System and method for identifying, analyzing and judging major accident potential based on AI large model

Abstract

The invention provides a major accident hidden danger identification analysis and judgment system and method based on an AI large model, and the system comprises a to-be-identified information obtaining subsystem which is used for obtaining to-be-identified information of a detection region; the model training subsystem is used for training an accident potential identification AI large model; the knowledge graph construction subsystem is used for constructing an industry field knowledge graph; and the hidden danger recognition subsystem is used for carrying out major accident hidden danger recognition analysis and judgment according to the to-be-recognized information on the basis of an accident hidden danger recognition AI large model and an industry domain knowledge graph. The invention discloses a major accident potential identification analysis and judgment system and method based on an AI large model. The method comprises the following steps: training an accident potential identification A I large model and constructing an industry field knowledge graph; after the to-be-identified information is obtained, the model automatically extracts the effective trigger word through a natural language processing technology in combination with the industry field knowledge graph, optimizes the hidden danger reasoning rule and automatically performs hidden danger reasoning, so that the hidden danger identification accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of AI large models, and particularly to a system and method for identifying, analyzing, and determining major accident hidden dangers based on an AI large model. Background Art

[0002] Currently, the criteria for determining major accident hidden dangers are simple and cannot effectively cover complex safety risks. Moreover, the determination method is relatively primitive and mainly relies on expert manual determination, making it impossible to effectively address increasingly complex safety hidden dangers. Affected by factors such as the lack of uniformity in inspection methods and standards and uneven personnel qualities, there are large blind spots and errors in manual inspections, affecting the accuracy and effectiveness of hidden danger inspections.

[0003] In view of this, there is an urgent need for a system and method for identifying, analyzing, and determining major accident hidden dangers based on an AI large model to at least address the above deficiencies. Summary of the Invention

[0004] One of the objectives of the present invention is to provide a system and method for identifying, analyzing, and determining major accident hidden dangers based on an AI large model, training an AI large model for identifying accident hidden dangers, and at the same time, constructing an industry domain knowledge graph by combining historical accident data, expert experience, and safety laws, regulations, and standards in the detection area; after the AI large model for identifying accident hidden dangers obtains the information to be identified, it automatically extracts effective trigger words through natural language processing technology in combination with the industry domain knowledge graph, optimizes the hidden danger reasoning rules, and automatically conducts hidden danger reasoning, avoiding manual inspection errors and improving the accuracy of hidden danger identification.

[0005] The system for identifying, analyzing, and determining major accident hidden dangers based on an AI large model provided by an embodiment of the present invention includes:

[0006] An information acquisition subsystem to be identified, configured to acquire the information to be identified in the detection area;

[0007] A model training subsystem, configured to train an AI large model for identifying accident hidden dangers;

[0008] A knowledge graph construction subsystem, configured to construct an industry domain knowledge graph;

[0009] A hidden danger identification subsystem, configured to perform identification, analysis, and determination of major accident hidden dangers based on the AI large model for identifying accident hidden dangers and the industry domain knowledge graph according to the information to be identified.

[0010] Preferably, the knowledge graph construction subsystem constructs an industry domain knowledge graph, including:

[0011] Obtaining historical accident data, expert experience, and safety laws, regulations, and standards;

[0012] Based on knowledge graph construction technology, identify and map the key factors and correlation relationships of major accident hazards according to historical accident data, expert experience, and safety laws, regulations, and standards, and obtain the knowledge graph of the industry domain.

[0013] Preferably, the hazard identification subsystem conducts identification analysis and determination of major accident hazards based on the accident hazard identification AI large model and the knowledge graph of the industry domain according to the information to be identified, including:

[0014] Based on natural language processing technology, extract the effective trigger words in the knowledge graph of the industry domain according to the information to be identified;

[0015] Optimize the hazard reasoning rules of the accident hazard identification AI large model according to the graph features of the effective trigger words in the knowledge graph of the industry domain.

