A method for recognizing group behaviors based on video data

Abstract

The present invention provides a method for identifying group behaviors based on video data, which relates to the technical field of group behavior identification. The method includes: obtaining a set of video images; extracting the behavior features of targets from the video images, where the targets include biological objects and non-biological objects; determining the behavior features of the group according to the behavior features of the biological objects; determining the influence weight of the biological objects on the group behavior through the behavior features of the biological objects and the group to obtain a first weight; using a tracking algorithm to track the targets to obtain the behavior trajectories of the targets; determining the influence weight of the non-biological objects on the group behavior through the behavior trajectories of the targets to obtain a second weight; and obtaining group behavior recognition data through the behavior features of the targets, the first weight, and the second weight. It not only considers biological objects but also non-biological objects, improving the comprehensiveness of group behavior recognition and thus improving the recognition accuracy of group behaviors.

Description

Technical Field

[0001] The present invention relates to the technical field of group behavior recognition, and particularly to a method for recognizing group behavior based on video data. Background Art

[0002] Group behavior recognition has been widely applied in social scenarios such as sports game video analysis, surveillance video recognition, and social behavior understanding. By recognizing group behaviors in videos, the occurrence of dangerous events can be prevented.

[0003] Currently, most efforts are dedicated to constructing the interaction relationships between individuals in a multi-person scenario to infer group behaviors. However, in actual scenarios, not only individual relationships affect group behaviors, but external interference events also have an impact. Due to the complexity of group relationships, the influence degrees of each individual and external interference events on group behaviors are different. If only individual characteristics are used to infer group behaviors, the recognition accuracy will be affected. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for recognizing group behavior based on video data, and the technical problem to be solved is how to improve the recognition accuracy of group behavior.

[0005] The present invention is achieved through the following technical solutions:

[0006] A method for recognizing group behavior based on video data, comprising the following steps:

[0007] S100. Obtain a group behavior video, sample the above group behavior video to obtain a video image set composed of a plurality of video images;

[0008] S200. Extract the behavior features of the targets from the above video images, where the targets include biological objects and non-biological objects; among them, a group is composed of a plurality of the above biological objects, and the group behavior features are determined according to the behavior features of the biological objects;

[0009] Determine the influence weight of the biological objects on the group behavior through the behavior features of the above biological objects and the group, and obtain the first weight;

[0010] S300. Use a tracking algorithm to track the above targets to obtain the behavior trajectories of the targets;

[0011] Determine the influence weight of the non-biological objects on the group behavior through the behavior trajectories of the above targets, and obtain the second weight;

[0012] S400. Obtain group behavior recognition data through the behavior features of the above targets, the first weight, and the second weight.

[0013] By evaluating the influence of biological objects and non-biological objects on group behavior, not only the behavioral characteristics of biological objects are considered, but also the influence of non-biological objects on group behavior is taken into account, which improves the comprehensiveness of group behavior recognition and further enhances the recognition accuracy of group behavior. In addition, using a tracking algorithm to obtain the behavioral trajectory of the target and understand the behavioral pattern of the target in time and space helps to identify the dynamic changes of group behavior.

[0014] Further, in S200, determining the behavioral characteristics of the group according to the behavioral characteristics of the above biological objects, the specific steps include:

[0015] S211. Classify the behavioral characteristics of the above biological objects to obtain biological behavior categories;

[0016] S212. Count the number of biological objects in the above biological behavior categories, sort the biological behavior categories from largest to smallest according to the number of biological objects to obtain a biological behavior sequence ; where represents the first biological behavior category, represents the biological behavior category, , represents the th biological object's behavioral characteristics in the

[0017] S213. Select the behavioral characteristics of the first biological behavior category from the above biological behavior sequence as the behavioral characteristics of the group.

