A method, system and device for generating emergency plans

Abstract

The present invention relates to the field of emergency management technology, and in particular to a method, system and device for generating an emergency plan. The method comprises the following steps: acquiring target environment image data; extracting target environment activities from the target environment image data to obtain target environment activity data; detecting abnormal information from the target environment activity data to obtain abnormal information data; obtaining emergency strategy data based on abnormal information data and a preset emergency strategy database; obtaining abnormal processing priority data based on the emergency strategy data and the abnormal information data; acquiring current resource data, and obtaining emergency strategy correction data based on the current resource data and the abnormal processing priority data; acquiring target object historical log data, and correcting it based on the target object historical log data and the emergency strategy correction data to obtain emergency strategy plan data. The present invention ensures the efficient execution of emergency strategies and minimizes risks and losses through a resource scheduling mechanism.

Description

Technical Field

[0001] The present invention relates to the technical field of emergency management, and in particular to an emergency plan generation method, system and device. Background Art

[0002] With the development and widespread application of information technology, timely response to abnormal events has become particularly important, especially in confidential places, protection areas and other environments. Abnormal events in such environments mainly involve illegal entry into protection areas (such as the top floor near confidential places), climbing over fences, illegal photography, etc. In order to ensure the safety of confidential areas and the effectiveness of protection facilities, corresponding emergency plans must be formulated and implemented for these abnormal events.

[0003] Existing emergency plans are usually based on preset fixed strategies and cannot flexibly respond to different types of abnormal events that occur in real time. Due to the lack of dynamic analysis of real-time data, the efficiency and effectiveness of emergency response are low, the processing priority is not clear, and timely adjustments cannot be made according to the severity and scope of the incident. When faced with complex scenarios in the protection area, fixed strategies will lead to waste of resources, slow execution of emergency measures, or in some cases, the inability to curb abnormal behavior in a timely manner. Summary of the invention

[0004] In order to solve the above technical problems, the present invention proposes an emergency plan generation method, system and device to solve at least one of the above technical problems.

[0005] The present application provides a method for generating an emergency plan, the method comprising:

[0006] S1. Collect target environment images through the monitoring system to obtain target environment image data; extract target environment activities from the target environment image data to obtain target environment activity data; perform abnormal information detection on the target environment activity data to obtain abnormal information data;

[0007] S2. Matching the abnormal information data with a preset emergency strategy database to obtain emergency strategy data; sorting the abnormal processing priorities according to the emergency strategy data and the abnormal information data to obtain abnormal processing priority data;

[0008] S3, obtaining current resource data, and matching resources according to the current resource data and exception handling priority data to obtain emergency strategy correction data;

[0009] S4. Obtain historical log data of the target object, and make corrections based on the historical log data of the target object and the emergency strategy correction data to obtain emergency strategy plan data.

[0010] In the present invention, the image data of the target environment is collected and processed by the monitoring system, so that real-time and efficient abnormal information detection can be achieved. By intelligently matching the abnormal information data with the emergency strategy database, the system can select the most appropriate emergency strategy for different abnormal situations, thereby improving the efficiency and pertinence of emergency response. Especially in complex or changeable environments, intelligent strategy matching can quickly form a processing plan to avoid repeated execution of inappropriate strategies. By matching and scheduling resources according to the priority of abnormal processing, it is ensured that important events are handled first and resources are optimally utilized.

[0011] Optionally, S1 includes:

[0012] Collect target environment images through the monitoring system to obtain target environment image data;

[0013] Perform multi-scale image feature extraction on the target environment image data to obtain target environment image feature data;

[0014] Perform multi-layer semantic segmentation on the target environment image feature data to obtain target environment segmentation image data;

[0015] Performing activity behavior detection on the target environment segmentation image data to obtain the target environment activity behavior data;

[0016] Perform abnormal information detection on the target environment activity behavior data to obtain abnormal information data.

[0017] In the present invention, real-time target environment image acquisition is performed through the monitoring system, which can ensure real-time monitoring of the on-site conditions in the environment of security-sensitive places, regardless of day or night. The system can capture abnormal activities occurring in the target environment, ensuring efficient security protection. Multi-scale feature extraction is performed on the target environment image data, which can capture the features of different detail levels in the image, ensuring that the system can identify targets or events of various scales. Multi-layer semantic segmentation can accurately classify different areas in the image and distinguish different scenes, objects and personnel. By segmenting the objects in the image, the system can accurately identify the activity status of each area in the target environment, which helps to better understand the actions and behaviors in the scene. By performing activity behavior detection on the target environment segmentation image data, the movement trajectory and behavior pattern of people or objects can be analyzed. This step can effectively identify potential security threats (such as abnormal actions of unauthorized personnel or abnormal movement of equipment), providing a basis for timely response. Abnormal information detection is performed on the activity behavior data, and activities that violate normal behavior patterns can be identified in a timely manner.

[0018] Optionally, the abnormal information detection includes:

[0019] Performing person recognition on the target environment activity behavior data to obtain person recognition data;

[0020] Perform object recognition on the target environment activity behavior data to obtain object recognition data;

[0021] Perform anomaly detection based on the person recognition data and the object recognition data to obtain first anomaly detection data;

[0022] Performing behavior component detection according to the person recognition data to obtain person behavior component data;

[0023] Performing anomaly detection on the character behavior component data to obtain second anomaly detection data;

[0024] The first anomaly detection data and the second anomaly detection data are integrated to obtain anomaly information data.

[0025] The present invention combines person recognition, object recognition and behavior component analysis, and comprehensively covers multiple dimensions of anomaly detection. By identifying people and objects, not only can active objects be detected, but also the behavior of these objects can be analyzed to see if they meet expectations, which significantly improves the accuracy of anomaly detection. Through person recognition and object recognition, the system can quickly capture the identities of people and objects in the monitoring scene and generate specific identification data. This step can effectively distinguish known people from strangers, or detect the presence of unauthorized objects, thereby triggering subsequent anomaly detection. By introducing person behavior component detection, the specific actions of the person are decomposed and analyzed, and subtle abnormal behavior changes can be detected. Through the detection of such behavior components, obvious abnormal behaviors can be found. The present invention integrates the first anomaly detection data (based on person and object recognition) with the second anomaly detection data (based on behavior component analysis) to analyze abnormal situations at multiple levels. This multi-level integration can reduce the false alarm rate and improve the reliability of anomaly detection. By combining person recognition data and object recognition data, the system can analyze the interactive behavior between people and objects. For example, when an unauthorized person is detected in contact with a sensitive object, the system can quickly identify and trigger an alarm.

[0026] Optionally, S2 includes:

[0027] Evaluate the impact scope of the event based on the abnormal information data to obtain the impact scope data of the abnormal event;

[0028] Extracting scenario features from abnormal information data to obtain abnormal scenario feature data;

[0029] The emergency strategy data is obtained by matching the abnormal event impact range data, the abnormal scenario feature data and the preset emergency strategy database;

[0030] Exception handling priorities are sorted according to the emergency strategy data and the exception information data to obtain exception handling priority data.

[0031] In the present invention, by evaluating the scope of impact of the abnormal information data, the system can predict the expansion scope and potential impact of the event, determine which areas, personnel or resources are affected, and help identify the areas most likely to be threatened, thereby ensuring that the emergency response can give priority to covering key areas. By extracting the scenario features of the abnormal information data, the system can analyze the specific scenarios (such as time, location, environmental conditions, etc.) of the abnormal event, which helps the system to better understand the background of the event during the emergency response, thereby adjusting the strategy. For example, events occurring at night require different emergency strategies than those occurring during the day. By matching the comprehensive abnormal event impact range data, abnormal scenario feature data and the preset emergency strategy database, the system can recommend the most suitable emergency strategy for abnormal events of different types and situations, ensuring that the system executes the most appropriate emergency plan according to the impact range and scenario characteristics of the event, and avoiding unnecessary waste of resources. By sorting the priority of abnormal processing according to the emergency strategy data and the abnormal information data, the system can dynamically adjust the processing order according to the urgency and impact range of the event. High-priority events will be processed in a timely manner, while low-priority events will be processed later according to the resource situation, ensuring the optimal use of limited resources and ensuring that key events can be responded to first.