[0016] Preferably, the hazard identification subsystem conducts identification analysis and determination of major accident hazards based on the accident hazard identification AI large model and the knowledge graph of the industry domain according to the information to be identified, and also includes:

[0017] Conduct image recognition analysis on the information to be identified to obtain the recognition image features;

[0018] Match the recognition image features with the suspicious hazard feature library obtained from the knowledge graph of the industry domain. If there is a match, obtain the hazard description corresponding to the matched suspicious hazard feature;

[0019] Continue to search in the knowledge graph of the industry domain for the determination criteria of the hazard type corresponding to the hazard description;

[0020] Based on the hazard description and the determination criteria, conduct identification analysis and determination of major accident hazards.

[0021] Preferably, the information to be identified acquisition subsystem acquires the information to be identified in the detection area, and also includes:

[0022] Acquire the area type label of the detection area;

[0023] Retrieve accident events according to the area type label;

[0024] Based on the accident events, obtain the basis for determining the target to be identified;

[0025] Acquire the area operation information of the detection area;

[0026] Based on the area operation information and the basis for determining the target to be identified, acquire the information to be identified in the detection area.

[0027] Preferably, the information to be identified acquisition subsystem obtains the basis for determining the target to be identified according to the accident events, including:

[0028] Analyze the accident event to obtain the first area task and the first manual analysis factor corresponding to the execution at the time of the accident event; the first manual analysis factor includes: the type of accident event, the analysis object that causes the accident event through subsequent manual analysis, and the analysis value of the analysis object.

[0029] Collect the first manual analysis factors of the first area tasks of the same task type to obtain the second manual analysis factor.

[0030] Divide the second manual analysis factors containing the same type of accident event into the same analysis group.

[0031] Obtain the occurrence frequency of the analysis object in the analysis group.

[0032] Traverse the analysis objects in descending order of the occurrence frequency, and use the currently traversed analysis object as the target analysis object.

[0033] Obtain the number of analysis values of the target analysis object.

[0034] If the number of analysis values is greater than or equal to the preset number threshold, directly obtain the standard deviation of the analysis values of the target analysis object.

[0035] If the number of analysis values is less than the preset number threshold, determine the reference analysis object that co-occurs with the target analysis object in the same analysis group.

[0036] Based on the transfer learning algorithm, according to the co-occurrence relationship of the analysis values of the reference analysis object and the target analysis object, obtain the extended analysis value of the target analysis object, and then obtain the standard deviation of the analysis values of the target analysis object according to the analysis values and the extended analysis values of the target analysis object.

[0037] If the standard deviation is less than or equal to the preset standard deviation threshold, associate the first area task, the type of accident event, the target analysis object, and the mean value of the analysis values corresponding to the target analysis object, and use it as the basis for determining the target to be recognized.

[0038] Preferably, the subsystem for obtaining information to be recognized obtains the information to be recognized in the detection area according to the regional operation information and the basis for determining the target to be recognized, including:

[0039] Analyze the regional operation information to obtain the second area task.

[0040] Match the second area task with the first area task in the basis for determining the target to be recognized. If they are the same, use the corresponding first area task as the third area task.

[0041] Based on the target recognition technology, detect the target analysis object corresponding to the third area task in the detection area to obtain the information to be recognized.

[0042] Preferably, the information acquisition subsystem to be recognized is based on the target recognition technology to detect the target analysis object corresponding to the third area task in the detection area, and obtain the information to be recognized, including:

[0043] Determine the regional position of the fixed target analysis object corresponding to the third area task;

[0044] Obtain the first observation condition of the average value of the first analysis value of the fixed target analysis object and associate it with the regional position;

[0045] Plan the mobile recognition device to go to the regional position closest to the position of the mobile recognition device in the execution area of the third area task, and mark the regional position as the initial position;

[0046] Plan the observation main line according to the position distribution of the initial position and the regional position in the execution area;

[0047] Correspondingly mark the first observation condition associated with the regional position on the observation main line to generate the first pilot information;

[0048] Control the mobile recognition device to move and take pictures based on the first pilot information;

[0049] Update the remaining pilot information of the current regional position according to the captured image of the current regional position, and control the mobile recognition device to move and take pictures based on the updated pilot information;

[0050] Repeat the system operation of the information acquisition subsystem to be recognized, which is to update the remaining pilot information of the current regional position according to the captured image of the current regional position, and control the mobile recognition device to move and take pictures based on the updated pilot information;

[0051] When all the regional positions have been photographed, the acquisition of the information to be recognized is completed.

