Outdoor monitoring automatic event classification system

Abstract

The invention discloses an outdoor monitoring automatic event classification system, which relates to the technical field of intelligent classification, and comprises a construction module, a division module and an output module, establishes a first mapping relation, obtains a grade corresponding to a classified event, and outputs a current time classification and a current grade to a monitoring center. According to the invention, a mapping relation is established between event classification preset by the building module and historical monitoring videos, current monitoring events are quickly identified and classified, the events are graded according to the grading logic, the monitoring videos can be analyzed in real time, abnormal behaviors can be automatically identified, and corresponding alarms are triggered according to event types and severity, so that the monitoring efficiency is improved. According to the method, rapid response of managers is facilitated, dependence on manual monitoring is reduced through automatic processing, event classification and grading logic can be continuously optimized based on AI and deep learning technologies, and the method adapts to different scenes and requirements.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent classification, and in particular to an outdoor surveillance automatic event classification system. Background Art

[0002] In recent years, deep learning technologies, especially convolutional neural networks (CNNs) and Transformer architectures, have become the core technologies for outdoor surveillance event classification. By combining various sensor data (such as cameras, millimeter-wave radars, etc.), the accuracy and reliability of event detection are improved, and abnormal behavior patterns are identified through clustering and analysis of target movement trajectories. Trajectory information analysis is not only used for event classification but also for behavior prediction and scene understanding.

[0003] Currently, in the Chinese invention patent with the publication number CN113313060A, a method for constructing an abnormal event classification model and a detection method based on video surveillance are disclosed. After clustering the normal event data set through the K-Means clustering algorithm, the normal event data located at the abnormal points are selected as the false abnormal event data set, and a sample training set is formed by combining other normal event data sets and abnormal event data sets; the SVM model is trained based on the above sample training set to obtain the trained SVM model. However, in the related technologies, no comprehensive mapping relationship is established by analyzing the relevant data of different people, things, and objects, which is not conducive to the comprehensiveness of abnormal event monitoring. There is no grading according to the different situations of abnormal events, which is not conducive to the clarity of early warning. At the same time, the dynamic development of event types is not considered. Summary of the Invention

[0004] The technical problem solved by the present invention is that in the related technologies, no comprehensive mapping relationship is established by analyzing the relevant data of different people, things, and objects, which is not conducive to the comprehensiveness of abnormal event monitoring. There is no grading according to the different situations of abnormal events, which is not conducive to the clarity of early warning. At the same time, the dynamic development of event types is not considered.

[0005] To solve the above technical problems, the present invention provides the following technical solution: An outdoor surveillance automatic event classification system, including a construction module, a division module, and an output module; The construction module establishes a first mapping relationship according to the preset event classification and historical surveillance videos; The division module divides the levels of the classified events in the historical surveillance videos according to the level division logic to obtain the levels corresponding to the classified events; The output module outputs the current time classification and the current level to the monitoring center.

[0006] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: event classification includes illegal intrusion events and abnormal behavior events, the illegal intrusion events include illegal entry of personnel, illegal entry of vehicles and illegal entry of wild animals, and the abnormal behavior events include fighting and falling; The construction module is connected to the security monitoring system, retrieves historical monitoring videos from the security monitoring system, and performs first preprocessing, second preprocessing and third preprocessing on each historical monitoring video.

[0007] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the first preprocessing includes: Select any historical monitoring video, set the first time interval as the segmentation step length, and segment the historical monitoring video into N historical sub-videos according to the segmentation step length; Perform second preprocessing on the sub-videos in the time order of the sub-videos; The second preprocessing includes: Obtain the brightness of the sub-video, set the first value as the brightness threshold, compare the brightness of the sub-video with the first value, when the brightness of the sub-video is less than the first value, perform enhancement processing on the sub-video, when the brightness of the sub-video is greater than or equal to the first value, jump to the next sub-video, and repeat the brightness comparison process until all sub-videos are processed, and then stop the second preprocessing; The third preprocessing includes: Perform frame windowing processing on the sub-videos after the second preprocessing to obtain each sub-image of the sub-video, delete the sub-images that do not have face shape features or vehicle shape features or wild animal shape features, and set the remaining sub-images after deletion as the first images corresponding to the sub-videos.

