that a construction scene poses a risk, the construction safety monitoring system can obtain the preset prompt word templates and fill in the prompt word templates with the risk detection results and visual feature information to form prompt words.

[0039] As an optional implementation method for constructing prompt words, the construction safety monitoring system identifies at least one information placeholder in the prompt word template related to the risk object, risk behavior, and / or risk status. Based on the risk detection results, it extracts the target object, the target object's attribute information, and its surrounding environment information from visual feature information. The target object, attribute information, and environmental information are then matched with the types of the aforementioned information placeholders. Based on the matching results, the target object, attribute information, and environmental information are filled into the prompt word template by replacing the placeholders, thus forming the prompt word.

[0040] As another optional implementation method for constructing prompt words, the construction safety detection system can also obtain the risk identification rules corresponding to the construction scene, match the risk detection results, risk identification rules and visual feature information with the fill words in the prompt word template, and replace the fill words in the prompt word template according to the matching results, risk detection results, risk identification rules and visual feature information to form prompt words.

[0041] Optionally, the construction safety monitoring system can obtain risk information such as risk objects, risk behaviors, and / or risk states from the risk monitoring results, match and replace the risk information with information placeholders to form preliminary prompt words, and combine the preliminary prompt words with risk identification rules and visual feature information to form prompt words for input into a large language model.

[0042] In one example, the construction safety monitoring system employs a direct-fill method based on a fixed frame. The system acquires a well-structured pre-defined template, which typically defines the model roles, task background, input data, and output format. For example, the template might be: Safety Analyst, please assess the safety status of the following construction scenario. Scenario Overview: [Scenario Description]. Observed Risk Signs: [Risk Description]. Please generate a report containing the risk level, regulatory basis, and remedial recommendations. The system summarizes the scenario description from visual feature information, such as "Two workers are installing formwork on the scaffolding platform at the southeast corner of the fifth floor of Building 2," and obtains the risk description from the risk detection results, such as "One of the workers did not fasten his safety belt to a secure anchor point." The system uses string search and replace to quickly generate usable prompts by replacing "[Scenario Description]" and "[Risk Description]" with the summarized and acquired text, respectively.

[0043] In another example, the construction safety monitoring system performs deep analysis on the selected prompt word template, identifying all preset information placeholders and their corresponding semantic role types, including but not limited to risk subjects, objects, behaviors, states, environments, and violated rule clauses. It extracts core risk assertions from the risk detection results and associates them with the original visual feature information that triggers those assertions, extracting detailed attributes that match each semantic role. For example, for the risk subject role, it extracts "worker" and supplements it with the possible "electrician" occupation attribute; for the environment role, it extracts information such as "damp ground" and "enclosed space." The system categorizes the extracted detailed attributes according to semantic roles and fills them into the corresponding placeholders with clear semantic annotations in the template. Simultaneously, the system fills the dedicated "reference rules" section of the template with the full text of the specific risk identification rules upon which this risk assessment is based, as key normative context, enabling the large language model to obtain direct regulatory guidance.

[0044] This application embodiment extracts real-time visual feature information of the construction scene through a visual perception model and determines preliminary risk detection results. When a risk is detected, the risk detection results are combined with the visual feature information to construct prompt words. Based on these prompt words, a large language model is used for deep understanding and reasoning to generate safety detection results. This improves the accuracy of the system's risk identification and allows users to obtain information based on the safety detection results that is completely corresponding to the specific construction scene and includes risk cause analysis, potential consequence assessment, and specific response suggestions in natural language. This enhances the auxiliary decision-making capability and the accuracy and efficiency of management response in construction safety detection.

[0045] Based on the same inventive concept, this application also provides a second embodiment, referring to... Figure 2 , Figure 2 This is a flowchart illustrating the second embodiment of the control method for the construction safety management system of this application.

[0046] In this embodiment, before obtaining the visual feature information extracted from the construction scene by the visual perception model as described in step S10, and determining the risk detection result of the visual feature information, steps S11 to S13 are also included: Step S11: Receive visual information of the construction scene collected by a visual sensor, wherein the visual sensor includes wearable devices, cameras and / or drones; Step S12: Identify the target objects in the visual information of the construction scene and the target feature information of the target objects through the visual perception model; Step S13: Construct visual feature information based on the target object and its target feature information.