[0032] Optionally, exception handling priority ranking includes:

[0033] Extracting abnormal features from abnormal information data to obtain abnormal feature data;

[0034] Perform abnormal expansion trend processing according to abnormal characteristic data to obtain abnormal expansion trend data;

[0035] Priority sorting is performed according to the abnormal expansion trend data and the emergency strategy data to obtain first abnormal processing priority data;

[0036] Perform utility priority sorting according to the emergency strategy data and the exception information data to obtain second exception handling priority data;

[0037] A weighted priority calculation is performed on the emergency strategy data according to the first exception handling priority data and the second exception handling priority data to obtain the exception handling priority data.

[0038] In the present invention, by extracting abnormal features from abnormal information data, the system can extract key information from the abnormal data, such as the type, severity, and scope of impact of the abnormality, so that the system can more accurately understand the nature of each abnormal event. By processing the abnormal expansion trend of the abnormal feature data, the system can predict the development trend of the abnormal event (such as the scope of diffusion, the speed of increasing severity, etc.). Priority sorting is performed according to the abnormal expansion trend data and the emergency strategy data to ensure that the most rapidly expanding abnormal events are given priority. Especially in the case of fast abnormal expansion speed or wide impact range, the response speed of the system can be effectively improved to prevent the situation from deteriorating. According to the utility priority sorting, the system can evaluate the utility of each emergency strategy (such as processing efficiency, resource consumption, success rate, etc.), so as to allocate appropriate resources for different abnormal events. By evaluating the utility of the emergency strategy, the system can ensure that resources are optimally utilized and reduce resource waste or unnecessary expenses. The weighted priority calculation realizes multi-dimensional priority calculation by combining the first abnormal processing priority data (based on the expansion trend) and the second abnormal processing priority data (based on the utility sorting).

[0039] Optionally, S3 includes:

[0040] Get current resource data;

[0041] Construct a resource scheduling model based on current resource data to obtain a resource scheduling model;

[0042] Calculate resource load according to the resource scheduling model to obtain the resource scheduling load model;

[0043] Resources are matched according to the resource scheduling load model and exception handling priority data to obtain emergency strategy correction data.

[0044] In the present invention, by acquiring the current resource data in real time, the system can fully grasp the status of the current available resources, including personnel, equipment, time and location, etc., so that the system can dynamically adjust the resource scheduling scheme, adapt to the complex and changeable emergency environment, and ensure rapid response when resources change. Based on the current resource data, the resource scheduling model is constructed, and the system can intelligently schedule appropriate resources according to the specific needs of the event (such as event type, scale and processing priority). Through the resource scheduling model, it can be ensured that the most suitable resources are optimally allocated, reduce resource waste, and improve the efficiency of overall emergency processing. Resource load calculation is a key step to ensure reasonable resource scheduling. By calculating the load of each resource (such as current task, available time, remaining capacity, etc.), the system can avoid resource overload or uneven distribution, ensure that all resources are balanced in the emergency processing, and neither cause resource waste nor avoid excessive use or shortage of resources. By combining the resource scheduling load model and exception processing priority data for resource matching, the system can prioritize the allocation of appropriate resources for the most urgent events.

[0045] Optionally, the resource scheduling model construction includes:

[0046] Classify resource functions according to current resource data to obtain current resource operation label data;

[0047] Calculate resource scheduling time and analyze resource spatial distribution based on current resource data and current resource job tag data to obtain resource scheduling time data and resource spatial distribution data;

[0048] A resource scheduling model is constructed according to resource scheduling time data and resource spatial distribution data to obtain a resource scheduling model.

[0049] In the present invention, by classifying resource functions according to current resource data, the system can identify the specific purpose, function and applicable scenario of each resource, and assign a job label to each resource, so that the system can clearly understand which resources are applicable to which tasks during scheduling, thereby improving the pertinence and efficiency of scheduling and avoiding the misuse or improper allocation of resources. By calculating the resource scheduling time according to the current resource data and the job label data, the system can accurately predict the time required for each resource from scheduling to executing the task. By analyzing the spatial distribution of the current resources, the system can understand the geographical location of the resources and their spatial relationship with the location of the emergency event. Combined with the resource scheduling time data, the system can optimize the resource scheduling route and reduce the moving time of the resources from the current location to the task location. By integrating the resource scheduling time data and the spatial distribution data to construct the resource scheduling model, the system can generate a comprehensive scheduling plan based on time and space. The scheduling model can achieve the optimal configuration of resources, not only ensuring that the resources can arrive at the task location in time, but also reducing the redundant use and excessive mobilization of resources through reasonable allocation.

[0050] Optionally, S4 includes:

[0051] Get the historical log data of the target object;

[0052] Extract similar events based on the historical log data and abnormal information data of the target object to obtain similar historical event data;

[0053] Based on similar historical event data, the emergency strategy effect is traced back to obtain historical strategy effect data;

[0054] Perform strategy comparison based on historical strategy effect data and emergency strategy revision data to obtain strategy comparison data;

[0055] The emergency strategy correction data is corrected according to the strategy comparison data to obtain the emergency strategy plan data.

[0056] In the present invention, by obtaining the historical log data of the target object, the system can access past abnormal events and emergency handling records, which provides valuable reference information for current events. Analysis based on historical data can avoid blind decision-making and ensure that the emergency response strategy is optimized and adjusted based on actual experience. By extracting similar events from historical log data, the system can automatically find historical events similar to the current abnormal event. Through the backtracking of emergency strategy effects, the system can evaluate the emergency handling effects in historical events, including the efficiency, success rate, response time, etc. of strategy execution. The system can quantitatively analyze the execution of past strategies to ensure that the current emergency strategy is optimized based on actual results. By comparing the current emergency strategy with the strategies in similar historical events, the system can identify the advantages and disadvantages of the current strategy and correct the existing strategy through historical successful experience. Strategy comparison ensures the rationality of decision-making and improves the reliability of emergency response through continuous optimization.

[0057] Optionally, the present application further provides an emergency plan generation system, which is used to execute the emergency plan generation method as described above, and the emergency plan generation system includes:

[0058] The target environment activity abnormal information detection module is used to collect target environment images through the monitoring system to obtain target environment image data; extract target environment activities from the target environment image data to obtain target environment activity data; and perform abnormal information detection on the target environment activity data to obtain abnormal information data;

[0059] The exception handling priority sorting module is used to match the exception information data with the preset emergency strategy database to obtain the emergency strategy data; to sort the exception handling priorities according to the emergency strategy data and the exception information data to obtain the exception handling priority data;

[0060] The emergency strategy correction module is used to obtain current resource data, and match resources according to the current resource data and exception handling priority data to obtain emergency strategy correction data;

[0061] The emergency strategy plan generation module is used to obtain the historical log data of the target object, and make corrections based on the historical log data of the target object and the emergency strategy correction data to obtain the emergency strategy plan data.

[0062] Optionally, the present application further provides an emergency plan generation device, the emergency plan generation device comprising:

[0063] and at least one processor;

[0064] A memory and a plurality of sensor modules in communication with the at least one processor, wherein the sensor modules include a target environment activity abnormal information detection module, an abnormality processing priority sorting module, an emergency strategy correction module, and an emergency strategy plan generation module;

[0065] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute an emergency plan generation method as described in any one of the above.

[0066] The purpose of the present invention is that the present invention can not only capture people, objects and their interactive behaviors, but also accurately extract activity data, which provides high-quality input data for anomaly detection and subsequent strategy formulation. Real-time image acquisition enables the system to respond to any potential abnormal events at any time, and activity extraction ensures the accuracy of anomaly detection. The present invention can quickly identify potential abnormal behaviors or events. The detection of abnormal information is based on multi-dimensional behavior analysis (such as human behavior, object location, situational characteristics, etc.), combined with historical data to form a comprehensive judgment. According to the matching of abnormal information with the preset emergency strategy database, the system can intelligently select the emergency strategy most relevant to the current situation. After obtaining the current resource data, the system matches resources according to the priority of the event and the availability of resources through the intelligent resource scheduling model. By obtaining the historical log data of the target object, the system can trace back and analyze similar events based on previous emergency response records and event processing results. The system can automatically extract similar historical events and review the effects of their emergency strategies, forming a self-learning mechanism, gradually accumulating experience and optimizing strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting implementations made with reference to the following drawings:

[0068] Figure 1 A flowchart showing a method for generating an emergency plan according to an embodiment of the present invention is provided;

[0069] Figure 2 A flowchart showing the steps of detecting abnormal information of target environment activities according to an embodiment is shown;

[0070] Figure 3 A flowchart of the steps of a method for prioritizing exception handling according to an embodiment is shown;

[0071] Figure 4 A flowchart showing a method for modifying an emergency strategy according to an embodiment of the present invention is shown;

[0072] Figure 5A flowchart showing a method for generating an emergency strategy plan according to an embodiment of the present invention is provided;

[0073] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0074] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0075] In addition, the drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.