[0018] Through classification, biological objects with similar behavioral characteristics are grouped into one category, thus identifying different behavioral patterns, which helps in the subsequent overall understanding of group behavior; by counting the number of biological objects in each category, the main behavioral patterns in the group are identified, and the behavior category with the largest number often represents the mainstream behavior of the group; the biological behavior sequence obtained by sorting according to the number not only highlights the mainstream behavior but also reflects the relative importance of other behaviors, which helps to more accurately grasp the hierarchical structure of group behavior in subsequent analysis; selecting the behavioral characteristics of the behavior category with the largest number as the behavioral characteristics of the group can most reflect the overall behavioral trend of the group. In a dynamically changing group, the mainstream behavior may change with time and environment. By real-time statistics and sorting, the behavioral characteristics of the group can be dynamically updated, thus enhancing the adaptability of recognition.

[0019] Further, in S200, determining the influence weight of biological objects on group behavior through the above behavioral characteristics of biological objects and groups, the specific steps include:

[0020] S221. The behavioral characteristics of the above biological objects include actions, postures, and positions in the group;

[0021] S222. Extract the group distribution range from the above video images, divide the group distribution range into a decision-making area, a following area, and a non-following area, and assign corresponding weights to the decision-making area, the following area, and the non-following area;

[0022] S223. Determine the additional weight of the biological object according to the position of the above biological object in the group;

[0023] S224. Compare the postures of the above biological object and the group. If the above postures are the same, the posture consistency number of the biological object is incremented by 1; otherwise, the posture inconsistency number of the biological object is incremented by 1; compare the actions of the above biological object and the group. If the above actions are the same, the action consistency number of the biological object is incremented by 1; otherwise, the action inconsistency number of the biological object is incremented by 1;

[0024] Traverse the above video image set, and based on the posture consistency number and posture inconsistency number of the above biological object, determine the posture consistency rate between the biological object and the group; based on the action consistency number and action inconsistency number of the above biological object, determine the action consistency rate between the biological object and the group;

[0025] Determine the behavior consistency rate between the biological object and the group through the above posture consistency rate and action consistency rate;

[0026] S225. Determine the influence weight of the biological object on the group behavior through the above additional weight of the biological object and the behavior consistency rate.

[0027] Assign an additional weight to the biological object according to its position in the group. The more important the position, the higher the additional weight; by comparing the postures and actions, determine the behavior consistency between the biological object and the group. If the behavior of an individual is highly consistent with the group behavior and can drive the group to act, then the biological object has a high influence.

[0028] Further, in S300, determine the influence weight of the non-biological object on the group behavior through the behavior trajectory of the above target. The specific steps include:

[0029] S311. Obtain the behavior trajectories of the above non-biological object and the biological object in the same time period, predict the behavior trajectory of the non-biological object in the next time period and the behavior trajectory of the biological object when not affected by the non-biological object; according to the predicted behavior trajectory of the above biological object, determine the group distribution range in the next time period;

[0030] S312. Determine the position where the non-biological object passes through the group distribution range in the next time period according to the predicted behavior trajectory of the above non-biological object and the group distribution range;

[0031] Determine the additional weight of the abiotic object according to the positions passed by the abiotic object within the population distribution range.

[0032] S313. Determine the moving speed of the abiotic object according to the obtained behavior trajectory and time period duration of the abiotic object; among them, multiple speed intervals are set for the abiotic object, and corresponding weights are assigned to each speed interval.

[0033] Determine the speed weight of the abiotic object according to the speed interval in which the moving speed of the abiotic object is located.

[0034] S314. Determine the influence weight of the abiotic object on the group behavior through the additional weight and speed weight of the abiotic object.

[0035] Combined with prediction, spatial analysis and weight assignment to determine the influence weight of the abiotic object on the group behavior. First, collect the behavior trajectories of abiotic objects and biotic objects within the same time period; use machine learning or physical models to predict the behavior trajectories of abiotic objects in the next time period. At the same time, predict the behavior trajectories of biotic objects when they are not affected by abiotic objects, and determine the population distribution range in the next time period according to the predicted behavior trajectories of biotic objects (that is, identify the areas where biotic objects may gather in the next time period). Then, according to the predicted behavior trajectory of the abiotic object, determine the positions within the population distribution range that it will pass through in the next time period, and evaluate the possible impact of the behavior of the abiotic object on the group behavior, that is, the additional weight. According to the obtained behavior trajectory and time period duration of the abiotic object, calculate its moving speed, and determine its speed weight according to the speed interval in which the moving speed of the abiotic object is located. Finally, calculate its influence weight on the group behavior by combining the additional weight and speed weight of the abiotic object.