[0008] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the face shape features, vehicle shape features and wild animal shape features are obtained by acquiring face images, vehicle images and wild animal images through an image database, and then extracting the first shape feature quantities of each image through machine vision. Extract the second shape feature quantity in the sub-image through machine vision, calculate the similarity between the first shape feature quantity and the second shape feature quantity through the cosine similarity formula, set the second value as the similarity threshold, compare the similarity with the second value, when the similarity is greater than or equal to the second value, it is determined that the sub-image has face shape features or vehicle shape features or wild animal shape features, and when the similarity is less than the second value, it is determined that the sub-image does not have face shape features or vehicle shape features or wild animal shape features.

[0009] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the construction logic of the first mapping relationship includes: Preset event classification, stipulate the rules of event classification, and establish the first mapping relationship between the stipulated rules and event classification; Obtain event classification by inputting rules into the first mapping relationship.

[0010] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the rules of the event classification include: Obtain the first image and set up an electronic fence; The rule for illegal intrusion of wild animals is: there are wild animals, and the wild animals are distributed on the boundary or inside the electronic fence; The rule for illegal intrusion of personnel is: there is a human face, the human face is distributed on the boundary of the electronic fence, and the posture of the person corresponding to the human face is a climbing posture or a striding posture; The rule for illegal entry of a vehicle is: there is a vehicle, the vehicle is distributed inside the electronic fence, there is no vehicle entry and exit record, or the license plate number recorded is different from the license plate number in the first picture; The rule for fighting is: there is a human face, the human face is distributed inside or on the boundary of the electronic fence, and the posture of the person is a posture of holding an instrument or a posture of physical collision; The rule for falling is: there is a human face, and the posture of the person corresponding to the human face is a lying posture; When there is a vehicle in the first image and the vehicle is distributed inside the electronic fence, identify the license plate number of the vehicle, obtain the monitoring time point at this time, retrieve the security record database, retrieve the vehicle entry and exit records of the corresponding event in the security record database, and when there is no vehicle entry and exit record or the license plate number recorded is different from the license plate number in the first picture, set the event as illegal entry of the vehicle; When there are two or more human faces in the first image, identify the postures of the persons corresponding to each human face through the first image in chronological order. When the posture is a physical collision, set the event classification as fighting.

[0011] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the recognition logic of the posture includes: By retrieving the posture database, retrieving the falling posture image, lying posture image, limb collision posture image, climbing posture image, and hand-held instrument image from the posture database, respectively extracting the third shape feature quantity of the falling posture image, lying posture image, limb collision posture image, climbing posture image, and hand-held instrument image, extracting the fourth shape feature quantity of the person corresponding to the face in the first image, calculating the similarity between the third shape feature quantity and the fourth shape feature quantity through the cosine similarity formula, comparing each similarity with the second value, when the similarity is greater than or equal to the second value, setting the corresponding posture image as the posture image of the person in the first image, when the similarity is less than the second value, jumping to the next posture image, and looping to compare the similarity, when all the posture images are jumped and all the similarities are less than or equal to the second value, determining the posture in the first image as a normal posture and not performing subsequent event classification.

[0012] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the level division logic includes: After event classification, select the first first image of any sub-video, count the first quantity of the face or wild animal or vehicle in the corresponding event classification, traverse the first quantity corresponding to the first image in the sub-video, calculate the growth rate of adjacent first vectors, and calculate the first average value of the growth rate of the sub-video; Set the third value and the fourth value as the growth rate thresholds, wherein the third value is less than the fourth value, compare the first average value with the growth rate thresholds, when the first average value is less than or equal to the third value, set the level as the first level, when the first average value is greater than the third value and less than or equal to the fourth value, set the level as the second level, when the first average value is greater than the fourth value, set the level as the third level, and the first level, the second level, and the third level represent the increasing order of the severity of the event classification.