[0047] In this embodiment, to ensure the comprehensiveness and accuracy of safety monitoring at the construction site, the construction safety management system can receive visual information about the construction scene, such as construction images or videos, from various visual sensing devices with photographic capabilities, such as wearable devices, cameras, and / or drones. These sensing devices include smart safety helmets or smart work clothes with cameras, which can collect specific construction scene information from the first-person perspective of construction workers. The construction safety management system further uses image recognition based on the construction scene information to identify target objects and their target feature information within the visual information, and constructs visual feature information based on these target objects and their feature information.

[0048] Optionally, when acquiring visual information of construction scenes from different dimensions collected by different sensing devices, information is extracted from each different construction scene visual information, and predetermined markers are identified. The visual perception model can then align the different construction scene visual information based on these markers.

[0049] Optionally, based on continuous image or video streams, the construction safety management system further determines the motion trend of target objects during video analysis. Motion trend refers to predicting future movement trajectories and state changes by analyzing the target's position, speed, and direction, or the object's shape, within a continuous frame sequence. By analyzing this motion trend, the construction safety management system correlates it in real-time with static environmental rules and dynamic equipment operating states, thereby identifying dynamic risk information that cannot be determined from a single frame image, such as personnel walking towards unprotected edges, the swaying trajectory of suspended objects potentially affecting the area below, or vehicles speeding into restricted areas.

[0050] In one example, the construction safety management system normalizes static images of the construction site captured by high-definition cameras and inputs them into a ResNet50 model. The ResNet50 model performs hierarchical feature extraction on the images through multiple sets of convolutional layers containing residual connections. Shallow networks identify edges and textures, while deep networks abstract high-dimensional features containing global semantics. The fully connected layers and Softmax activation function at the end of the model map these features to a preset scene category probability distribution, selecting the category with the highest probability as the output. For example, based on the comprehensive features of deep pits, support structures, and excavation equipment in the image, the model determines that the scene is "deep foundation pit excavation stage." This classification result directly triggers the system to load a set of targeted risk identification rules such as "foundation pit collapse" and "edge protection."

[0051] In another example, the construction safety management system directly inputs each frame of the video stream into the YOLOv8 model to extract common features from the images. The feature pyramid network fuses feature maps of different scales to simultaneously detect both large and small targets. The model's core detection head predicts multiple bounding boxes in parallel on each cell of the image grid. Each box contains precise center coordinates, width, height, object category confidence, and the probability of each category. For example, for a frame of tower crane operation image, the model simultaneously outputs "Tower Crane Boom" (box coordinates (x1, y1, w1, h1), confidence 0.99), "Suspended Rebar Bundle" (box coordinates (x2, y2, w2, h2), confidence 0.97), and below it, "Worker Walking" (box coordinates (x3, y3, w3, h3), category 'Personnel', confidence 0.98). By recording these discrete targets and calculating the geometric relationships of the bounding boxes, the model determines in real time whether the bounding box of the "worker" intersects with the predicted swing trajectory area of ​​the "suspended object," thereby achieving risk assessment.

[0052] Since the system described in Embodiment 2 of this application is a system used to implement the method of Embodiment 1 of this application, those skilled in the art can understand the specific structure and variations of the system based on the method described in Embodiment 1 of this application, and therefore will not be described again here. All systems used in the method of Embodiment 1 of this application fall within the scope of protection of this application.

[0053] Based on the same inventive concept, this application also provides a third embodiment, referring to... Figure 3 , Figure 3 This is a flowchart illustrating the third embodiment of the control method for the construction safety management system of this application.

[0054] In this embodiment, as described in step S30, generating the safety detection result corresponding to the construction scenario based on the prompt words using a large language model further includes steps S31 to S34: Step S31: Construct a search query based on the prompt words; Step S32: Match the search query with the construction safety knowledge base to retrieve at least one safety code provision or historical handling case related to the risk detection results, thereby forming enhanced contextual information; Step S33: Combine the enhanced contextual information with the prompt words to form enhanced prompt words; Step S34: Using a large language model and based on enhanced prompt words, generate safety inspection results corresponding to the construction scenario.