[0036] Furthermore, use the following formula to calculate the behavior consistency rate between the above biotic object and the group:

[0037] ,

[0038] where represents the behavior consistency rate between the th biotic object and the group; represents the posture consistency rate between the th biotic object and the group; represents the number of consistent postures between the th The number of actions of the biological object consistent with the group; Indicates the The number of actions of the biological object inconsistent with the group; Indicates the weight of the posture in the action consistency rate; Indicates the weight of the action in the action consistency rate.

[0039] Furthermore, the influence weight of the above biological object on the group behavior is calculated using the following formula:

[0040] ,

[0041] Among them, Indicates the Influence weight of the biological object on the group behavior; Indicates the additional weight of the biological object; Indicates the decision region weight; Indicates the following region weight; Indicates the non-following region weight.

[0042] Furthermore, the following formula is used to calculate the additional weight of the above non-biological object in the next time period:

[0043] ,

[0044] Among them, Indicates the additional weight of the Non-biological object in the (T + 1)th period; Indicates the decision region weight; Indicates the following region weight; Indicates the Area of the decision region covered by the non-biological object in the (T + 1)th period; Indicates the total area of the decision region in the (T + 1)th period; Indicates the Area of the following region covered by the non-biological object in the (T + 1)th period; Indicates the total area of the following region in the (T + 1)th period.

[0045] Furthermore, the following formula is used to calculate the movement speed of the above non-biological object:

[0046] , where, Indicates the Movement speed of the non-biological object; Indicates the Behavior trajectory of the non-biological object in the Tth period; Indicates the duration of the Tth period;

[0047] Judge the above Retrieve the corresponding speed weight within the speed range to which the movement speed of the non-biological object belongs, and obtain the speed weight of the non-biological object.

[0048] Furthermore, use the following formula to calculate the influence weight of the above non-biological object on group behavior:

[0049] ,

[0050] represents the influence weight of the non-biological object on group behavior; represents the speed weight of the non-biological object.

[0051] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0052] By evaluating the influence of biological objects and non-biological objects on group behavior, not only the behavioral characteristics of biological objects are considered, but also the influence of non-biological objects on group behavior is considered, improving the comprehensiveness of group behavior recognition, and thus improving the recognition accuracy of group behavior; in addition, using a tracking algorithm to obtain the behavioral trajectory of the target and understand the behavioral pattern of the target in time and space helps to identify the dynamic changes of group behavior. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following will briefly introduce the drawings required for the embodiments. It should be understood that the following drawings only show some embodiments of the present invention, and thus should not be regarded as limiting the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings. In the drawings:

[0054] Figure 1 is the main flowchart. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] To make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the embodiments and the drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0056] First Embodiment:

[0057] In combination with Figure 1 , a group behavior recognition method based on video data is provided in the first aspect, including the following steps:

[0058] S100. Obtain a video of group behavior, sample the video of group behavior, and obtain a video image set composed of a number of video images;

[0059] S200. Extract the behavior features of the target from the above video images, where the target includes biological objects and non-biological objects; among them, a group is composed of a number of the above biological objects, and the behavior features of the group are determined according to the behavior features of the biological objects;

[0060] Determine the influence weight of the biological object on the group behavior through the behavior features of the above biological object and the group, and obtain the first weight;

[0061] The above biological objects can be people, fish, birds, etc., and the non-biological objects can be buildings, vehicles, balls, etc. By determining the influence weight of the biological object on the group behavior, the accuracy of recognition can be improved; for example, in a crowd, the behavior of decision-makers has a significant impact on the behavior of the entire group.

[0062] S300. Use a tracking algorithm to track the above target to obtain the behavior trajectory of the target;

[0063] Determine the influence weight of the non-biological object on the group behavior through the behavior trajectory of the above target, and obtain the second weight;

[0064] Consider the influence of non-biological objects (such as road layout, building location, etc.) on the group behavior, so as to improve the comprehensiveness of recognition. For example, in a traffic scenario, the road layout will significantly affect the flow patterns of vehicles and people.