[0013] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the calculation logic of the growth rate includes: In chronological order, subtract the previous first quantity from the next first quantity to obtain a second difference, calculate the first ratio of the second difference to the previous first quantity, and set the first ratio as the growth rate of the next first quantity relative to the previous first quantity.

[0014] As a preferred solution of an outdoor monitoring automatic event classification system according to the present invention, wherein: the output module obtains the current event classification of the current monitoring video according to the first mapping relationship, obtains the current level corresponding to the current event classification according to the level division logic, and outputs the current time classification and the current level to the monitoring center.

[0015] Advantages of the present invention: By establishing a mapping relationship between the event classifications preset in the construction module and historical surveillance videos, the system can quickly identify and classify current surveillance events, grade the events according to the grading logic, ensure that the monitoring center can prioritize the handling of high-priority events, can analyze the surveillance videos in real time, automatically identify abnormal behaviors (such as fighting, falling, abnormal parking, etc.), and trigger corresponding alarms according to the event type and severity, and transmit the classified events and grade information to the monitoring center in real time, facilitating rapid response by management personnel. The automated processing reduces the dependence on manual monitoring, avoids the fatigue and negligence of manual monitoring. Based on AI and deep learning technologies, the system can continuously optimize the event classification and grading logic to adapt to different scenarios and requirements. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 FIG. is a schematic diagram of the basic process of an outdoor surveillance automatic event classification system provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific embodiments of the present invention is made in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention.

[0018] Embodiment, referring to Figure 1 , which is an embodiment of the present invention, provides an outdoor surveillance automatic event classification system, including a construction module, a division module, and an output module; The construction module establishes a first mapping relationship according to the preset event classifications and historical surveillance videos; The division module grades the grades of the classified events in the historical surveillance videos according to the grading logic to obtain the grades corresponding to the classified events; The output module outputs the current classification and the current grade to the monitoring center.

[0019] By establishing a mapping relationship between the event classifications preset in the construction module and historical surveillance videos, the system of the present invention can quickly identify and classify current surveillance events, grade the events according to the grading logic, ensure that the monitoring center can prioritize the handling of high-priority events, can analyze the surveillance videos in real time, automatically identify abnormal behaviors (such as fighting, falling, abnormal parking, etc.), and trigger corresponding alarms according to the event type and severity, and transmit the classified events and grade information to the monitoring center in real time, facilitating rapid response by management personnel. The automated processing reduces the dependence on manual monitoring, avoids the fatigue and negligence of manual monitoring. Based on AI and deep learning technologies, the system can continuously optimize the event classification and grading logic to adapt to different scenarios and requirements.

[0020] The event classification includes illegal intrusion events and abnormal behavior events. The illegal intrusion events include illegal entry of personnel, illegal entry of vehicles, and illegal entry of wild animals. The abnormal behavior events include fighting and falling. The construction module is connected to the security monitoring system, retrieves historical monitoring videos from the security monitoring system, and performs first preprocessing, second preprocessing, and third preprocessing on each historical monitoring video.

[0021] The first preprocessing includes: Select any historical monitoring video, set the first time interval as the segmentation step size, and segment the historical monitoring video into N historical sub-videos according to the segmentation step size. Perform second preprocessing on the sub-videos in the time order of the sub-videos. The second preprocessing includes: Obtain the brightness of the sub-video, set the first value as the brightness threshold. When the brightness of the sub-video is compared with the first value, if the brightness of the sub-video is less than the first value, perform enhancement processing on the sub-video. When the brightness of the sub-video is greater than or equal to the first value, jump to the next sub-video and repeat the brightness comparison process until all sub-videos are processed, then stop the second preprocessing. The third preprocessing includes: Perform frame windowing processing on the sub-videos after the second preprocessing to obtain each sub-image of the sub-video. Delete the sub-images that do not have human face shape features, vehicle shape features, or wild animal shape features, and set the remaining sub-images after deletion as the first image corresponding to the sub-video.