[0076] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0077] See also Figures 1 to 5 , the present application provides a method for generating an emergency plan, the method comprising:

[0078] S1. Collect target environment images through the monitoring system to obtain target environment image data; extract target environment activities from the target environment image data to obtain target environment activity data; perform abnormal information detection on the target environment activity data to obtain abnormal information data;

[0079] Specifically, the monitoring system is used to capture image data in the target environment; image segmentation technology is used to separate different areas and objects in the target environment and extract the activity area. Then, object tracking is used to identify the activity behavior in the target environment and generate activity trajectory data, including the movement of objects, human behavior, and the operating status of equipment in the detection area.

[0080] Based on the historical data of normal target environment activities, by comparing the activity trajectory data, it detects activities that are inconsistent with normal behavior. Abnormal information detection uses thresholds to determine the activity frequency, range, or abnormal device status, and obtains abnormal activity data, such as illegal movement of objects, abnormal device opening, etc.

[0081] S2. Matching the abnormal information data with a preset emergency strategy database to obtain emergency strategy data; sorting the abnormal processing priorities according to the emergency strategy data and the abnormal information data to obtain abnormal processing priority data;

[0082] Specifically, the exception information data is matched with the preset emergency strategy database to find the emergency strategy most relevant to the current exception. This process includes gradually screening in the database based on features such as exception type, exception area, and exception severity to find the emergency plan with the highest matching degree. Based on multiple matched emergency strategies and current exception information data, the priority of each emergency strategy is evaluated. The basis for priority evaluation includes the severity of the exception, the frequency of the exception, the degree of potential risk, etc. According to this information, the emergency strategies are sorted and the exception handling priority data is generated.

[0083] S3, obtaining current resource data, and matching resources according to the current resource data and exception handling priority data to obtain emergency strategy correction data;

[0084] Specifically, currently available resource data is obtained, including human resources, equipment, material reserves, etc. Resource data can be collected through database query, real-time resource monitoring system, etc.

[0085] According to the exception handling priority data, currently available resources are matched to ensure that appropriate resources are allocated to each emergency strategy, including a comparison of resource requirements and resource status, including the quantity, availability, and allocation time of resources.

[0086] Dynamically adjust the emergency strategy based on the resource matching results. In the case of insufficient or excessive resources, modify the execution steps and emergency timetable of the emergency strategy, or change the priority of the strategy to generate revised emergency strategy data.

[0087] S4. Obtain historical log data of the target object, and make corrections based on the historical log data of the target object and the emergency strategy correction data to obtain emergency strategy plan data.

[0088] Specifically, historical log data of the target object is obtained, and these log data include historical abnormal data of the target environment, emergency strategy execution history, resource scheduling history, etc. These log data are extracted from the historical database through the data analysis system.

[0089] Further strategy optimization is performed based on the historical log data of the target object and the current revised emergency strategy. The emergency treatment effect, resource utilization efficiency, and abnormality resolution speed in the historical data are analyzed, and the existing emergency strategy is optimized and adjusted using this historical information to generate emergency strategy plan data. The plan not only considers the current abnormal situation, but also performs empirical optimization based on historical data.

[0090] In the present invention, the image data of the target environment is collected and processed by the monitoring system, so that real-time and efficient abnormal information detection can be achieved. By intelligently matching the abnormal information data with the emergency strategy database, the system can select the most appropriate emergency strategy for different abnormal situations, thereby improving the efficiency and pertinence of emergency response. Especially in complex or changeable environments, intelligent strategy matching can quickly form a processing plan to avoid repeated execution of inappropriate strategies. By matching and scheduling resources according to the priority of abnormal processing, it is ensured that important events are handled first and resources are optimally utilized.

[0091] Optionally, S1 includes:

[0092] S11, collecting target environment images through the monitoring system to obtain target environment image data;

[0093] Specifically, real-time image data collection is performed through the deployed monitoring system to ensure that the camera can cover the key areas of the target environment and capture sufficiently clear images. The collected image data is transmitted to the processing center in real time.

[0094] S12, performing multi-scale image feature extraction on the target environment image data to obtain target environment image feature data;

[0095] Specifically, multi-scale feature extraction is performed on the collected image data to extract image details and structural features at different scales. First, the image is processed at different resolutions and scales, so that information from small details (such as object edges) to global features (such as the overall environment outline) can be obtained. Next, feature point detection is performed on the image at each scale to generate local feature descriptors.

[0096] A set of feature point data is generated at different scales of the image. The feature point data includes the position, direction, scale information, etc. of the key points, and then the target environment image feature data is obtained by calculating the local feature descriptor.

[0097] S13, performing multi-layer semantic segmentation on the target environment image feature data to obtain target environment segmentation image data;

[0098] Specifically, a hierarchical image segmentation method is used to distinguish different regions in the image and extract meaningful semantic regions (such as people, equipment, desktop, etc.). The specific process is as follows: First, a preliminary region division is performed based on the low-level features of the image, such as dividing the foreground and background by features such as color and texture. Then, a higher-level semantic segmentation is further performed in each region to divide the image into multiple semantic categories, such as "person", "object", "table", etc.

[0099] Generate multi-level segmentation data, each layer represents the segmentation results of different semantic levels in the image. The low-level segmentation results provide details, while the high-level segmentation results provide overall semantic information, and output the target environment segmentation image data, representing the semantic category of each pixel in the image.

[0100] S14, performing activity behavior detection on the target environment segmentation image data to obtain target environment activity behavior data;

[0101] Specifically, based on the segmented image data, the activity behavior in the image is detected. First, the activity trajectory data is obtained by tracking the dynamic objects (such as people, objects, etc.) in the image. Then, the activity trajectory data is analyzed to extract the characteristics of the activity behavior, such as speed, moving direction, stay time, etc. Combined with the analysis of multiple frames of images, the behavior of people in the environment can be judged, such as walking, staying, object movement, etc.

[0102] By analyzing the image sequence, activity behavior trajectory data is generated, and through the trajectory and behavior analysis model, the target environment activity behavior data is obtained, which includes the time, location, category and other information of each activity behavior.

[0103] S15. Perform abnormal information detection on the target environment activity behavior data to obtain abnormal information data.

[0104] Specifically, abnormal information detection is performed on the detected activity behavior data to identify activities that do not conform to normal behavior patterns. For example, by comparing with the preset normal behavior pattern, it is analyzed whether certain activities have abnormal characteristics, such as abnormal movement paths, too high or too low activity frequencies, and people staying in sensitive areas for too long. The set threshold or behavior pattern model is used for judgment during detection.

[0105] Compare the activity behavior data with the normal behavior pattern, mark the abnormal activity behavior, and generate abnormal information data. The abnormal information data includes a specific description of the abnormal behavior, the time and location of occurrence, and the possible risk level.

[0106] In the present invention, real-time target environment image acquisition is performed through the monitoring system, which can ensure real-time monitoring of the on-site conditions in the environment of security-sensitive places, regardless of day or night. The system can capture abnormal activities occurring in the target environment, ensuring efficient security protection. Multi-scale feature extraction is performed on the target environment image data, which can capture the features of different detail levels in the image, ensuring that the system can identify targets or events of various scales. Multi-layer semantic segmentation can accurately classify different areas in the image and distinguish different scenes, objects and personnel. By segmenting the objects in the image, the system can accurately identify the activity status of each area in the target environment, which helps to better understand the actions and behaviors in the scene. By performing activity behavior detection on the target environment segmentation image data, the movement trajectory and behavior pattern of people or objects can be analyzed. This step can effectively identify potential security threats (such as abnormal actions of unauthorized personnel or abnormal movement of equipment), providing a basis for timely response. Abnormal information detection is performed on the activity behavior data, and activities that violate normal behavior patterns can be identified in a timely manner.