[0065] S400. Obtain group behavior recognition data through the behavior features of the above target, the first weight, and the second weight. Comprehensively consider the information of biological objects and non-biological objects to improve the recognition accuracy of group behavior. For example, in a complex scenario, the behavior of a single biological object may be difficult to accurately identify, but after combining the information of its non-biological object, the behavior intention and group behavior pattern of the biological object can be judged.

[0066] By evaluating the influence of biological objects and non-biological objects on group behavior, not only the behavior features of biological objects are considered, but also the influence of non-biological objects on group behavior is considered, which improves the comprehensiveness of group behavior recognition and further improves the recognition accuracy of group behavior; in addition, using a tracking algorithm to obtain the behavior trajectory of the target and understand the behavior pattern of the target in time and space helps to identify the dynamic changes of group behavior.

[0067] A reference usage scenario is the monitoring scenario of group activities in a public square, where there are both biological objects (crowds) and non-biological objects (moving public facilities, such as vending machines, moving screens, etc.). The video data of this scenario is used to identify and analyze group behaviors, such as the flow direction of the crowd, aggregation patterns, reactions to specific events, etc.; by identifying the behavioral characteristics of biological objects (crowds) and non-biological objects (public facilities) and their impact on group behaviors, the group dynamics are understood to optimize the management and safety measures in public places.

[0068] Specifically, the surveillance camera records the video of the activities in the public square for a day. One frame is extracted from the activity video every 5 seconds to form a video image set consisting of hundreds of frames. The video image set contains the positions and states of the crowds and public facilities in the square at different time periods; calculate the moving speed of each crowd in each frame, and identify the areas where the crowds gather through spatial density analysis; determine the overall flow direction based on the moving direction of the majority of the crowds, analyze the relationship between the change in crowd speed and the group flow direction, and obtain the first weight; if the crowd speed increases and the direction is consistent, it is considered that this crowd has a greater impact on group behavior and a higher weight is assigned.

[0069] Use object detection and tracking algorithms (such as YOLO+SORT) to track the crowds and public facilities between consecutive frames to obtain their behavioral trajectories. By analyzing the trajectories, record the number of people in front of the vending machine and its usage in each frame, and detect whether the content played on the screen attracts the crowd to stop. If a large number of people gather in front of the vending machine, it indicates that it has a certain impact on group behavior, and a weight is assigned according to the number of people gathered and the time to obtain the second weight; if a certain moving screen attracts the crowd to stay for a long time, its weight should be increased accordingly.

[0070] Combine the behavioral characteristics of biological objects and non-biological objects, the first weight and the second weight for comprehensive analysis; if the group flow direction is mainly affected by the crowd speed (high first weight), and the vending machine only affects the crowd distribution in a small range (low second weight), then the group behavior recognition data will mainly reflect the flow pattern of the crowd. Generate a report containing the results of group behavior recognition, including the group flow direction, aggregation hotspots, reaction speed to specific events (such as promotional activities, emergencies), etc., to provide decision-making support for the management of public places.

[0071] In volleyball, the movement trajectory of the volleyball usually determines the positions and behavioral tendencies of the athletes, which plays an important role in judging the group behavior of the athletes.

[0072] Second Embodiment:

[0073] Based on the first embodiment, in S200, determine the behavioral characteristics of the group according to the behavioral characteristics of the above biological objects. The specific steps include:

[0074] S211. Classify the behavioral characteristics of the above biological objects to obtain biological behavior categories;

[0075] S212. Count the number of biological objects in the above biological behavior categories, sort the biological behavior categories from largest to smallest according to the number of biological objects, and obtain a biological behavior sequence ; where represents the first biological behavior category, represents the biological behavior category, , represents the in the biological behavior category the behavioral characteristics of the

[0076] S213. Select the behavioral characteristics of the first biological behavior category from the above biological behavior sequence as the behavioral characteristics of the group.