[0022] In specific implementation, splitting a long video into multiple shorter sub-videos can reduce the amount of data processed at one time and improve the processing speed. The division of sub-videos makes subsequent brightness processing and feature extraction more flexible, and can perform targeted processing according to the characteristics of different sub-videos. The video content in different time periods may vary greatly, and after splitting, it can better adapt to scene changes. Performing enhancement processing on sub-videos with insufficient brightness can improve the image quality, reduce the loss of details caused by insufficient illumination, and the enhanced image is more conducive to subsequent feature detection (such as human faces, vehicles, etc.). Skipping sub-videos with normal brightness directly can avoid unnecessary processing and save computing resources. Deleting irrelevant sub-images can reduce the amount of data processed subsequently and improve the system efficiency.

[0023] The human face shape feature, vehicle shape feature, and wild animal shape feature are obtained by acquiring human face images, vehicle images, and wild animal images from an image database, and then extracting the first shape feature quantity of each image through machine vision. The second shape feature quantity in the sub-image is extracted through machine vision, and the similarity between the first shape feature quantity and the second shape feature quantity is calculated using the cosine similarity formula. The second value is set as the similarity threshold, and the similarity is compared with the second value. When the similarity is greater than or equal to the second value, it is determined that there is a human face shape feature, vehicle shape feature, or wild animal shape feature in the sub-image. When the similarity is less than the second value, it is determined that there is no human face shape feature, vehicle shape feature, or wild animal shape feature in the sub-image.

[0024] In specific implementation, the shape feature quantities of human faces, vehicles, and wild animals extracted from the image database through machine vision technology can more accurately describe the shape information of the target, adapt to different types of images, including human faces, vehicles, and wild animals, etc., and have good versatility. By adjusting the similarity threshold, it can adapt to the recognition requirements in different scenarios and improve the adaptability of the system.

[0025] The construction logic of the first mapping relationship includes: Preset event classification, stipulate the rules of event classification, and establish the first mapping relationship between the stipulated rules and event classification; By inputting rules into the first mapping relationship, event classification is obtained.

[0026] In specific implementation, through preset event classification, the event types are clarified, avoiding ambiguity and uncertainty, enabling the system to quickly identify and respond to different types of events. Through preset rules and mapping relationships, the system can quickly classify monitoring events, reducing the time and complexity of event processing. Through quick classification, the monitoring center can promptly obtain event type and priority information, thereby quickly responding to and processing events. The classified event data can serve as the basis for intelligent analysis, such as behavior pattern analysis, event trend prediction, etc.

[0027] The rules of the event classification include: Obtain the first image and set up an electronic fence; The rule for illegal intrusion of wild animals is: there are wild creatures, wild animals are distributed on the boundary or inside the electronic fence; The rule for illegal intrusion of personnel is: there is a human face, the human face is distributed on the boundary of the electronic fence, and the posture of the person corresponding to the human face is a climbing posture or a straddling posture; The rule for illegal entry of a vehicle is: there is a vehicle, the vehicle is distributed inside the electronic fence, there is no vehicle entry and exit record, or the license plate number recorded is different from the license plate number in the first picture; The rules for fighting are as follows: there is a human face, the human face is distributed inside the electronic fence or distributed inside the electronic fence, and the human posture is a posture of holding an instrument or a posture of physical collision; The rules for falling are as follows: there is a human face, and the posture of the person corresponding to the human face is a lying posture; When there is a vehicle in the first image and the vehicle is distributed inside the electronic fence, identify the license plate number of the vehicle, obtain the monitoring time point at this time, retrieve the security record database, and retrieve the vehicle entry and exit records of the corresponding event in the security record database. When there is no vehicle entry and exit record or the recorded license plate number is different from the license plate number in the first picture, set the event as illegal vehicle entry; When there are two or more human faces in the first image, identify the postures of the persons corresponding to each human face through the first image in chronological order. When the posture is a physical collision, classify the event as a fight.