[0052] Preferably, updating the remaining pilot information of the current regional position according to the captured image of the current regional position includes:

[0053] According to the captured image, detect the target analysis object corresponding to the third area task in the detection area, and obtain the first detection target;

[0054] Obtain the first detection targets that appear in the same captured image and for which the information to be recognized has not been obtained, and use them as the second detection targets;

[0055] If the acquisition is successful, obtain the second observation condition according to the average value of the second analysis value corresponding to the second detection target; the second observation condition includes: the observation point position and the observation posture;

[0056] According to the captured image, calculate the real-time distance between the mobile recognition device and the observation point position of the second detection target;

[0057] Plan the observation order of the second detection target in ascending order of the real-time distance;

[0058] Determine the second pilotage information according to the observation order and the observation attitude;

[0059] Place the second pilotage information in the front of the remaining pilotage information to obtain the updated pilotage information.

[0060] The method for identifying, analyzing and determining major accident hazards based on the AI large model provided by the embodiments of the present invention includes:

[0061] Step 1: Obtain the information to be identified in the detection area;

[0062] Step 2: Train the AI large model for identifying accident hazards;

[0063] Step 3: Construct the knowledge graph of the industry field;

[0064] Step 4: Based on the AI large model for identifying accident hazards and the knowledge graph of the industry field, conduct identification analysis and determination of major accident hazards according to the information to be identified.

[0065] The beneficial effects of the present invention are:

[0066] The present invention trains the AI large model for identifying accident hazards. At the same time, it constructs the knowledge graph of the industry field by combining the historical accident data, expert experience and safety laws, regulations and standards in the detection area. After the AI large model for identifying accident hazards obtains the information to be identified, it automatically extracts effective trigger words through natural language processing technology in combination with the knowledge graph of the industry field, optimizes the hidden danger reasoning rules and automatically conducts hidden danger reasoning, avoiding manual investigation errors and improving the accuracy of hidden danger identification.

[0067] Other features and advantages of the present invention will be described in the following specification, and, in part, will be obvious from the specification, or will be understood by implementing the present invention. The objectives and other advantages of the present invention can be realized and obtained by the structure specifically pointed out in this application document.

[0068] The technical solutions of the present invention will be further described in detail below through the drawings and embodiments. Description of the Drawings

[0069] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0070] Figure 1 It is a schematic diagram of the system for identifying, analyzing and determining major accident hazards based on the AI large model in the embodiments of the present invention;

[0071] Figure 2This is a schematic diagram of the method for identifying, analyzing, and determining major accident hazards based on an AI large model in an embodiment of the present invention. Detailed implementation manners

[0072] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0073] The embodiment of the present invention provides a system for identifying, analyzing, and determining major accident hazards based on an AI large model, as Figure 1 shown, including:

[0074] A to-be-identified information acquisition subsystem 1, configured to acquire to-be-identified information of a detection area;

[0075] Among them, the to-be-identified information acquisition subsystem acquires to-be-identified information of a detection area, including:

[0076] Acquiring to-be-identified information through a mobile terminal, an inspection robot, and a drone;

[0077] A model training subsystem 2, configured to train an accident hazard identification AI large model;

[0078] A knowledge graph construction subsystem 3, configured to construct a knowledge graph of the industry field;

[0079] Among them, the knowledge graph construction subsystem constructs a knowledge graph of the industry field, including:

[0080] Obtaining historical accident data, expert experience, and safety laws, regulations, and standards;

[0081] Based on knowledge graph construction technology, identifying and mapping key factors and correlation relationships of major accident hazards according to historical accident data, expert experience, and safety laws, regulations, and standards to obtain a knowledge graph of the industry field;

[0082] A hazard identification subsystem 4, configured to perform identification, analysis, and determination of major accident hazards based on the accident hazard identification AI large model and the knowledge graph of the industry field according to the to-be-identified information;

[0083] Among them, the hazard identification subsystem performs identification, analysis, and determination of major accident hazards based on the accident hazard identification AI large model and the knowledge graph of the industry field according to the to-be-identified information, including:

[0084] Extracting effective trigger words in the knowledge graph of the industry field based on natural language processing technology according to the to-be-identified information;

[0085] Optimizing the hazard inference rules of the accident hazard identification AI large model according to the graph features of the effective trigger words in the knowledge graph of the industry field.