[0077] Through classification, biological objects with similar behavioral characteristics are grouped into one category, thus identifying different behavioral patterns, which helps to grasp the overall group behavior in the follow-up; by counting the number of biological objects in each category, the main behavioral patterns in the group are identified. The behavioral category with the largest number often represents the mainstream behavior of the group; the biological behavior sequence obtained by sorting according to the number not only highlights the mainstream behavior but also reflects the relative importance of other behaviors, which helps to more accurately grasp the hierarchical structure of the group behavior in the follow-up analysis; selecting the behavioral characteristics of the behavioral category with the largest number as the behavioral characteristics of the group can reflect the overall behavioral trend of the group to the greatest extent. In a dynamically changing group, the mainstream behavior may change with time and environment. By real-time counting and sorting, the behavioral characteristics of the group can be dynamically updated, thus enhancing the adaptability of identification.

[0078] Third Embodiment:

[0079] On the basis of any of the above embodiments, in S200, determine the influence weight of the biological object on the group behavior through the above behavioral characteristics of the biological object and the group. The specific steps include:

[0080] S221. The behavioral characteristics of the above biological object include actions, postures, and positions in the group;

[0081] S222. Extract the group distribution range from the above video image, divide the group distribution range into a decision-making area, a following area, and a non-following area, and assign corresponding weights to the decision-making area, the following area, and the non-following area;

[0082] S223. Determine the additional weight of the biological object according to the position of the biological object in the group.

[0083] S224. Compare the postures of the biological object and the group. If the postures are the same, increment the posture consistency count of the biological object by 1; otherwise, increment the posture inconsistency count of the biological object by 1. Compare the actions of the biological object and the group. If the actions are the same, increment the action consistency count of the biological object by 1; otherwise, increment the action inconsistency count of the biological object by 1.

[0084] Traverse the above video image set. Based on the posture consistency count and posture inconsistency count of the biological object, determine the posture consistency rate between the biological object and the group. Based on the action consistency count and action inconsistency count of the biological object, determine the action consistency rate between the biological object and the group.

[0085] Determine the behavior consistency rate between the biological object and the group through the above posture consistency rate and action consistency rate.

[0086] S225. Determine the influence weight of the biological object on the group behavior through the additional weight of the biological object and the behavior consistency rate.

[0087] Assign an additional weight to the biological object according to its position in the group. The more important the position, the higher the additional weight. By comparing postures and actions, determine the behavioral consistency between the biological object and the group. If an individual's behavior is highly consistent with the group behavior and can drive the group to act, then the biological object has a high influence.

[0088] A reference usage scenario is the on-site monitoring scenario of a large music festival. The biological objects are mainly the audience participating in the music festival, and the non-biological objects include the stage, sound equipment, etc. The monitoring video data is used to identify and analyze the behavioral characteristics of the audience and their influence on the group behavior, so as to optimize the on-site management and improve the audience experience. At the music festival site, the audience is active in different areas such as the stage, in front of the stage, and the rest area. Extract the actions (such as dancing, walking, standing), postures (such as facing the stage, side-facing the stage, back to the stage), and positions in the group (such as on the stage, in front of the stage, rest area) of each audience from the video images. Through spatial density analysis, identify the high-density area (such as the stage) as the decision area, the medium-density area (such as in front of the stage) as the following area, and the low-density area (such as the rest area) as the non-following area. According to the position of the audience in the group, for example, the audience located in the decision area has a higher additional weight, followed by the following area, and the lowest in the non-following area. In addition, factors such as the influence of the audience in the group can also be considered. For example, the additional weight of an obvious leader (such as a person holding a flag or leading the chant) is increased additionally.

[0089] Traverse the video image set. For each viewer, compare their postures with those of the surrounding group. If the postures are similar (such as all facing the stage and dancing), then the number of consistent postures is incremented by 1; otherwise, the number of inconsistent postures is incremented by 1. Similarly, compare the actions of the viewer with those of the group. If the actions are the same (such as all dancing to the music rhythm), then the number of consistent actions is incremented by 1; otherwise, the number of inconsistent actions is incremented by 1. Calculate the posture consistency rate based on the number of consistent postures and the number of inconsistent postures, and calculate the action consistency rate based on the number of consistent actions and the number of inconsistent actions. Then, combine the two to obtain the behavior consistency rate of the biological object and the group.

[0090] Combine the additional weight of the biological object and the behavior consistency rate to calculate the influence weight of the biological object on the group behavior. For example, a viewer located in the decision area and with highly consistent behavior will have a greater impact on the group behavior, so its influence weight is higher.