[0028] In specific implementation, by setting multiple conditions (such as posture, position, license plate number comparison, etc.) to judge the event type, the misjudgment that may be caused by a single condition is reduced, the efficiency of outdoor monitoring is improved, and the comprehensiveness of monitoring is facilitated.

[0029] The recognition logic of the posture includes: By retrieving the posture database, retrieve the falling posture image, lying posture image, physical collision posture image, climbing posture image and instrument-holding image from the posture database, respectively extract the third shape feature quantity of the falling posture image, lying posture image, physical collision posture image, climbing posture image and instrument-holding image, extract the fourth shape feature quantity of the person corresponding to the human face in the first image, calculate the similarity between the third shape feature quantity and the fourth shape feature quantity through the cosine similarity formula, compare each similarity with the second value. When the similarity is greater than or equal to the second value, set the corresponding posture image as the posture image of the person in the first image. When the similarity is less than the second value, jump to the next posture image and loop to compare the similarity. When all the posture images are jumped and all the similarities are less than or equal to the second value, judge the posture in the first image as a normal posture and do not perform subsequent event classification.

[0030] The level division logic includes: After the event is classified, select the first first image of any sub-video, count the first quantity of the human face or wild animal or vehicle in the corresponding event classification, traverse the first quantity corresponding to the first image in the sub-video, calculate the growth rate of adjacent first vectors, and calculate the first average value of the growth rate of the sub-video; Set the third value and the fourth value as the growth rate thresholds, where the third value is less than the fourth value. Compare the first average value with the growth rate thresholds. When the first average value is less than or equal to the third value, set the level as the first level. When the first average value is greater than the third value and less than or equal to the fourth value, set the level as the second level. When the first average value is greater than the fourth value, set the level as the third level. The severity of the event classification represented by the first level, the second level, and the third level is in ascending order.

[0031] The calculation logic of the growth rate includes: In chronological order, subtract the previous first quantity from the subsequent first quantity to obtain a second difference. Calculate the first ratio of the second difference to the previous first quantity, and set the first ratio as the growth rate of the subsequent first quantity relative to the previous first quantity.

[0032] In specific implementation, counting the first quantity of human faces, wild animals, or vehicles in the event classification provides basic data for evaluating the severity of the event. Calculating the growth rate of adjacent first vectors and calculating the first average value of the growth rate of the sub-video can reflect the development speed and trend of the event. Through level classification, the monitoring center can give priority to handling events with higher severity, improving the response efficiency.

[0033] The output module obtains the current event classification of the current monitoring video according to the first mapping relationship, obtains the current level corresponding to the current event classification according to the level classification logic, and outputs the current time classification and the current level to the monitoring center.

[0034] In specific implementation, the output module obtains the event classification of the current monitoring video according to the first mapping relationship, clarifies the type of the event (such as illegal intrusion, abnormal behavior, etc.), enabling the monitoring center to quickly understand the nature of the event. According to the level classification logic, the output module classifies the severity of the event into different levels (such as the first level, the second level, the third level), providing more detailed event information for the monitoring center, being able to output event information in real time, ensuring that the monitoring center always grasps the latest monitoring status and takes measures in a timely manner.