[0107] Optionally, the abnormal information detection includes:

[0108] Performing person recognition on the target environment activity behavior data to obtain person recognition data;

[0109] Specifically, character detection is performed on the target environment activity behavior data to identify all character objects in the image. By analyzing the character's appearance, facial features, posture information, etc., the corresponding character identity information and behavior information are generated. The character recognition process will compare with the preset character database to determine the identity of the person. The recognition process can also detect the specific location and movement trajectory of the person. Generate character recognition data, including the identity, location, and action information of the person, such as "Character A is standing in the southeast corner of the room, facing north."

[0110] Perform object recognition on the target environment activity behavior data to obtain object recognition data;

[0111] Specifically, object recognition is performed on the activity behavior data in the target environment. First, movable or fixed objects (such as equipment, files, tables and chairs, etc.) in the image are extracted through object detection, and then the category of the object is determined by comparing the known object information in the database. When identifying an object, the appearance, shape, color and other features of the object are analyzed at the same time to generate object recognition data. Object recognition data includes information such as object category, location, and status. For example, it is recognized that a device is in an active state or an object has been moved.

[0112] Perform anomaly detection based on the person recognition data and the object recognition data to obtain first anomaly detection data;

[0113] Specifically, a person A approaches a designated sensitive device B, which is only accessible to specific people according to the preset person recognition settings. Detect whether object C has been moved to an area where it should not be stored, or whether it has been operated by unauthorized personnel. The system compares the interaction pattern between the person and the object with the preset normal behavior pattern. For example, under normal circumstances, device B can only be operated by designated personnel, and object C should be stored in a specific area and should not be accessed by unauthorized personnel. When abnormal interaction behavior is detected, the system generates abnormal detection data. For example: "Person A approached device B without authorization." "Object C was moved to a non-designated area by an unauthorized person."

[0114] Performing behavior component detection according to the person recognition data to obtain person behavior component data;

[0115] Specifically, based on the character recognition data, the character's behavior components are further analyzed and detected. Behavior components refer to the specific actions of the character, such as walking, picking up objects, touching devices, etc. The start, end, duration, frequency and other features of each action are extracted to construct the character's behavior pattern. Then, based on the known behavioral norms, it is detected whether the behavior is abnormal, such as frequent contact with sensitive objects or staying in a certain area for a long time. Generate character behavior component data to record the specific behavior of each character, such as "Character A walked 5 times and touched device B."

[0116] Performing anomaly detection on the character behavior component data to obtain second anomaly detection data;

[0117] Specifically, the extracted character behavior component data is subjected to anomaly detection to determine whether these behavior components conform to the norm. The main analysis is whether the frequency, duration, and behavior object of the behavior exceed the preset threshold. For example, a character's behavior components such as long stays, abnormal repetitive actions, or contact with objects that should not be touched will be marked as abnormal. Generate second anomaly detection data, including a specific description of the abnormal behavior, such as "character A stays in a sensitive area for a time longer than the normal range" or "character B repeats the same action many times."

[0118] The first anomaly detection data and the second anomaly detection data are integrated to obtain anomaly information data.

[0119] Specifically, the first anomaly detection data (based on the anomaly of the interaction between the person and the object) and the second anomaly detection data (based on the anomaly of the person's behavior component) generated previously are integrated. By analyzing whether the person's interaction behavior and behavior component echo each other, a comprehensive judgment is made based on these data to improve the accuracy of anomaly detection.

[0120] The integrated abnormal information data contains a complete description of abnormal behavior, which can more comprehensively reflect the risk situation in the target environment. For example, "person A not only touched device B, but also stayed in a sensitive area for a long time, which is suspected to be illegal operation."

[0121] The present invention combines person recognition, object recognition and behavior component analysis, and comprehensively covers multiple dimensions of anomaly detection. By identifying people and objects, not only can active objects be detected, but also the behavior of these objects can be analyzed to see if they meet expectations, which significantly improves the accuracy of anomaly detection. Through person recognition and object recognition, the system can quickly capture the identities of people and objects in the monitoring scene and generate specific identification data. This step can effectively distinguish known people from strangers, or detect the presence of unauthorized objects, thereby triggering subsequent anomaly detection. By introducing person behavior component detection, the specific actions of the person are decomposed and analyzed, and subtle abnormal behavior changes can be detected. Through the detection of this behavior component, not only obvious abnormal behaviors can be discovered. The present invention integrates the first anomaly detection data (based on person and object recognition) with the second anomaly detection data (based on behavior component analysis) to analyze abnormal situations at multiple levels. This multi-level integration can reduce the false alarm rate and improve the reliability of anomaly detection. By combining person recognition data and object recognition data, the system can analyze the interactive behavior between people and objects. For example, when an unauthorized person is detected in contact with a sensitive object, the system can quickly identify and trigger an alarm.

[0122] Optionally, S2 includes:

[0123] S21. Evaluate the impact scope of the event based on the abnormal information data to obtain the impact scope data of the abnormal event;

[0124] Specifically, based on the abnormal information data, the impact scope of the event is evaluated. The evaluation process takes into account the geographical location, physical space, time dimension, and potential chain reactions of the abnormal event. First, by analyzing the location and event type in the abnormal information data, the direct impact scope on the surrounding area is determined. Then, combined with environmental data (such as personnel density, equipment distribution, network connection, etc.), an extended analysis is performed to evaluate other areas or devices affected by the abnormal event. Finally, these impact range data are combined to determine the overall impact area of ​​the abnormal event.

[0125] Generate data on the impact range of abnormal events, including the spatial scope of the event, the people or devices affected. For example, an abnormal behavior affects the operation of confidential equipment, or has a chain effect on multiple surrounding devices.

[0126] S22, extracting scenario features from the abnormal information data to obtain abnormal scenario feature data;

[0127] Specifically, scenario features are extracted from the abnormal information data. Scenario features refer to the specific environmental conditions and associated contextual information under which the abnormal event occurred. During the extraction process, basic information such as the time point, spatial location, and abnormality type when the abnormal event occurred are first identified. Then, the external conditions when the event occurred are analyzed, such as environmental status (light, temperature, humidity), personnel and equipment distribution, network status, etc. At the same time, historical data related to the event is extracted to determine whether the scenario has similarities or correlations with previous abnormal events.

[0128] After the scene feature extraction is completed, the abnormal scene feature data is generated, which contains information such as the time, space, equipment status, personnel activities, etc. when the event occurred. For example, "the abnormal event occurred at 2 pm, equipment A was running at high temperature, and there were 3 staff members around."

[0129] Specifically, the system extracts the basic scenario features of the abnormal event from the abnormal information data, including the time of occurrence, spatial location, and type of abnormal event. For example: "The abnormal event occurred at 2 pm, the location was the fence on the east side of the building, and the event type was fence climbing." The system further analyzes the external environmental conditions when the abnormal event occurred. For example, environmental parameters such as lighting conditions, temperature, and humidity: "The lighting conditions at that time were cloudy, the temperature was 25 degrees Celsius, and the humidity was 60%." At the same time, the distribution information of personnel when the abnormal event occurred is extracted. For example, the number and location of personnel inside and outside the event area are identified: "When the incident occurred, there were 2 staff members near the fence, and 3 security personnel patrolled within 500 meters." The system checks whether the scenario is related to similar abnormal events in history. For example, whether other fence climbing events have occurred at the same location or under similar conditions: "In the past week, there have been 2 fence climbing events at the same fence location, with similar scenario conditions." After the extraction is completed, the system generates abnormal scenario feature data, including detailed information such as the time, location, personnel distribution, and environmental conditions of the event. For example: "The abnormal incident occurred at 2 pm at the east fence, the temperature was 25 degrees Celsius, there were 2 staff members around, and 3 security guards were patrolling nearby."

[0130] S23, matching the abnormal event impact range data, the abnormal scenario feature data and the preset emergency strategy database to obtain emergency strategy data;

[0131] Specifically, based on the abnormal information data, the system first evaluates the severity of each abnormal event. For example, if an illegal intrusion is detected in a certain area, the system will analyze the sensitivity of the area (such as whether it is close to an important protection area), the potential harmfulness of the incident, etc. The scope of impact caused by the abnormal event is further analyzed. For example, if an intruder attempts to enter multiple protection areas or is approaching a crowded place, such events will be marked as events with a wide impact range and high danger. Based on the real-time data of abnormal events, the urgency of the event is evaluated. For example, whether an intrusion incident is developing rapidly or whether there is a potential risk of rapid spread;

[0132] The system combines the emergency strategy data to analyze the execution time of each strategy, the number of security personnel required, and the difficulty of on-site response. For example, for a serious intrusion incident, if the mobilized security personnel can arrive quickly and solve the problem, the priority of the incident will be increased. By evaluating the effectiveness of the emergency strategy, it is determined whether the strategy can effectively contain or solve the abnormal incident. For example, a strategy involves mobilizing multiple security personnel to block a specific area. If the strategy can quickly control the situation, the priority of the strategy will also be increased accordingly. The system generates exception handling priority data based on the risk level of the incident, the availability of security resources, and the urgency of incident handling. The data contains the priority order of each exception handling strategy. For example: "Prioritize the intrusion incident (location: east fence), mobilize 2 nearby security personnel, and arrive at the scene within 3 minutes. Secondly, handle the fence climbing incident (location: west gate), and block the relevant entrances and exits."