[0091] In terms of application, by identifying viewers with high influence weights, their needs and behaviors can be given priority attention. If the behavior consistency rate of viewers in a certain area suddenly drops, it may mean that an emergency has occurred or the emotions of the viewers have fluctuated, and measures need to be taken in a timely manner to ensure safety; by analyzing the behavior characteristics of the viewers and their impact on the group behavior, the on-site layout, event arrangements, etc. can be adjusted to improve the overall experience of the viewers.

[0092] In a specific embodiment, the following formula is used to calculate the behavior consistency rate of the above biological object and the group:

[0093] ,

[0094] Among them, represents the behavior consistency rate of the biological object and the group; represents the posture consistency rate of the biological object and the group; represents the number of consistent postures of the biological object and the group; represents the number of inconsistent postures of the biological object and the group; represents the action consistency rate of the biological object and the group; represents the number of consistent actions of the biological object and the group; represents the number of inconsistent actions of the biological object and the group; represents the weight of the posture in the behavior consistency rate; represents the weight of the action in the behavior consistency rate.

[0095] In a specific embodiment, the following formula is used to calculate the influence weight of the above biological object on the group behavior:

[0096] ,

[0097] Among them, represents the influence weight of the biological object on the group behavior; represents the additional weight of the biological object; represents the decision region weight; represents the following region weight; represents the non-following region weight.

[0098] Fourth Embodiment:

[0099] Based on any of the above embodiments, in S300, the influence weight of the non-biological object on the group behavior is determined through the behavior trajectory of the above target. The specific steps include:

[0100] S311. Obtain the behavior trajectories of the above non-biological object and the biological object in the same time period, predict the behavior trajectory of the non-biological object in the next time period and the behavior trajectory of the biological object when not affected by the non-biological object; determine the group distribution range in the next time period according to the predicted behavior trajectory of the above biological object;

[0101] S312. Determine the position where the non-biological object passes through the group distribution range in the next time period according to the predicted behavior trajectory of the above non-biological object and the group distribution range;

[0102] Determine the additional weight of the non-biological object according to the position where the above non-biological object passes through the group distribution range;

[0103] S313. Determine the movement speed of the non-biological object according to the obtained behavior trajectory of the above non-biological object and the time period duration; among them, multiple speed intervals are set for the above non-biological object, and corresponding weights are assigned to each speed interval;

[0104] Determine the speed weight of the non-biological object according to the speed interval in which the movement speed of the above non-biological object is located;

[0105] S314. Determine the influence weight of the non-biological object on the group behavior through the additional weight and speed weight of the above non-biological object.

[0106] ​Combines prediction, spatial analysis, and weight assignment to determine the influence weight of non-biological objects on group behavior. First, collect the behavior trajectories of non-biological objects and biological objects within the same time period, which can be obtained through video surveillance, sensor data, or other tracking technologies; use machine learning or physical models to predict the behavior trajectories of non-biological objects in the next time period. At the same time, predict the behavior trajectories of biological objects when not affected by non-biological objects. According to the predicted behavior trajectories of biological objects, determine the group distribution range in the next time period (i.e., identify the areas where biological objects may gather in the next time period). Then, according to the predicted behavior trajectories of non-biological objects, determine the positions of the group distribution ranges that they will pass through in the next time period, and evaluate the possible impact of the behavior of non-biological objects on group behavior, that is, the additional weight. According to the obtained behavior trajectories of non-biological objects and the time period duration, calculate their movement speeds, and determine their speed weights according to the speed intervals in which the movement speeds of non-biological objects are located. Finally, by combining the additional weight and speed weight of non-biological objects, calculate their influence weight on group behavior.