[0035] The present invention establishes a mapping relationship between the event classification preset by the module and the historical surveillance videos, enabling the system to quickly identify and classify current surveillance events, grade the events according to the grading logic, ensure that the surveillance center can give priority to handling high-priority events, analyze the surveillance videos in real time, automatically identify abnormal behaviors (such as fighting, falling, abnormal parking, etc.), trigger corresponding alarms according to the event type and severity, and transmit the classified events and grade information to the surveillance center in real time, facilitating the rapid response of management personnel. The automated processing reduces the dependence on manual surveillance, avoids the fatigue and negligence of manual surveillance. Based on AI and deep learning technologies, the system can continuously optimize the event classification and grading logic to adapt to different scenarios and requirements.

[0036] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media containing computer-usable program code. Among them, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (Static Random Access Memory, abbreviated as SRAM), electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, abbreviated as EEPROM), erasable programmable read-only memory (Erasable Programmable Read Only Memory, abbreviated as EPROM), programmable read-only memory (Programmable Red-Only Memory, abbreviated as PROM), read-only memory (Read-Only Memory, abbreviated as ROM), magnetic memory, flash memory, a magnetic disk, or an optical disc. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured article including an instruction device, and the instruction device implements the functions specified in one process Figure 1 one process or multiple processes and / or boxes Figure 1 specified in the boxes or multiple boxes.

[0037] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. An outdoor monitoring automatic event classification system, characterized in that: Includes building module, partitioning module and output module; The construction module establishes a first mapping relationship according to a preset event classification and historical surveillance video; The classification module classifies the levels of the classified events in the historical surveillance video according to the level classification logic to obtain the levels corresponding to the classified events; The output module outputs the current time classification and the current level to the monitoring center.

2. The outdoor monitoring automatic event classification system according to claim 1, characterized in that: Event classification includes illegal intrusion events and abnormal behavior events. The illegal intrusion events include illegal intrusion by people, illegal entry by vehicles and illegal intrusion by wild animals. The abnormal behavior events include fighting and falling. The construction module is connected to the security monitoring system, retrieves historical monitoring videos from the security monitoring system, and performs first preprocessing, second preprocessing and third preprocessing on each historical monitoring video.

3. An outdoor monitoring automatic event classification system as claimed in claim 2, characterized in that: The first preprocessing comprises: Select any historical surveillance video, set the first time interval as a segmentation step, and segment the historical surveillance video into N historical sub-videos according to the segmentation step; Performing a second preprocessing on the sub-videos according to the time sequence of the sub-videos; The second preprocessing comprises: Acquire the brightness of the sub-video, set the first value as the brightness threshold, compare the brightness of the sub-video with the first value, and when the brightness of the sub-video is less than the first value, enhance the sub-video, and when the brightness of the sub-video is greater than or equal to the first value, jump to the next sub-video, and repeat the brightness comparison process until all the sub-videos are processed, and then stop the second preprocessing; The third preprocessing comprises: The second preprocessed sub-video is frame-divided and windowed to obtain sub-images of the sub-video, and the sub-images without facial shape features, vehicle shape features, or wild animal shape features are deleted, and the remaining sub-images after the deletion are set as the first images corresponding to the sub-video.

4. The outdoor monitoring automatic event classification system according to claim 3, characterized in that: The facial shape features, vehicle shape features and wild animal shape features are obtained by acquiring facial images, vehicle images and wild animal images from an image database, extracting the first shape feature quantity of each image through machine vision, extracting the second shape feature quantity in the sub-image through machine vision, and calculating the similarity between the first shape feature quantity and the second shape feature quantity through a cosine similarity formula, setting the second value as a similarity threshold, comparing the similarity with the second value, and when the similarity is greater than or equal to the second value, judging that the facial shape features or the vehicle shape features or the wild animal shape features exist in the sub-image, and when the similarity is less than the second value, judging that the facial shape features or the vehicle shape features or the wild animal shape features do not exist in the sub-image.

5. The outdoor monitoring automatic event classification system according to claim 3, characterized in that: The construction logic of the first mapping relationship includes: Preset event classification, define event classification rules, and establish a first mapping relationship between the defined rules and event classification; By inputting rules into the first mapping relationship, event classification is obtained.