[0133] S24. Sort exception handling priorities according to the emergency strategy data and the exception information data to obtain exception handling priority data.

[0134] Specifically, based on the abnormal information data, the system first evaluates the severity of each abnormal event (index mapping). For example, if an illegal intrusion is detected in a certain area, the system will analyze the sensitivity of the area (such as whether it is close to an important protection area), the potential harmfulness of the incident, etc. The scope of impact caused by the abnormal event is further analyzed (based on the probability calculation of probabilistic events and threshold comparison). For example, if the intruder attempts to enter multiple protection areas or is approaching a crowded place, such events will be marked as events with a wide impact range and high danger. Based on the real-time data of abnormal events, the urgency of the event is evaluated. For example, whether an intrusion event is developing rapidly or whether there is a potential risk of rapid spread;

[0135] The system combines the emergency strategy data to analyze the execution time of each strategy, the number of security personnel required, and the difficulty of on-site response. For example, for a serious intrusion incident, if the mobilized security personnel can arrive quickly and solve the problem, the priority of the incident will be increased. By evaluating the effectiveness of the emergency strategy, it is determined whether the strategy can effectively contain or solve the abnormal incident. For example, a strategy involves mobilizing multiple security personnel to block a specific area. If the strategy can quickly control the situation, the priority of the strategy will also be increased accordingly. The system finally generates exception handling priority data based on the risk level of the incident, the availability of security resources, and the urgency of incident handling. The data contains the priority order of each exception handling strategy. For example: "Prioritize the intrusion incident (location: east fence), mobilize 2 nearby security personnel, and arrive at the scene within 3 minutes. Secondly, handle the fence climbing incident (location: west gate), and block the relevant entrances and exits."

[0136] In the present invention, by evaluating the scope of impact of the abnormal information data, the system can predict the expansion scope and potential impact of the event, determine which areas, personnel or resources are affected, and help identify the areas most likely to be threatened, thereby ensuring that the emergency response can give priority to covering key areas. By extracting the scenario features of the abnormal information data, the system can analyze the specific scenarios (such as time, location, environmental conditions, etc.) of the abnormal event, which helps the system to better understand the background of the event during the emergency response, thereby adjusting the strategy. For example, events occurring at night require different emergency strategies than those occurring during the day. By matching the comprehensive abnormal event impact range data, abnormal scenario feature data and the preset emergency strategy database, the system can recommend the most suitable emergency strategy for abnormal events of different types and situations, ensuring that the system executes the most appropriate emergency plan according to the impact range and scenario characteristics of the event, and avoiding unnecessary waste of resources. By sorting the priority of abnormal processing according to the emergency strategy data and the abnormal information data, the system can dynamically adjust the processing order according to the urgency and impact range of the event. High-priority events will be processed in a timely manner, while low-priority events will be processed later according to the resource situation, ensuring the optimal use of limited resources and ensuring that key events can be responded to first.

[0137] Optionally, exception handling priority ranking includes:

[0138] Extracting abnormal features from abnormal information data to obtain abnormal feature data;

[0139] Specifically, the system first extracts basic feature information related to the event from the abnormal information data. For example, the type, time of occurrence, and spatial location of the abnormal event are extracted: "The type of abnormal event is fence climbing, the time of occurrence is 3 pm, and the location is the east fence." The system analyzes the dynamic characteristics of the personnel and surrounding environment involved in the abnormal event. For example, the number of personnel with abnormal behavior and their specific location are identified: "When the incident occurred, there was an unauthorized person climbing over the fence, and there were 2 staff members near the fence." At the same time, the system extracts changes in environmental conditions, such as lighting conditions and weather changes: "When the incident occurred, the sky gradually darkened and the light intensity gradually decreased." In-depth analysis of the dynamic characteristics of abnormal events, such as the rate and trend of abnormal behavior. For example, the speed of personnel climbing over the fence and its duration: "The duration of the fence climbing action is 30 seconds, and the speed of personnel activities gradually increases during the climbing process." The system generates detailed abnormal feature data, including the main characteristics of the event, personnel activity information, environmental state changes, etc. For example: "The abnormal event is climbing over the east fence, the event occurred at 3 pm, and the unauthorized personnel climbing action lasted for 30 seconds, during which the lighting conditions gradually decreased."

[0140] Perform abnormal expansion trend processing according to abnormal characteristic data to obtain abnormal expansion trend data;

[0141] Specifically, the system first analyzes the development speed of abnormal events based on the extracted abnormal feature data. For example, the change in the movement speed or activity intensity of illegal intruders in the area: "After the unauthorized person climbs over the fence, the movement speed gradually increases, and the probability of entering the building increases." The system combines the current abnormal feature data with environmental information (such as personnel distribution and regional layout) to analyze the propagation path of abnormal events. For example, the action path taken by illegal intruders in the building: "After entering from the east fence, the intruder may move along the main corridor to the south of the building to avoid the monitoring equipment." Based on the propagation path of the abnormal event, the system further predicts the area to which the abnormal event extends or the personnel affected. For example: "The intruder approaches the safety exit in the next 5 minutes and threatens 3 surrounding staff members." The system generates abnormal expansion trend data, including the development direction of the abnormal event, the potential affected areas and personnel. For example: "The expansion trend of the abnormal event shows that the unauthorized person will approach the south side of the building in the next 3 minutes and threaten the nearby security personnel."

[0142] Priority sorting is performed according to the abnormal expansion trend data and the emergency strategy data to obtain first abnormal processing priority data;

[0143] Specifically, the system analyzes the speed of event expansion and potential impact range based on the abnormal expansion trend data. For example, "After climbing over the fence, the intruder is moving quickly to the south side of the building and is expected to reach the safety exit within 3 minutes." If the abnormal expansion speed is fast and the impact range is wide, the priority of the event will be increased;

[0144] The system compares the abnormal expansion trend with the preset emergency strategies to evaluate the applicability and processing efficiency of each emergency strategy. For example, in response to the current expansion trend of the intruder, the strategy of blocking the south exit of the building can effectively prevent the intruder from moving further, with high applicability and fast response speed;

[0145] According to the execution time and effect of the strategy, the system assigns a preliminary priority to each emergency strategy. Strategies that quickly and effectively contain abnormal expansion have a higher priority. For example, "Strategy 1 (mobilizing security personnel to block the south exit) can be executed within 2 minutes and can directly stop intruders, so it has the highest priority." The system generates the first abnormal processing priority data and records the priority order of each emergency strategy. The basis for priority sorting is the urgency of the abnormal expansion trend and the effectiveness of each emergency strategy. For example: "Emergency Strategy 1 (blocking the south exit) has the highest priority, Strategy 2 (mobilizing security personnel to patrol the southern area) has the second priority, and Strategy 3 (closing other entrances to the building) has the third priority.".

[0146] Perform utility priority sorting according to the emergency strategy data and the exception information data to obtain second exception handling priority data;

[0147] Specifically, the system evaluates the utility of each emergency strategy based on the emergency strategy data and abnormal information data. For example, whether a certain strategy can directly prevent the expansion of abnormal events or effectively mitigate the impact. For each strategy, the system evaluates its resource requirements and execution time, and gives priority to strategies with low resource consumption and quick response. For example, "Strategy 1 (mobilizing security personnel to block the south exit) requires 2 security personnel, a response time of 2 minutes, can effectively prevent the movement of intruders, has low resource consumption, and high utility." Generate utility priority data for each emergency strategy. Strategies with higher utility priorities are usually the solutions that can most effectively solve the problem in the current situation and consume the least resources. For example, "Strategy 1 has the highest utility priority because it has a short response time and low resource consumption, and can directly prevent the expansion of abnormal events."