[0107] A reference usage scenario is to study the influence of different vehicles (such as cars, buses, bicycles, etc.) on the behavior of people in an open square in order to provide decision-making support for urban planning and management. There are often crowds gathering and vehicles shuttling in the square. Collect the behavior trajectories of vehicles and people within the same time period through the surveillance cameras in the square. The behavior trajectory data includes information such as vehicle type, speed, position, and the walking paths and speeds of the crowd. Use machine learning algorithms (such as deep learning models) to predict the behavior trajectories of vehicles in the next time period. At the same time, predict the possible walking paths of the crowd when not affected by vehicles. According to the predicted behavior trajectories of the crowd, determine the areas where the crowd may gather in the next time period. The gathering areas may include the entrances, exits, rest areas, etc. of the square. According to the predicted behavior trajectories of vehicles, determine the areas where the crowd gathers that the vehicles will pass through in the next time period. If a bus traveling at a high speed passes near a crowd gathering area, its additional weight is relatively high; in addition, the determination of the additional weight can also consider factors such as vehicle type (for example, the impact of a car on the crowd may be less than that of a bus), speed, and relative position to the crowd.

[0108] According to the collected vehicle behavior trajectory data, calculate the average speed of each vehicle, set multiple speed intervals for the vehicles (such as low speed, medium speed, high speed), and assign corresponding weights to each interval. The weight of the high-speed interval is greater than that of the low-speed interval. Determine the speed weight of each vehicle according to the speed interval in which its speed is located. Combine the additional weight and speed weight of the vehicle to calculate the influence weight of each vehicle on the behavior of the crowd.

[0109] Through the above process, the influence weights of different vehicles on crowd behavior are obtained. These weights can be used to evaluate the impact of vehicles on aspects such as the flow and safety of the crowd in the square, and provide decision-making support for urban planners and managers. For example, the vehicle driving routes can be adjusted, the vehicle speeds can be restricted, or crowd protection measures can be increased based on these weights.

[0110] In a specific embodiment, the following formula is used to calculate the additional weight of the above non-biological object in the next time period:

[0111] ,

[0112] where, represents the additional weight of the th non-biological object in the T+1 time period; represents the decision area weight; represents the following area weight; represents the area of the th non-biological object covering the decision area in the T+1 time period; represents the total area of the decision area in the T+1 time period; represents the area of the th non-biological object covering the following area in the T+1 time period; represents the total area of the following area in the T+1 time period.

[0113] In a specific embodiment, the following formula is used to calculate the movement speed of the above non-biological object:

[0114] , where, represents the movement speed of the th non-biological object; represents the behavior trajectory of the th non-biological object in the T time period; represents the duration of the T time period;

[0115] Judge the speed interval to which the movement speed of the above th non-biological object belongs, retrieve the corresponding speed weight, and obtain the speed weight of the th non-biological object.

[0116] In a specific embodiment, the following formula is used to calculate the influence weight of the above non-biological object on group behavior:

[0117] ,

[0118] represents the influence weight of the th non-biological object on group behavior; represents the speed weight of the th non-biological object.

[0119] The specific embodiments described above further elaborate on the objective, technical solution, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for identifying group behaviors based on video data, characterized in that, Including the following steps: S100. Obtain a group behavior video, sample the group behavior video to obtain a video image set composed of a number of video images; S200. Extract the behavior characteristics of the target from the video images, where the target includes biological objects and non-biological objects; among them, a group is composed of a number of the biological objects, and determine the behavior characteristics of the group according to the behavior characteristics of the biological objects; Determine the influence weight of the biological object on the group behavior through the behavior characteristics of the biological object and the group to obtain a first weight, and the specific steps include: S221. The behavior characteristics of the biological object include actions, postures, and positions in the group; S222. Extract the group distribution range from the video images, divide the group distribution range into a decision-making area, a following area, and a non-following area, and assign corresponding weights to the decision-making area, the following area, and the non-following area; S223. Determine the additional weight of the biological object according to the position of the biological object in the group; S224. Compare the postures of the biological object and the group. If the postures are the same, the posture consistency number of the biological object is incremented by 1; otherwise, the posture inconsistency number of the biological object is incremented by 1; compare the actions of the biological object and the group. If the actions are the same, the action consistency number of the biological object is incremented by 1; otherwise, the action inconsistency number of the biological object is incremented by 1; Traverse the video image set, and determine the posture consistency rate between the biological object and the group based on the posture consistency number and the posture inconsistency number of the biological object; determine the action consistency rate between the biological object and the group based on the action consistency number and the action inconsistency number of the biological object; Determine the behavior consistency rate between the biological object and the group through the posture consistency rate and the action consistency rate; S225. Determine the influence weight of the biological object on the group behavior through the additional weight of the biological object and the behavior consistency rate; S300. Use a tracking algorithm to track the target to obtain the behavior trajectory of the target; Determine the influence weight of the non-biological object on the group behavior through the behavior trajectory of the target to obtain a second weight, and the specific steps include: S311. Obtain the behavior trajectory of the non-biological object and the behavior trajectory of the biological object in the same time period, predict the behavior trajectory of the non-biological object in the next time period and the behavior trajectory of the biological object when not affected by the non-biological object; determine the group distribution range in the next time period according to the predicted behavior trajectory of the biological object; S312. Determine the position where the non-biological object passes through the group distribution range in the next time period according to the predicted behavior trajectory of the non-biological object and the group distribution range; Determine the additional weight of the non-biological object according to the position where the non-biological object passes through the group distribution range; S313. Determine the movement speed of the non-biological object according to the obtained behavior trajectory of the non-biological object and the time period duration; among them, multiple speed intervals are set for the non-biological object, and corresponding weights are assigned to each speed interval; Determine the speed weight of the non-biological object according to the speed interval where the movement speed of the non-biological object is located; S314. Determine the influence weight of the non-biological object on the group behavior based on the additional weight and speed weight of the non-biological object; S400. Obtain the group behavior recognition data based on the behavior characteristics of the target, the first weight, and the second weight.