6. An outdoor monitoring automatic event classification system as claimed in claim 5, characterized in that: The event classification rules include: Acquire the first image and set up an electronic fence; The rules for illegal wildlife intrusion are: the presence of wildlife, wildlife distribution on the boundary of the electronic fence or inside the electronic fence; The rules for illegal intrusion are: the presence of a face, the face is distributed on the boundary of the electronic fence, and the posture of the person corresponding to the face is climbing or crossing; The rules for illegal vehicle entry are: the vehicle exists, the vehicle is located inside the electronic fence, there is no vehicle entry or exit record, or the recorded license plate number is different from the license plate number in the first picture; The rules for fighting are: the presence of a human face, the human face is distributed inside the electronic fence or distributed inside the electronic fence, and the person is holding a weapon or in a physical collision posture; The rule for falling is: there is a face, and the person's posture corresponding to the face is lying down; When there is a vehicle in the first image and the vehicle is distributed inside the electronic fence, the license plate number of the vehicle is identified, the monitoring time point at that time is obtained, the security record database is retrieved, and the vehicle entry and exit records of the corresponding event in the security record database are retrieved; when there is no vehicle entry and exit record or the recorded license plate number is different from the license plate number in the first image, the event is set as illegal vehicle entry; When there are two or more faces in the first image, the postures of the persons corresponding to the faces are identified through the first images in time sequence, and when the postures are physical collisions, the event classification is set as a fight.

7. An outdoor monitoring automatic event classification system as claimed in claim 6, characterized in that: The gesture recognition logic includes: By retrieving a posture database, falling posture images, lying posture images, limb collision posture images, climbing posture images and hand-held instrument images are retrieved from the posture database, and the third shape feature quantities of the falling posture images, lying posture images, limb collision posture images, climbing posture images and hand-held instrument images are respectively extracted, and the fourth shape feature quantity of the person corresponding to the face in the first image is extracted, and the similarity of the third shape feature quantity and the fourth shape feature quantity is calculated by the cosine similarity formula, and each similarity is compared with the second value. When the similarity is greater than or equal to the second value, the corresponding posture image is set to the posture image of the person in the first image. When the similarity is less than the second value, jump to the next posture image and cyclically compare the similarities. When all posture images are jumped and the similarities are all less than or equal to the second value, the posture in the first image is judged as a normal posture, and subsequent event classification is not performed.

8. The outdoor monitoring automatic event classification system according to claim 1, characterized in that: The grading logic includes: After the event is classified, the first first image of any sub-video is selected, the first number of human faces, wild animals or vehicles in the corresponding event classification is counted, the first number corresponding to the first image in the sub-video is traversed, the growth rate of adjacent first vectors is calculated, and the first average value of the growth rate of the sub-video is calculated; The third value and the fourth value are set as growth rate thresholds, wherein the third value is smaller than the fourth value, and the first average value is compared with the growth rate threshold; when the first average value is less than or equal to the third value, the level is set to the first level; when the first average value is greater than the third value and less than or equal to the fourth value, the level is set to the second level; when the first average value is greater than the fourth value, the level is set to the third level; the first level, the second level and the third level represent the severity of the event classification in ascending order.

9. An outdoor monitoring automatic event classification system as claimed in claim 8, characterized in that: The calculation logic of the growth rate includes: In chronological order, a second difference is obtained by subtracting a later first quantity from a previous first quantity, a first ratio of the second difference to the previous first quantity is calculated, and the first ratio is set as a growth rate of the later first quantity relative to the previous first quantity.

10. The outdoor monitoring automatic event classification system according to claim 1, characterized in that: The output module obtains the current event classification of the current monitoring video according to the first mapping relationship, obtains the current level corresponding to the current event classification according to the level classification logic, and outputs the current time classification and the current level to the monitoring center.

Images