[0148] A weighted priority calculation is performed on the emergency strategy data according to the first exception handling priority data and the second exception handling priority data to obtain the exception handling priority data.

[0149] Specifically, the system performs weighted calculation on the first exception handling priority data (based on the expansion trend) and the second exception handling priority data (based on utility). The weight setting is based on the current emergency situation and resource status of the system. If the expansion trend is urgent, the weight of the expansion trend priority is higher; if resources are scarce, the weight of the utility priority is higher. For example, in an emergency, the weight of the expansion trend is set to 70%, and the weight of the utility is set to 30%. The system generates comprehensive priority data through weighted calculation. Based on the weighted results, the system generates exception handling priority data and records the comprehensive priority ranking of each emergency strategy. For example, "Strategy 1 (mobilizing security personnel to block the south exit) has the highest comprehensive priority and should be executed first."

[0150] In the present invention, by extracting abnormal features from abnormal information data, the system can extract key information from the abnormal data, such as the type, severity, and scope of impact of the abnormality, so that the system can more accurately understand the nature of each abnormal event. By processing the abnormal expansion trend of the abnormal feature data, the system can predict the development trend of the abnormal event (such as the scope of diffusion, the speed of increasing severity, etc.). Priority sorting is performed according to the abnormal expansion trend data and the emergency strategy data to ensure that the most rapidly expanding abnormal events are given priority. Especially in the case of fast abnormal expansion speed or wide impact range, the response speed of the system can be effectively improved to prevent the situation from deteriorating. According to the utility priority sorting, the system can evaluate the utility of each emergency strategy (such as processing efficiency, resource consumption, success rate, etc.), so as to allocate appropriate resources for different abnormal events. By evaluating the utility of the emergency strategy, the system can ensure that resources are optimally utilized and reduce resource waste or unnecessary expenses. The weighted priority calculation realizes multi-dimensional priority calculation by combining the first abnormal processing priority data (based on the expansion trend) and the second abnormal processing priority data (based on the utility sorting).

[0151] Optionally, S3 includes:

[0152] S31, obtaining current resource data;

[0153] Specifically, currently available resource data is obtained from the resource management system or database, including human resources, equipment resources, material reserves, network bandwidth, energy supply, etc. Through real-time data collection and query, the current status of resources is updated to ensure accurate information. The acquisition of resource data also needs to consider the status, location and availability of resources. For example, the available working hours of human resources, whether the equipment is under maintenance, the inventory of materials, etc. Generate current resource data, including the type, quantity, current status and location of each resource. For example, "Equipment A is idle and can be called immediately; Equipment B is under maintenance and is temporarily unavailable."

[0154] S32, constructing a resource scheduling model according to current resource data to obtain a resource scheduling model;

[0155] Specifically, a resource scheduling model is constructed based on the current resource data obtained. The construction of the resource scheduling model needs to comprehensively consider factors such as resource availability, location, function, and call time. Different types of resources are classified, such as manpower, equipment, and materials. The scheduling relationship between resources is established in combination with the usage rules and call restrictions of each resource. The model also needs to consider the priority of resources, such as the priority call strategy for key equipment and personnel in an emergency. Generate a resource scheduling model, which defines the call order and scheduling rules for different resources. For example, "Device A is called first; when device A is not available, device C is called as an alternative."

[0156] S33, calculating resource load according to the resource scheduling model to obtain a resource scheduling load model;

[0157] Specifically, based on the resource scheduling model, perform resource load calculations to evaluate the load level of current resources in emergency situations. Load calculations need to take into account the frequency of resource use, call time, time required to complete a task, and the available amount of resources. For each resource, calculate its current workload, such as the operating time of the equipment, the rate at which materials are consumed, etc. Evaluate the resource requirements for emergency handling, and compare the requirements with existing resources to determine the load level of the resources. If the load is too large, the model will indicate a risk of resource shortage. Generate a resource scheduling load model to record the load status of each resource. For example, "The load of device A is 80%, and device B is unavailable, so the overall load is high and additional backup equipment needs to be called."

[0158] S34. Perform resource matching according to the resource scheduling load model and exception handling priority data to obtain emergency strategy correction data.

[0159] Specifically, according to the resource scheduling load model and exception handling priority data, resource matching is performed to ensure that each emergency strategy can get appropriate resource support. High-priority exception handling strategies are matched with current resources to ensure that key tasks get priority resource support. Whether resources are sufficient is determined based on the resource load model. If some resources are insufficient, the scheduling strategy will be replaced or optimized. According to the matching results, the emergency strategy is revised and the processing flow is adjusted to ensure reasonable resource allocation. Generate emergency strategy revision data, including revised emergency handling plans and resource allocation plans. For example, "Since the load of device A is too high, use backup device C to process the task, and adjust personnel scheduling to optimize resource allocation."

[0160] In the present invention, by acquiring the current resource data in real time, the system can fully grasp the status of the current available resources, including personnel, equipment, time and location, etc., so that the system can dynamically adjust the resource scheduling scheme, adapt to the complex and changeable emergency environment, and ensure rapid response when resources change. Based on the current resource data, the resource scheduling model is constructed, and the system can intelligently schedule appropriate resources according to the specific needs of the event (such as event type, scale and processing priority). Through the resource scheduling model, it can be ensured that the most suitable resources are optimally allocated, reduce resource waste, and improve the efficiency of overall emergency processing. Resource load calculation is a key step to ensure reasonable resource scheduling. By calculating the load of each resource (such as current task, available time, remaining capacity, etc.), the system can avoid resource overload or uneven distribution, ensure that all resources are balanced in the emergency processing, and neither cause resource waste nor avoid excessive use or shortage of resources. By combining the resource scheduling load model and exception processing priority data for resource matching, the system can prioritize the allocation of appropriate resources for the most urgent events.

[0161] Optionally, the resource scheduling model construction includes:

[0162] Classify resource functions according to current resource data to obtain current resource operation label data;

[0163] Specifically, based on the current resource data, classify the resources by their functions. Analyze the types of resources, such as manpower, equipment, and materials. Then, classify the resources according to their functions or application scenarios. For example, human resources are divided into technical personnel, logistics personnel, etc.; equipment resources can be divided into production equipment, maintenance equipment, etc. By marking the specific role of the resources, a job label corresponding to each resource is generated, and the label clarifies the specific role of each resource in emergency handling. Generate current resource job label data, and each resource has corresponding role label information. For example, "Equipment A label: maintenance equipment; Personnel B label: emergency technical support personnel."

[0164] Calculate resource scheduling time and analyze resource spatial distribution based on current resource data and current resource job tag data to obtain resource scheduling time data and resource spatial distribution data;

[0165] Specifically, the scheduling time of each resource is calculated based on the current resource data and job tag data. The scheduling time includes the response time of the resource from the current state to being scheduled to the emergency treatment. For example, the time for the computing device to move from the current location to the target location, or the response time for the computing personnel from receiving the instruction to entering the working state. The scheduling time of resources needs to be evaluated in combination with factors such as traffic, path, and equipment startup time. Based on the location information of the current resources, the spatial distribution of resources is analyzed, including the current physical location of each resource, the distance from the emergency location, and the degree of resource concentration or dispersion. The analysis results can help determine the convenience and response speed of resource scheduling. Resource scheduling time data: records the response time or scheduling time of each resource, for example, "it takes 5 minutes for device A to be scheduled from the current location to the emergency location." Resource spatial distribution data: records the spatial distribution information of resources, such as "device A is located in the east district, 500 meters away from the emergency location; device B is located in the west district, 200 meters away from the emergency location."

[0166] A resource scheduling model is constructed according to resource scheduling time data and resource spatial distribution data to obtain a resource scheduling model.

[0167] Specifically, the resource scheduling model is constructed based on the resource scheduling time data and resource spatial distribution data. The priority and path of resource scheduling are determined in combination with the scheduling time and spatial distribution. For example, priority is given to scheduling resources with the shortest response time and the closest distance. At the same time, the job tags of the resources are considered, and appropriate resources are allocated according to the specific needs of the emergency event. The scheduling model also needs to include a resource conflict detection mechanism to ensure that resources are not repeatedly scheduled or overloaded. Alternative plans for emergency situations should also be set in the model to ensure that when key resources are unavailable, there are backup resources to supplement them. Generate a resource scheduling model, which includes information such as resource scheduling order, path, and timetable. For example, "equipment A and personnel B are optimally scheduled, equipment A will arrive at the emergency location within 5 minutes, and personnel B will arrive within 3 minutes."