2. The group behavior recognition method according to claim 1, wherein In S200, determine the behavior characteristics of the group according to the behavior characteristics of the biological object. The specific steps include: S211. Classify the behavior characteristics of the biological object to obtain the biological behavior categories; S212. Count the number of biological objects in the biological behavior categories, sort the biological behavior categories in descending order according to the number of biological objects to obtain a biological behavior sequence ; where represents the first biological behavior category, represents the biological behavior category, , represents the behavior characteristics of the th biological object in the biological behavior category; S213. Select the behavioral characteristics of the first biological behavior category from the biological behavior sequence as the behavioral characteristics of the group. as the behavioral characteristics of the group.

3. The group behavior recognition method according to claim 1, characterized in that Calculate the behavior consistency rate of the biological object and the group using the following formula: ， Among them, represents the consistency rate of the behavior of the biological object and the group; represents the consistency rate of the posture of the biological object and the group; represents the number of consistent postures of the biological object and the group; represents the number of inconsistent postures of the biological object and the group; represents the consistency rate of the actions of the biological object and the group; represents the number of consistent actions of the biological object and the group; represents the number of inconsistent actions of the biological object and the group; represents the weight of the posture in the behavior consistency rate; represents the weight of the action in the behavior consistency rate.

4. The group behavior recognition method according to claim 3, wherein Calculate the influence weight of the biological object on the group behavior using the following formula: ， Among them, represents the influence weight of the biological object on the group behavior; represents the additional weight of the biological object; represents the decision region weight; represents the following region weight; represents the non-following region weight.

5. The group behavior recognition method according to claim 1, characterized in that, Calculate the additional weight of the non-biological object in the next time period using the following formula: ， Among them, represents the additional weight of the non-biological object in the T+1 period at the non-biological object; represents the decision area weight; represents the following area weight; represents the area of the non-biological object covering the decision area in the T+1 period at the non-biological object; represents the total area of the decision area in the T+1 period; represents the area of the non-biological object covering the following area in the T+1 period at the non-biological object; represents the total area of the following area in the T+1 period.

6. The group behavior recognition method according to claim 5, wherein Calculate the movement speed of the non-biological object using the following formula: , where represents the movement speed of the non-biological object; represents the behavior trajectory of the non-biological object in the T period; represents the duration of the T period; Determine the speed interval to which the movement speed of the non-biological object belongs, retrieve the corresponding speed weight, and obtain the speed weight of the non-biological object. speed weight of the non-biological object.

7. The group behavior recognition method according to claim 6, characterized in that Calculate the influence weight of the non-biological object on the group behavior using the following formula: ， Indicates the influence weight of non-biological objects on group behavior; Indicates the speed weight of non-biological objects.

Images