[0055] In this embodiment, the construction safety management system can extract keywords or vectorized query statements (i.e., retrieval statements) for querying the construction safety knowledge base based on the initial prompts. The construction safety knowledge base is a dedicated database storing authoritative texts such as regulations, standards, operating procedures, and accident cases. It may contain historical monitoring data, domain rule data, regulations, standards, and operating procedures from the construction safety management system. The construction safety management system can retrieve original text paragraphs or case summaries highly relevant to the current risk from the knowledge base, forming enhanced contextual information, and then combine this enhanced contextual information with the initial prompts to form enhanced prompts.

[0056] Specifically, after obtaining initial prompts describing risk scenarios, the construction safety management system analyzes them to extract core entities and risk concepts, and constructs a precise search query. Based on this query, the system performs a matching query in a local or cloud-based construction safety knowledge base, quickly locating and returning the most relevant regulatory provisions or historical case texts as enhanced contextual information. This contextual information is then combined with the initial prompts according to a predetermined template to form enhanced prompts containing specific references. The system submits these enhanced prompts to a large language model. When generating safety inspection results, the model prioritizes and integrates this authoritative knowledge, thereby outputting a well-founded and professionally recommended report.

[0057] The embodiments of this application enhance the generation through retrieval, making up for the shortcomings of large language models such as knowledge illusion and lack of timeliness, and greatly improving the accuracy, standardization and credibility of the final security detection results and decision support.

[0058] Since the system described in Embodiment 2 of this application is a system used to implement the method of Embodiment 1 of this application, those skilled in the art can understand the specific structure and variations of the system based on the method described in Embodiment 1 of this application, and therefore will not be described again here. All systems used in the method of Embodiment 1 of this application fall within the scope of protection of this application.

[0059] This application provides a control device for a construction safety management system. The device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the control method of the construction safety management system in the first embodiment described above.

[0060] The following is for reference. Figure 4The diagram illustrates a structural schematic of a control device suitable for implementing the construction safety management system of the embodiments of this application. The control device of the construction safety management system in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), vehicle terminals (e.g., vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 4 The control equipment of the construction safety management system shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0061] like Figure 4 As shown, the control equipment of the construction safety management system may include a processing unit 1001 (e.g., a core processor, a graphics processor, etc.), which can perform various appropriate actions and processes according to programs stored in read-only memory (ROM) 1002 or programs loaded from storage device 1003 into random access memory (RAM) 1004. The random access memory 1004 also stores various programs and data required for the operation of the control equipment of the construction safety management system. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the control equipment of the construction safety management system to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows control equipment for a construction safety management system with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented or possessed alternatively.

[0062] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from read-only memory 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0063] The control device for the construction safety management system provided in this application, employing the control method of the construction safety management system in the above embodiments, can solve the technical problem of insufficient comprehensive assessment of risks and hazards in construction safety inspections. Compared with the prior art, the beneficial effects of the control device for the construction safety management system provided in this application are the same as those of the control method for the construction safety management system provided in the above embodiments, and other technical features in the control device for the construction safety management system are the same as those disclosed in the method of the previous embodiment, and will not be repeated here.

[0064] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0065] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0066] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the control method of the construction safety management system in the above embodiments.

[0067] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, radio frequency (RF), etc., or any suitable combination thereof.

[0068] The aforementioned computer-readable storage medium may be included in the control equipment of the construction safety management system; or it may exist independently and not be installed in the control equipment of the construction safety management system.

[0069] The aforementioned computer-readable storage medium carries one or more programs. When the aforementioned one or more programs are executed by the control equipment of the construction safety management system, the control equipment of the construction safety management system: acquires visual feature information extracted from the construction scene by the visual perception model, determines the risk detection result of the visual feature information, and if the risk detection result indicates that there is a risk, constructs prompt words based on the risk detection result and the visual feature information, and generates the safety detection result corresponding to the construction scene based on the prompt words through a large language model.