[0086] In this embodiment, the detection area is the area where potential accident hazards need to be investigated, such as the construction area of a substation, the construction area of a construction site, etc.

[0087] In this embodiment, the information to be recognized is the relevant information used to determine whether there are potential accident hazards, such as video data, image data, etc. of the detection area. The information to be recognized can be obtained by shooting the detection area with devices equipped with cameras such as mobile trolleys and inspection robots;

[0088] In this embodiment, the AI large model for identifying potential accident hazards is an AI model obtained by learning the inference and analysis records of potential accident hazards by humans based on deep learning technology, and is used to analyze the risk of potential accident hazards in the detection area according to the input analysis basis (such as hazard analysis images, etc.).

[0089] In this embodiment, the historical accident data is the accident records that have occurred in the detection area in history;

[0090] In this embodiment, the effective trigger word is: first, obtain the description word by describing the information to be recognized in words, and then match the description word with the entity description words and relationship description words in the industry domain knowledge graph to obtain the entity description word or relationship description word that matches;

[0091] In this embodiment, the graph feature of the effective trigger word in the industry domain knowledge graph is: the entity or relationship related to the entity or relationship corresponding to the effective trigger word in the industry domain knowledge graph.

[0092] The working principle and beneficial effects of the above technical solution are:

[0093] The present invention trains the AI large model for identifying potential accident hazards, and at the same time constructs an industry domain knowledge graph by combining the historical accident data, expert experience and safety laws, regulations and standards of the detection area; after the AI large model for identifying potential accident hazards obtains the information to be recognized, it automatically extracts effective trigger words by combining the industry domain knowledge graph through natural language processing technology, optimizes the hazard inference rules and automatically performs hazard inference, avoiding manual investigation errors and improving the accuracy of hazard identification.

[0094] In one embodiment, the hazard identification subsystem is based on the AI large model for identifying potential accident hazards and the industry domain knowledge graph, and performs major potential accident hazard identification analysis and determination according to the information to be recognized, and further includes:

[0095] Perform image recognition analysis on the information to be recognized to obtain the recognition image features; wherein, the recognition image features are: the texture, color, etc. of the image in the information to be recognized;

[0096] Match the identified image features with the suspicious hidden danger feature library obtained from the industrial domain knowledge graph. If there is a match, obtain the hidden danger description corresponding to the suspicious hidden danger feature that matches; among them, the suspicious hidden danger feature library includes: image features extracted from accident images that may cause hidden dangers, such as: image features of worn safety ropes; the hidden danger description is, for example: "The safety rope is worn, and the safety rope may break."

[0097] Continue to search in the industrial domain knowledge graph for the determination criteria for the hidden danger type corresponding to the hidden danger description; among them, the hidden danger type corresponding to the hidden danger description is: the type of hidden danger of the hidden danger description, for example: the hidden danger description is "The safety rope may break", and its corresponding hidden danger type is: safety rope break; the determination criteria are: the overall determination basis for the hidden danger type, for example: to what extent the wear degree reaches, and how heavy the load-bearing object is to determine the accident hidden danger of the safety rope break

[0098] Based on the hidden danger description and the determination criteria, conduct identification analysis and determination of major accident hidden dangers.

[0099] The working principle and beneficial effects of the above technical solution are as follows:

[0100] The present invention performs image recognition analysis on the information to be recognized to obtain the identified image features; docks with the suspicious hidden danger feature library obtained from the industrial domain knowledge graph, matches the identified image features with the suspicious hidden danger features in the suspicious hidden danger feature library, determines the hidden danger description corresponding to the suspicious hidden danger feature that matches; after obtaining the hidden danger description, the hidden danger type can be determined, then search in the industrial domain knowledge graph for the determination criteria corresponding to the hidden danger type, and then compare the hidden danger description with the determination criteria. If it meets the criteria, confirm and output the suspicious hidden danger, otherwise determine that there is no hidden danger, optimize the hidden danger reasoning rules and automatically perform hidden danger reasoning, which is more intelligent.

[0101] In one embodiment, the information acquisition subsystem for information to be recognized acquires the information to be recognized in the detection area, including:

[0102] Acquire the information to be recognized through mobile terminals, inspection robots, and drones.