[0168] Specifically, according to the needs of the emergency event, combined with the job tags of the resources, select the resources that meet the needs. The job tags clarify the functions, uses and applicable scenarios of the resources. The emergency event requires a maintenance device, and both equipment A and equipment C belong to the job tags of maintenance equipment. The maintenance tag of equipment A describes that it is suitable for high-precision maintenance, while equipment C is suitable for general maintenance. Therefore, according to the needs of the emergency scenario, a more suitable device is selected.

[0169] According to the resource scheduling time data and spatial distribution data, resources are prioritized and paths are planned. Priority is given to resources with the shortest response time and the closest distance. The scheduling path of each resource is calculated to determine the optimal path and the backup plan for the suboptimal resources. After analyzing the time and distance, the response time of device A is longer, but the maintenance accuracy meets the requirements; device C responds faster but has slightly lower accuracy. Therefore, according to actual needs, device A can be given priority, but device C can be kept as a suboptimal alternative to ensure that device C can be quickly called when device A is unavailable.

[0170] A resource conflict detection mechanism is added to the model to ensure that the same resource will not be repeatedly called or overloaded in multiple emergency events. By checking the resource scheduling model, it is confirmed whether each resource has a task conflict during the scheduling process. Device A is in maintenance and cannot be called. Therefore, when a conflict is detected, the model automatically sets device C as the main scheduling device and updates the scheduling order.

[0171] Set up alternatives for key resources to deal with resource unavailability or scheduling failure. Ensure that when the optimal resource is not in place, the system can call alternative resources for emergency processing in a timely manner to avoid interruption of emergency response. If device A fails or is overloaded during the scheduling process, the model automatically calls device C as a backup device. At the same time, update the personnel scheduling and change personnel B from the main task scheduling to the auxiliary task scheduling.

[0172] According to the above steps, a resource scheduling model is generated. The model contains specific information such as resource scheduling order, scheduling path, timetable, alternative plans, etc. to ensure smooth and efficient resource scheduling. The scheduling plan generated by the model is "Equipment A will be dispatched to the emergency location and is expected to arrive within 5 minutes; Personnel B is expected to arrive at the scene to assist within 3 minutes. Equipment C is an alternative equipment and can be replaced within 2 minutes if equipment A cannot be used."

[0173] In the present invention, by classifying resource functions according to current resource data, the system can identify the specific purpose, function and applicable scenario of each resource, and assign a job label to each resource, so that the system can clearly understand which resources are applicable to which tasks during scheduling, thereby improving the pertinence and efficiency of scheduling and avoiding the misuse or improper allocation of resources. By calculating the resource scheduling time according to the current resource data and the job label data, the system can accurately predict the time required for each resource from scheduling to executing the task. By analyzing the spatial distribution of the current resources, the system can understand the geographical location of the resources and their spatial relationship with the location of the emergency event. Combined with the resource scheduling time data, the system can optimize the resource scheduling route and reduce the moving time of the resources from the current location to the task location. By integrating the resource scheduling time data and the spatial distribution data to construct the resource scheduling model, the system can generate a comprehensive scheduling plan based on time and space. The scheduling model can achieve the optimal configuration of resources, not only ensuring that the resources can arrive at the task location in time, but also reducing the redundant use and excessive mobilization of resources through reasonable allocation.

[0174] Optionally, S4 includes:

[0175] S41, obtaining historical log data of the target object;

[0176] Specifically, historical log data of the target object is obtained from the system log or database. These data include the status of the target object in different time periods, operation records, exception handling records, etc. Log data usually records the time, location, resources involved, and results of emergency handling when the emergency occurs. Filter out historical records related to the current abnormal event to ensure the timeliness and relevance of the data.

[0177] Generate historical log data for the target object, which includes detailed records of the target object in previous events, such as "On May 1, 2022, device A failed and the processing time was 10 minutes."

[0178] S42, extracting similar events based on the historical log data and abnormal information data of the target object to obtain similar historical event data;

[0179] Specifically, similar events are extracted based on the historical log data of the target object and the current abnormal information data. The extraction process first compares the characteristics of the current abnormal event (such as event type, impact range, resource call status, etc.) with the event characteristics recorded in the historical log to find similar historical events. Matching is performed through dimensions such as event type, time, processing strategy, and event consequences to extract the historical case that is closest to the current abnormal event.

[0180] Generate similar historical event data and record the historical event information that is closest to the current abnormal event. For example, "The failure of device A on May 1, 2022 is similar to the current abnormal event, with the same failure type and the same equipment involved."

[0181] S43, backtracking the effect of the emergency strategy based on similar historical event data to obtain historical strategy effect data;

[0182] Specifically, based on similar historical event data, the effects of historical emergency strategies are traced back. The traceback process first checks the emergency strategies adopted in similar historical events, and then analyzes the effects of these strategies, including processing time, problem-solving efficiency, resource consumption, etc. Evaluate whether the execution of the emergency strategy encounters obstacles and whether there is room for optimization. By comparing the effect data of historical emergency strategies, a reference can be provided for the optimization of the current strategy.

[0183] Generate historical strategy effect data to record the processing effects of emergency strategies in historical events, for example, "On May 1, 2022, the emergency strategy for device A failure was effective, but the processing time was long and resource utilization was low."

[0184] S44, performing strategy comparison based on historical strategy effect data and emergency strategy correction data to obtain strategy comparison data;

[0185] Specifically, compare the historical strategy effect data with the current emergency strategy revision data to analyze the differences between the two in terms of resource allocation, processing time, problem solving efficiency, etc. Compare the effects of emergency strategies item by item to find out the differences in the execution effects of historical strategies and current strategies. Evaluate whether the current strategy can avoid the problems encountered in the historical strategy or whether it can optimize the execution method of the historical strategy.

[0186] Generate policy comparison data to record the comparison results between the historical policy and the current revised policy. For example, "The current policy has improved resource utilization compared to the historical policy, but the processing time is still too long."

[0187] S45. Correct the emergency strategy correction data according to the strategy comparison data to obtain emergency strategy plan data.

[0188] Specifically, based on the policy comparison data, the current emergency policy revision data is further optimized. Based on the successful experiences or failed lessons in the historical policies, the execution steps, resource calling methods or processing priorities of the current policies are adjusted. The policy revision process needs to be combined with the actual situation of the current abnormal event to ensure that the revised policy can not only solve the problems in the historical policy, but also better deal with the current abnormal event.

[0189] Generate emergency strategy plan data, which includes optimized processing steps, resource allocation plan, emergency response time, etc. For example, "the revised emergency plan recommends adding additional resource calls to shorten the processing time and optimizes the scheduling priority of device A."

[0190] Specifically, the system first obtains the historical log data of the target object from the database, including abnormal events that occurred in the past, execution records of emergency response strategies, and event processing results. For example, the security system of a building has recorded several fence climbing events in the past and the corresponding security response measures. Based on the current abnormal information data, the system analyzes the abnormal events in the historical logs and extracts historical events similar to the current abnormal events. The similarity judgment criteria include event type, location, personnel behavior patterns, etc. For example: "The current fence climbing event is similar to three similar events in the past month. They occurred at the same location and the intruders had similar behavior patterns.";

[0191] Based on similar historical event data, the system traces back to the emergency strategies used at the time and evaluates the effectiveness of these strategies. The effectiveness evaluation criteria include event processing time, resource consumption, whether the event is effectively controlled, etc. For example: "The security personnel blockade strategy used last time effectively blocked the intruder within 10 minutes, and the resource consumption was 2 security personnel." The system compares the historical strategy effect data with the current emergency strategy to analyze the differences in resource consumption, execution time and effect. For example, "The historical strategy has a good effect, but the resource consumption is large. The current strategy has been adjusted in resource optimization.";

[0192] Based on the strategy comparison data, the system modifies the current emergency strategy to ensure the optimal resource allocation and response speed in the current environment. For example, "The revised strategy will reduce the number of security personnel and strengthen monitoring of specific areas." The system generates revised emergency strategy plan data, recording the adjusted emergency handling steps, required resources and time requirements. For example: "The revised emergency strategy is to mobilize one security personnel to block the south exit and another to strengthen patrols in the surrounding areas."