[0070] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0071] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0072] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0073] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., computer programs) for executing the control method of the above-described construction safety management system, thereby solving the technical problem that the assessment of risks and hazards in construction safety inspections is not comprehensive enough. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the control method of the construction safety management system provided in the above embodiments, and will not be repeated here.

[0074] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A control method of a construction safety management system, characterized by, The method comprises the following steps: obtaining visual feature information extracted by a visual perception model based on a construction scene, and determining a risk detection result of the visual feature information; if the risk detection result is that there is a risk, constructing a prompt word according to the risk detection result and the visual feature information; generating a safety detection result corresponding to the construction scene based on the prompt word through a large language model.

2. The control method of the construction safety management system according to claim 1, characterized by, The step of obtaining visual feature information extracted by a visual perception model based on a construction scene, and determining a risk detection result of the visual feature information comprises: receiving the visual feature information output by the visual perception model, and determining construction scene information according to the visual feature information; obtaining risk identification rules corresponding to the construction scene according to the construction scene information; matching the risk identification rules with the visual feature information, and determining the risk detection result according to the matching result.

3. The control method of the construction safety management system according to claim 2, characterized by, The step of matching the risk identification rules with the visual feature information, and determining the risk detection result according to the matching result comprises: determining target position data of the construction scene according to the visual feature information or construction scene visual information; performing correlation analysis on the target position data, the visual feature information and the risk identification rules to generate risk information and a confidence degree corresponding to the risk information; when the confidence degree is lower than a confidence degree threshold, marking the risk information as existing risk, and generating the risk detection result according to the marking result.

4. The control method of the construction safety management system according to claim 1, characterized by, Before the step of obtaining visual feature information extracted by a visual perception model based on a construction scene, and determining a risk detection result of the visual feature information, the method further comprises: receiving construction scene visual information collected by a visual sensor, wherein the visual sensor comprises a wearable device, a camera and / or a drone; identifying a target object in the construction scene visual information and target feature information of the target object through the visual perception model; constructing the visual feature information according to the target object and the target feature information of the target object.

5. The control method of the construction safety management system according to claim 1, wherein The step of constructing a prompt word according to the risk detection result and the visual feature information if the risk detection result is that there is a risk comprises: if it is judged that there is a risk in the construction scene according to the risk detection result, obtaining a preset prompt word template; filling the risk detection result and the visual feature information into the prompt word template to form the prompt word.

6. The control method of the construction safety management system according to claim 5, wherein The step of filling the risk detection result and the visual feature information into the prompt word template to form the prompt word comprises: identifying at least one information placeholder related to a risk object, a risk behavior and / or a risk state in the prompt word template; extracting a target object, attribute information of the target object and environmental information from the visual feature information according to the risk detection result; matching the target object, the attribute information and the environmental information with the type of the information placeholder, and filling them into the prompt word template to form the prompt word.

7. The control method of the construction safety management system according to claim 5, wherein The step of filling the risk detection result and the visual feature information into the prompt word template to form the prompt word comprises: obtaining a risk identification rule corresponding to the construction scene; matching the risk detection result, the risk identification rule and the visual feature information with the fill-in word in the prompt word template; replacing the fill-in word in the prompt word template with the risk detection result, the risk identification rule and the visual feature information according to the matching result to form the prompt word.

8. The control method of the construction safety management system according to claim 1, wherein The step of generating the safety detection result corresponding to the construction scene based on the prompt word through the large language model further comprises: constructing a retrieval formula based on the prompt word; matching the retrieval formula with a construction safety knowledge base to retrieve at least one safety specification provision or historical disposal case related to the risk detection result to form enhanced context information; combining the enhanced context information with the prompt word to form an enhanced prompt word; generating the safety detection result corresponding to the construction scene based on the enhanced prompt word through the large language model.

9. A control device of a construction safety management system, characterized by, The device comprises a memory, a processor and a computer program stored on the memory and executable on the processor, and the computer program is configured to implement the steps of the control method of the construction safety management system according to any one of claims 1 to 8.

10. A storage medium, characterized by The storage medium is a computer readable storage medium, and the storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the control method of the construction safety management system according to any one of claims 1 to 8.

Images