Unmanned aerial vehicle and robot dog collaborative inspection system

Abstract

The invention discloses an unmanned aerial vehicle and robot dog cooperative inspection system, and relates to the technical field of unmanned aerial vehicles, the unmanned aerial vehicle and robot dog cooperative inspection system comprises an unmanned aerial vehicle inspection module, a robot dog inspection module and a cooperative control module, the unmanned aerial vehicle inspection module is used for aerial aerial view of large-scale crowds, and real-time calculation of crowd density, velocity field and congestion hotspots; the robot dog inspection module is used for carrying out close inspection in a ground crack corridor, detecting gait stability and falling individuals in real time, forming a soft boundary to carry out differential diversion guidance and carrying out physical intervention on dangerous situations, and carrying out real-time inspection on the ground crack corridor. The cooperative control module is used for integrating the aerial view angle of the unmanned aerial vehicle and the ground detection information of the robot dog, dynamically updating a routing inspection path and triggering sound-light alarm and linkage guidance when a risk is identified, so that air-ground integrated risk prevention and control are realized, the unmanned aerial vehicle routing inspection module comprises a visual perception unit, and the unmanned aerial vehicle routing inspection system has the characteristic of near execution.

Description

Technical Field

[0001] The present invention relates to the technical field of drones, specifically to a collaborative inspection system of drones and robotic dogs. Background Art

[0002] In large-scale crowd activities such as New Year's Eve celebrations and parades in squares, there are risks of crowd crush and trampling. When the crowd becomes agitated, the lateral pressure wave caused by external forces or internal fluctuations will push individuals towards fixed obstacles. Most deaths are caused by hypoxia due to chest compression and trampling.

[0003] Drones play a role in early detection, providing global guidance and warnings in large-scale crowd parades and activities. Drones rely on technologies such as overhead vision or infrared to calculate the density and velocity fields of the crowd in each area, as well as the reverse flow ratio and congestion hotspots, make short-term congestion evolution predictions, and conduct long-distance guidance through aerial sound amplification and ground projection arrows.

[0004] However, relying solely on drones for inspection, their aerial perspective is restricted by occlusion and depression angles, making it difficult to accurately locate visible ground hazards, and it is also impossible to directly intervene in the crowd. Coupled with the herd mentality and inertia of the crowd leading to implementation gaps, it is difficult to transform the global solution into local behavior changes. Therefore, it is necessary to design a collaborative inspection system of drones and robotic dogs that can perform close-up operations. Summary of the Invention

[0005] The purpose of the present invention is to provide a collaborative inspection system of drones and robotic dogs to solve the problems raised in the above background art.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A collaborative inspection system of drones and robotic dogs, including a drone inspection module, a robotic dog inspection module, and a collaborative control module. The drone inspection module is used to overlook a large-scale crowd from the air, calculate the crowd density, velocity field, and congestion hotspots in real time, conduct congestion evolution prediction, and provide global guidance through sound amplification or projection. The robotic dog inspection module is used to conduct close-up inspections in the ground fissure corridor, detect gait stability and fallen individuals in real time, form a soft boundary for micro-divergence guidance, and physically intervene in dangerous situations. The collaborative control module is used to integrate the aerial perspective of the drone and the ground detection information of the robotic dog, dynamically update the inspection path, trigger audible and visual alarms and linkage guidance when identifying risks, and achieve risk prevention and control of air-ground integration.

[0007] According to the above technical solution, the UAV inspection module includes a visual perception unit, a crowd density calculation module, a velocity field calculation module, and a reverse flow analysis module. The visual perception unit is electrically connected to the crowd density calculation module and the velocity field calculation module, and the velocity field calculation module is electrically connected to the reverse flow analysis module. The visual perception unit uses a top-down camera and an infrared sensor to acquire the position of individuals in the crowd. The crowd density calculation module is used to calculate the crowd distribution density in each area based on the perception data. The velocity field calculation module is used to perform field modeling of the overall movement speed and direction of the crowd. The reverse flow analysis module is used to detect and evaluate the proportion of individuals moving in reverse flow in the crowd to identify potential congestion or disturbance risks. The robot dog inspection module includes a walking control module, a gait analysis module, a hazardous individual detection module, a guidance and intervention module, and a movement and encirclement control module. The gait analysis module is electrically connected to the guidance and intervention module, and the hazardous individual detection module is electrically connected to the movement and encirclement control module. The walking control module is used to control the robot dog to move stably in the crevice corridor and perform inspection tasks. The gait analysis module is used to analyze the gait stability of individuals in real time to identify abnormal movements. The hazardous individual detection module is used to detect individuals in a state of falling or imbalance and mark their positions. The guidance and intervention module is used to form a soft boundary through posture and light signals to guide the crowd to achieve micro-diversion. The movement and encirclement control module is used to form a temporary protective circle around hazardous individuals to prevent secondary trampling. The collaborative control module includes a crowd segmentation module, a gap corridor generation module, a path update module, and a data transmission module. The crowd segmentation module is electrically connected to the gap corridor generation module and the path update module in sequence. The data transmission module is electrically connected to all modules of the UAV inspection module and the robot dog inspection module. The crowd segmentation module is used to dynamically divide a large crowd into several relatively independent diversion units. The gap corridor generation module is used to generate corridors with safe intervals between adjacent diversion units for inspection and guidance. The path update module is used to dynamically adjust the robot dog's inspection path according to the real-time movement of the crowd and the congestion situation. The data transmission module realizes information interaction and linkage command transmission between the UAV and the robot dog.