[0103] The working principle and beneficial effects of the above technical solution are as follows:

[0104] The present invention introduces multiple service methods to provide the function of identifying major accident hidden dangers, improving the comprehensiveness of identification.

[0105] In one embodiment, the information acquisition subsystem for information to be recognized acquires the information to be recognized in the detection area, and further includes:

[0106] Obtain the area type label of the detection area; the area type label is: an ID identifying the category of the detection area, such as: "substation", "automobile manufacturing factory", etc.

[0107] Retrieve accident events according to the area type label; where the accident events are anomalies and accidents that occurred in the area of this area type in history, such as: electric shock of staff, burning of substation equipment, etc.;

[0108] Obtain the basis for determining the target to be recognized according to the accident events;

[0109] Among them, obtaining the basis for determining the target to be recognized according to the accident events includes:

[0110] Analyze the accident events to obtain the first area task and the first artificial analysis factor executed when the accident events occurred; the first artificial analysis factor includes: the type of accident event, the analysis object of the artificial subsequent analysis that caused the accident event, and the analysis value of the analysis object; the first area task is the task being executed in the area where the accident event occurred when the accident event occurred, such as: "substation equipment maintenance"; the type of accident event is: the type of accident event, such as: "electric shock"; the analysis object is: the analysis target in the area corresponding to the first area task when making a post-event attribution of the accident event by artificial, such as: staff, equipment being maintained; the analysis value of the analysis object is: the state of the analysis object that caused the accident when making a post-event attribution of the accident event by artificial, such as: non-standard operation of staff (such as: not wearing insulating gloves), the state of the internal components of the equipment being maintained that caused the equipment being maintained to leak electricity (such as: the insulation skin at the junction of line A and line B is damaged);

[0111] Collect the first artificial analysis factors of the first area tasks of the same task type to obtain the second artificial analysis factor;

[0112] Divide the second artificial analysis factors containing the same type of accident event into the same analysis group;

[0113] Obtain the occurrence frequency of the analysis object in the analysis group; the occurrence frequency is: the result obtained by dividing the number of second artificial analysis factors in which the analysis object exists in the same analysis group by the number of all second artificial analysis factors in the analysis group;

[0114] Traverse the analysis objects in order from largest to smallest according to the occurrence frequency, and use the currently traversed analysis object as the target analysis object;

[0115] Obtain the number of analysis values of the target analysis object;

[0116] If the number of analysis values is greater than or equal to the preset number threshold, directly obtain the standard deviation of the analysis values of the target analysis object; where the preset number threshold is set in advance by artificial;

[0117] If the number of analysis values is less than a preset number threshold, determine the reference analysis objects that co-occur with the target analysis object in the same analysis group; the reference analysis objects that co-occur with the target analysis object are: other analysis objects that accompany the target analysis object in the second artificial analysis factor in the same analysis group;

[0118] Based on the transfer learning algorithm, according to the co-occurrence relationship of the analysis values of the reference analysis object and the target analysis object, obtain the extended analysis value of the target analysis object, and then obtain the standard deviation of the analysis value of the target analysis object according to the analysis value and the extended analysis value of the target analysis object; the co-occurrence relationship of the analysis values is: the correlation relationship between the analysis value of the reference analysis object and the analysis value of the target analysis object, for example: the relationship between the adding step of the artificial insulating oil and the content of the insulating oil leaked into the environment; the extended analysis value is: using the transfer learning algorithm to learn the co-occurrence relationship of the analysis values of the reference analysis object and the target analysis object and applying it to the second artificial analysis factor that does not include the target analysis object but includes the reference analysis object, and predicting the prediction result of the state of the accident caused by the missing analysis of the target analysis object in the corresponding second artificial analysis factor;

[0119] If the standard deviation is less than or equal to a preset standard deviation threshold, associate the first area task, the type of accident event, the target analysis object, and the mean value of the analysis value corresponding to the target analysis object, and use it as the basis for determining the target to be identified; the preset standard deviation threshold is set manually by the user;

[0120] Obtain the area operation information of the detection area; the area operation information is: the current task type and execution location of the detection area;

[0121] According to the area operation information and the basis for determining the target to be identified, obtain the information to be identified in the detection area;

[0122] Among them, according to the area operation information and the basis for determining the target to be identified, obtaining the information to be identified in the detection area includes:

[0123] Parse the area operation information to obtain the second area task; the second area task is: the task currently being executed in the detection area, for example: transformer inspection;

[0124] Match the second area task with the first area task in the basis for determining the target to be identified. If they are the same, use the corresponding first area task as the third area task;

[0125] Based on the target recognition technology, detect the target analysis object corresponding to the third area task in the detection area to obtain the information to be identified.