[0193] In the present invention, by obtaining the historical log data of the target object, the system can access past abnormal events and emergency handling records, which provides valuable reference information for current events. Analysis based on historical data can avoid blind decision-making and ensure that the emergency response strategy is optimized and adjusted based on actual experience. By extracting similar events from historical log data, the system can automatically find historical events similar to the current abnormal event. Through the backtracking of emergency strategy effects, the system can evaluate the emergency handling effects in historical events, including the efficiency, success rate, response time, etc. of strategy execution. The system can quantitatively analyze the execution of past strategies to ensure that the current emergency strategy is optimized based on actual results. By comparing the current emergency strategy with the strategies in similar historical events, the system can identify the advantages and disadvantages of the current strategy and correct the existing strategy through historical successful experience. Strategy comparison ensures the rationality of decision-making and improves the reliability of emergency response through continuous optimization.

[0194] Optionally, the present application further provides an emergency plan generation system, which is used to execute the emergency plan generation method as described above, and the emergency plan generation system includes:

[0195] The target environment activity abnormal information detection module is used to collect target environment images through the monitoring system to obtain target environment image data; extract target environment activities from the target environment image data to obtain target environment activity data; and perform abnormal information detection on the target environment activity data to obtain abnormal information data;

[0196] The exception handling priority sorting module is used to match the exception information data with the preset emergency strategy database to obtain the emergency strategy data; to sort the exception handling priorities according to the emergency strategy data and the exception information data to obtain the exception handling priority data;

[0197] The emergency strategy correction module is used to obtain current resource data, and match resources according to the current resource data and exception handling priority data to obtain emergency strategy correction data;

[0198] The emergency strategy plan generation module is used to obtain the historical log data of the target object, and make corrections based on the historical log data of the target object and the emergency strategy correction data to obtain the emergency strategy plan data.

[0199] Optionally, the present application further provides an emergency plan generation device, the emergency plan generation device comprising:

[0200] and at least one processor;

[0201] A memory and a plurality of sensor modules in communication with the at least one processor, wherein the sensor modules include a target environment activity abnormal information detection module, an abnormality processing priority sorting module, an emergency strategy correction module, and an emergency strategy plan generation module;

[0202] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute an emergency plan generation method as described in any one of the above.

[0203] The purpose of the present invention is that the present invention can not only capture people, objects and their interactive behaviors, but also accurately extract activity data, which provides high-quality input data for anomaly detection and subsequent strategy formulation. Real-time image acquisition enables the system to respond to any potential abnormal events at any time, and activity extraction ensures the accuracy of anomaly detection. The present invention can quickly identify potential abnormal behaviors or events. The detection of abnormal information is based on multi-dimensional behavior analysis (such as human behavior, object location, situational characteristics, etc.), combined with historical data to form a comprehensive judgment. According to the matching of abnormal information with the preset emergency strategy database, the system can intelligently select the emergency strategy most relevant to the current situation. After obtaining the current resource data, the system matches resources according to the priority of the event and the availability of resources through the intelligent resource scheduling model. By obtaining the historical log data of the target object, the system can trace back and analyze similar events based on previous emergency response records and event processing results. The system can automatically extract similar historical events and review the effects of their emergency strategies, forming a self-learning mechanism, gradually accumulating experience and optimizing strategies.

[0204] Therefore, from any point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the attached application documents rather than the above description, and it is intended that all changes falling within the meaning and scope of equivalent elements of the application documents are included in the present invention.

[0205] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for generating an emergency plan, characterized in that: The method comprises: S1. Collect target environment images through the monitoring system to obtain target environment image data; extract target environment activities from the target environment image data to obtain target environment activity data; perform abnormal information detection on the target environment activity data to obtain abnormal information data; S2. Matching the abnormal information data with a preset emergency strategy database to obtain emergency strategy data; sorting the abnormal processing priorities according to the emergency strategy data and the abnormal information data to obtain abnormal processing priority data; S3, obtaining current resource data, and matching resources according to the current resource data and exception handling priority data to obtain emergency strategy correction data; S4, obtaining historical log data of the target object, and revising the emergency strategy revision data according to the historical log data of the target object to obtain emergency strategy plan data; S1 includes: Collect target environment images through the monitoring system to obtain target environment image data; Perform multi-scale image feature extraction on the target environment image data to obtain target environment image feature data; Perform multi-layer semantic segmentation on the target environment image feature data to obtain target environment segmentation image data; Performing activity behavior detection on the target environment segmentation image data to obtain the target environment activity behavior data; Perform abnormal information detection on the target environment activity behavior data to obtain abnormal information data; The abnormal information detection includes: Performing person recognition on the target environment activity behavior data to obtain person recognition data; Perform object recognition on the target environment activity behavior data to obtain object recognition data; Perform anomaly detection based on the person recognition data and the object recognition data to obtain first anomaly detection data; Performing behavior component detection according to the person recognition data to obtain person behavior component data; Performing anomaly detection on the character behavior component data to obtain second anomaly detection data; Integrate the first anomaly detection data and the second anomaly detection data to obtain anomaly information data; S2 includes: Evaluate the impact scope of the event based on the abnormal information data to obtain the impact scope data of the abnormal event; Extracting scenario features from abnormal information data to obtain abnormal scenario feature data; The emergency strategy data is obtained by matching the abnormal event impact range data, the abnormal scenario feature data and the preset emergency strategy database; Exception handling priorities are sorted according to the emergency strategy data and the exception information data to obtain exception handling priority data.

2. The method according to claim 1, characterized in that Exception handling priorities include: Extracting abnormal features from abnormal information data to obtain abnormal feature data; Perform abnormal expansion trend processing according to abnormal characteristic data to obtain abnormal expansion trend data; Priority sorting is performed according to the abnormal expansion trend data and the emergency strategy data to obtain first abnormal processing priority data; Perform utility priority sorting according to the emergency strategy data and the exception information data to obtain second exception handling priority data; A weighted priority calculation is performed on the emergency strategy data according to the first exception handling priority data and the second exception handling priority data to obtain the exception handling priority data.

3. The method according to claim 1, characterized in that S3 include: Get current resource data; Construct a resource scheduling model based on current resource data to obtain a resource scheduling model; Calculate resource load according to the resource scheduling model to obtain the resource scheduling load model; Resources are matched according to the resource scheduling load model and exception handling priority data to obtain emergency strategy correction data.

4. The method according to claim 3, characterized in that Resource scheduling model construction includes: Classify resource functions according to current resource data to obtain current resource operation label data; Calculate resource scheduling time and analyze resource spatial distribution based on current resource data and current resource job tag data to obtain resource scheduling time data and resource spatial distribution data; A resource scheduling model is constructed according to resource scheduling time data and resource spatial distribution data to obtain a resource scheduling model.

5. The method according to claim 1, characterized in that S4 includes: Get the historical log data of the target object; Extract similar events based on the historical log data and abnormal information data of the target object to obtain similar historical event data; Based on similar historical event data, the emergency strategy effect is traced back to obtain historical strategy effect data; Perform strategy comparison based on historical strategy effect data and emergency strategy revision data to obtain strategy comparison data; The emergency strategy correction data is corrected according to the strategy comparison data to obtain the emergency strategy plan data.

6. An emergency plan generation system, characterized in that: Used to execute the emergency plan generation method according to claim 1, the emergency plan generation system comprises: The target environment activity abnormal information detection module is used to collect target environment images through the monitoring system to obtain target environment image data; extract target environment activities from the target environment image data to obtain target environment activity data; and perform abnormal information detection on the target environment activity data to obtain abnormal information data; The exception handling priority sorting module is used to match the exception information data with the preset emergency strategy database to obtain the emergency strategy data; to sort the exception handling priorities according to the emergency strategy data and the exception information data to obtain the exception handling priority data; The emergency strategy correction module is used to obtain current resource data, and match resources according to the current resource data and exception handling priority data to obtain emergency strategy correction data; The emergency strategy plan generation module is used to obtain the historical log data of the target object, and to amend the emergency strategy amendment data according to the historical log data of the target object to obtain the emergency strategy plan data.

7. An emergency plan generating device, characterized in that: The emergency plan generating device comprises: at least one processor; A memory and a plurality of sensor modules in communication with the at least one processor, wherein the sensor modules include a target environment activity abnormal information detection module, an abnormality processing priority sorting module, an emergency strategy correction module, and an emergency strategy plan generation module; The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute an emergency plan generation method as described in any one of claims 1-5.

Images