[0126] The working principle and beneficial effects of the above technical solution are:

[0127] The types of accident events that occur in different types of detection areas are also different. First, the present invention retrieves the accident events in the history of areas of the same area type based on the area type label, and extracts the basis for determining the target to be recognized from the accident events. The specific extraction process is as follows:

[0128] Parse and obtain the first area task and the first manual analysis factor in the accident event, collect the first manual analysis factors of the same task type, and divide the second manual analysis factors including the same accident event type into the same analysis group; however, not all manual analysis factors are credible. Therefore, the analysis results of different analysts for the same type of analysis task can be obtained, and the analysis results are compared. Here, the standard deviation is used to characterize the difference in the analysis results of different analysts. However, when the analysis result data of different analysts for the same type of analysis task is insufficient (that is, the number of analysis values is less than the number threshold), it is not accurate enough to use the standard deviation to characterize the difference in the analysis results of different analysts. Considering the situation of missing analysis by analysts in the analysis data of accident events, the transfer learning algorithm and existing data are used to obtain the extended analysis values of the target analysis object, and then the standard deviation of the analysis values of the target analysis object is obtained based on the analysis values and extended analysis values of the target analysis object;

[0129] Compare the standard deviation with the introduced standard deviation threshold. If the standard deviation is less than or equal to the preset standard deviation threshold, associate the first area task, the accident event type, the target analysis object, and the mean value of the analysis values corresponding to the target analysis object as the basis for determining the target to be recognized.

[0130] After determining the basis for determining the target to be recognized, determine the second area task based on the area operation information, and match the second area task with the first area task in the basis for determining the target to be recognized. If they are consistent, use the target analysis object corresponding to the first area task as the target to be recognized; detect the target to be recognized in the detection area based on the target recognition technology for subsequent acquisition of the information to be recognized, improving the recognition efficiency.

[0131] In one embodiment, the information acquisition subsystem to be recognized, based on the target recognition technology, detects the target analysis object corresponding to the third area task in the detection area to obtain the information to be recognized, including:

[0132] Determine the regional location of the fixed target analysis object corresponding to the third area task; the fixed target analysis object is: the target analysis object with an unchanged position corresponding to the third area task;

[0133] Obtain the first observation condition of the first mean value of the analysis values of the fixed target analysis object and associate it with the regional location; the first mean value of the analysis values is the mean value of the analysis values corresponding to the fixed target analysis object, for example: transformer oil leakage; the first observation condition is: the condition under which the first mean value of the analysis values can be obtained, for example: shooting towards the tank seal ring, shooting distance less than 50 cm;

[0134] Plan the area position in the execution area of the task of moving the mobile recognition device to the third area that is the closest to the position of the mobile recognition device, and mark the area position as the initial position;

[0135] According to the distribution of the initial position and the area positions in the execution area, plan the main observation line; when planning the main observation line, based on the route planning technology, use the initial position as the starting point of the route planning and plan the shortest route passing through each area position;

[0136] Correspondingly mark the first observation conditions associated with the area positions on the main observation line to generate the first pilotage information;

[0137] Control the mobile recognition device to move and take pictures based on the first pilotage information;

[0138] Update the remaining pilotage information of the current area position according to the captured image of the current area position, and control the mobile recognition device to move and take pictures based on the updated pilotage information;

[0139] Repeat the system operation of updating the remaining pilotage information of the current area position according to the captured image of the current area position and controlling the mobile recognition device to move and take pictures described in the subsystem for obtaining information to be recognized;

[0140] When all area positions have been photographed, complete the acquisition of the information to be recognized;

[0141] Among them, updating the remaining pilotage information of the current area position according to the captured image of the current area position includes:

[0142] According to the captured image, detect the target analysis object corresponding to the third area task in the detection area to obtain the first detection target; implemented based on the target detection technology;

[0143] Obtain the first detection targets that appear in the same captured image and for which the information to be recognized has not been obtained, and use them as the second detection targets;

[0144] If the acquisition is successful, obtain the second observation condition according to the average value of the second analysis values corresponding to the second detection targets; the second observation condition includes: the observation point position and the observation posture; the average value of the second analysis values is: the average value of the analysis values corresponding to the target analysis objects at non-fixed positions, such as: the non-standard operation behaviors of maintenance personnel;

[0145] According to the captured image, calculate the real-time distance between the mobile recognition device and the observation point position of the second detection target;

[0146] Plan the observation order of the second detection targets in ascending order of the real-time distance;

[0147] Determine the second pilotage information according to the observation sequence and the observation attitude;

[0148] Prepose the second pilotage information in the remaining pilotage information to obtain updated pilotage information.

[0149] The working principle and beneficial effects of the above technical solution are as follows:

[0150] The target analysis object is divided into a movable object (such as a substation worker) and a fixed object. The observation position of the fixed object is fixed and unchanged. Therefore, based on the regional position of the fixed target analysis object corresponding to the third regional task, according to the regional position, determine the position distribution of the initial position and the regional position in the execution area and plan the main observation line. Mark the first observation condition of the average value of the first analysis value of the fixed target analysis object on the main observation line to obtain the first pilotage information. Control the mobile recognition device to move and take pictures based on the first pilotage information. When moving and taking pictures, obtain the movable object (the second detection target) according to the captured image of the current regional position. Determine the second observation condition (the observation point and the observation attitude of the movable object) according to the average value of the second analysis value of the movable object; use the depth camera to calculate the real-time distance between the mobile recognition device and the observation point of the second detection target according to the captured image, and plan the observation sequence of the second detection target in ascending order of the real-time distance. Based on the observation sequence and the observation attitude, determine the second pilotage information, prepose the second pilotage information in the remaining pilotage information to obtain updated pilotage information, and repeatedly execute the system operation of the subsystem for obtaining information to be recognized to update the remaining pilotage information of the current regional position according to the captured image of the current regional position and control the movement and shooting of the mobile recognition device based on the updated pilotage information. During the process of analyzing the fixed target analysis object, the movable object is captured and analyzed based on the necessary captured images, and the pilotage information is dynamically determined, greatly improving the analysis and determination efficiency.

[0151] The embodiment of the present invention provides a method for identifying, analyzing and determining major accident hazards based on an AI large model, as Figure 2 shown, including:

[0152] Step 1: Obtain the information to be recognized in the detection area;

[0153] Step 2: Train the AI large model for accident hazard identification;

[0154] Step 3: Construct an industry domain knowledge graph;

[0155] Step 4: Based on the AI large model for accident hazard identification and the industry domain knowledge graph, conduct major accident hazard identification, analysis and determination according to the information to be recognized.

[0156] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalent technologies, the present invention is also intended to include these modifications and variations.

Claims

1. A major accident hazard identification, analysis and determination system based on AI big model, characterized by: include: The to-be-identified information acquisition subsystem is used to acquire the to-be-identified information of the detection area; Model training subsystem, used to train the large AI model for accident hazard identification; The knowledge graph construction subsystem is used to construct industry knowledge graphs; The hidden danger identification subsystem is used to identify, analyze and judge major accident hazards based on the accident hazard identification AI large model and industry knowledge graph and the information to be identified.

2. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 1 is characterized in that: The knowledge graph construction subsystem constructs industry knowledge graphs, including: Obtain historical accident data, expert experience, and safety laws, regulations and standards; Based on knowledge graph construction technology, according to historical accident data, expert experience and safety laws, regulations and standards, the key factors and correlations of major accident hazards are identified and mapped to obtain industry knowledge graphs.

3. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 1 is characterized in that: The hidden danger identification subsystem is based on the accident hidden danger identification AI big model and industry knowledge graph, and performs major accident hidden danger identification, analysis and judgment according to the information to be identified, including: Based on natural language processing technology and the information to be identified, effective trigger words are extracted from the industry knowledge graph; Optimize the hidden danger reasoning rules of the accident hidden danger identification AI model based on the graph features of effective trigger words in the industry knowledge graph.

4. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 1 is characterized in that: The hidden danger identification subsystem is based on the accident hidden danger identification AI big model and industry knowledge graph, and performs major accident hidden danger identification, analysis and judgment according to the information to be identified. It also includes: Perform image recognition analysis on the information to be identified to obtain identification image features; Match the identified image features with the suspicious hidden danger feature library obtained from the industry knowledge graph. If there is a match, obtain the hidden danger description corresponding to the matched suspicious hidden danger feature; Continue to search for the criteria for determining the hidden danger type corresponding to the hidden danger description in the industry knowledge graph; Identify, analyze and determine major accident hazards based on hazard descriptions and determination criteria.

5. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 1 is characterized in that: The information to be identified acquisition subsystem acquires the information to be identified in the detection area, including: Obtain the information to be identified through mobile terminals, inspection robots and drones.

6. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 1 is characterized in that: The to-be-identified information acquisition subsystem acquires the to-be-identified information of the detection area, and also includes: Get the region type label of the detection region; Retrieve accident events based on area type labels; According to the accident event, obtain the basis for determining the target to be identified; Obtain regional operation information of the detection area; According to the regional operation information and the basis for determining the target to be identified, the information to be identified in the detection area is obtained.

7. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 6 is characterized in that: The information acquisition subsystem to be identified obtains the information to be identified in the detection area according to the regional operation information and the basis for determining the target to be identified, including: Analyze regional operation information and obtain the second regional tasks; Match the second area task with the first area task in the basis for determining the target to be identified. If they are consistent, use the corresponding first area task as the third area task; Based on the target recognition technology, the target analysis object corresponding to the third area task of the detection area is detected to obtain the information to be recognized.

8. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 7 is characterized in that: The information to be identified acquisition subsystem detects the target analysis object corresponding to the third area task of the detection area based on the target recognition technology, and obtains the information to be identified, including: Determine the regional location of the fixed target analysis object corresponding to the third regional task; Obtaining a first observation condition of a first analysis value mean of a fixed target analysis object and associating it with a regional position; Planning the mobile identification device to go to a regional position closest to the position of the mobile identification device in the execution area of ​​the third regional task, and marking the regional position as the initial position; Plan the main observation line according to the distribution of the initial position and the regional position in the execution area; Marking the first observation condition associated with the regional position on the observation main line accordingly to generate first pilotage information; Controlling the mobile identification device to move and shoot based on the first pilotage information; Update the remaining pilotage information of the current area location according to the captured image of the current area location, and control the movement and shooting of the mobile identification device based on the updated pilotage information; Repeat the steps of updating the remaining navigation information of the current area according to the captured image of the current area described in the information acquisition subsystem to be identified, and controlling the movement and shooting of the mobile identification device based on the updated navigation information; When all the area locations are photographed, the acquisition of the information to be identified is completed.

9. The major accident hazard identification, analysis and determination system based on AI large model as claimed in claim 8, characterized in that: The information acquisition subsystem to be identified updates the remaining pilotage information of the current area location according to the captured image of the current area location, including: Detecting a target analysis object corresponding to a third area task in the detection area according to the captured image to obtain a first detection target; Acquire a first detection target that appears in the same captured image and has not acquired information to be identified, and use it as a second detection target; If the acquisition is successful, the second observation condition is obtained according to the second analysis value mean corresponding to the second detection target; the second observation condition includes: observation point position and observation posture; Calculate the real-time distance between the observation point of the mobile identification device and the second detection target according to the captured image; Plan the observation order of the second detection target according to the order of real-time distance from small to large; Determine the second pilotage information according to the observation sequence and observation attitude; The second pilotage information is placed in front of the remaining pilotage information to obtain updated pilotage information.

10. A method for identifying, analyzing and determining major accident hazards based on an AI large model, characterized in that: include: Step 1: Obtain the information to be identified in the detection area; Step 2: Train the AI ​​model for accident hazard identification; Step 3: Build industry domain knowledge graph; Step 4: Based on the accident hazard identification AI big model and industry knowledge graph, conduct major accident hazard identification, analysis and judgment according to the information to be identified.

Images