[0008] According to the above technical solution, the working steps of the system include: S1. The drone flies above the crowd and detects the crowd, generates a Cartesian coordinate map of the ground, and collects data on the location, density, velocity field and reverse flow ratio of individuals in the crowd; S2. Divide the crowd into several flow units. After determining the outline and number of people in each flow unit, generate safe intervals between adjacent units along the outline edge. The fissure corridor serves as the inspection path for the robot dog; S3. Multiple robot dogs patrol and inspect the rift corridor. When someone blocks the robot dog's path, they use gestures and light signals to remind the crowd to avoid it. They perform gait analysis and dangerous individual detection, and promptly identify and protect individuals who fall or lose balance. S4. When the system determines that there is a risk of congestion and stampede, it triggers an audible and visual alarm and guides the crowd to divert through drone amplification, projection, robot dog gestures, and light signals to reduce local pressure. S5. Based on the data collected in real time by drones and robot dogs, the risk of trampling in each diversion unit is analyzed in real time, and the gap corridor is updated in real time in combination with the changes in the position of the crowd.

[0009] According to the above technical solution, in step S1, the data collection of individual location, density, velocity field, and reverse flow ratio in the crowd specifically includes: S1-1. Mark the location of each individual in the crowd, and generate corresponding coordinates for each individual. ,in Assign numbers to individuals in the population, mark each coordinate in the coordinate system, count the density of coordinate points in each area, track and record the position of each individual, and display the change of coordinate position over time in the coordinate system; S1-2. Obtain the velocity vector based on the positional changes of the individual coordinates over time. For any defined region, take the velocity vector of all individuals within that region. The averaged velocity direction is taken as the mainstream flow direction of the population in the area. When the angle between the velocity vector of an individual and the mainstream direction exceeds a set threshold, the population dynamics are considered to be in the mainstream direction. When an individual is identified as a reverse-flowing individual, the proportion of reverse-flowing individuals to the total number of individuals in the region is calculated as the reverse-flowing proportion.

[0010] According to the above technical solution, in step S2, dividing the crowd into several diversion units specifically involves: S2-1. If there are obstacles on the ground, use the outline of the obstacles as the boundary line of the diversion unit; if there are no obstacles on the ground, set a threshold for the number of people. As a basis for division, when the total number of people is At that time, it was divided into Each part, and will Round up to get , As the initial number of diversion units, the number of people in each diversion unit is roughly divided. S2-2, The drone scan obtains the overall coverage area of ​​the crowd. Connect the coordinates of the edge individuals within the diversion unit to obtain the initial contour, and then calculate the area of ​​the initial contour of the diversion unit. and the initial number of people To calculate population density, if If the population density in this diversion unit is high, then... To determine the density threshold, the outline size and number of people in this diversion unit need to be reduced, and individuals in the edge region with the highest population density within this diversion unit should be removed. If the population density of this subdivision unit is low, this subdivision unit will accommodate individuals removed from other subdivision units and generate new profiles. This process will continue until the population density of each subdivision unit approaches a certain level. Number of people in each diversion unit and outline area It can be determined.

[0011] According to the above technical solution, in step S3, the gait analysis and high-risk individual detection are specifically performed as follows: S3-1. Scan the leg movements of all individuals in the crowd and record the frequency of each individual's leg movements. When a scan detects... When the frequency of leg movements of more than one individual changes abruptly, and due to the difficulty in direct location, the lateral angle between the individual experiencing the frequency change and the current robot dog is calculated. This is then done by scanning all individuals at this lateral angle using a drone from above. When multiple individuals within the group are simultaneously scanned... When all individual coordinates are of individuals flowing in the opposite direction, it is determined that there is a risk of stampede in the current diversion unit. S3-2. The robot dog identifies the legs and torso of the human body. When a person's torso is found within the leg height range of a crowd during scanning, it determines that the person has fallen. The robot dog issues a warning and shoos away the crowd ahead, while simultaneously adjusting its patrol path to temporarily deviate from the crevice corridor. The drone then uses this location to pinpoint the coordinates of the fallen person, with a radius of [missing information]. All robot dog inspection paths within the area change to the direction towards this coordinate. When they reach the person who has fallen, the robot dog moves along the radius... Conduct a circular inspection until maintenance personnel arrive.

[0012] According to the above technical solution, in step S4, when it is determined that there is a risk of congestion and stampede, the proportion of reverse flow individuals in the diversion unit is calculated, that is, the number of reverse flow individuals in a certain diversion unit. Per capita proportion The higher the value, and the higher the absolute value of the velocity vector of all individuals in the splitting unit. The larger the value, the greater the risk of a stampede. The larger it is, the more complex the calculation formula is. The decibel level of the drone's amplification and the brightness of the light signals were both related to The current stampede risk value is directly proportional to the risk level. The higher the level, the more prominent the warnings and guidance should be, in order to avoid hindering the normal activities of the crowd.

[0013] According to the above technical solution, in step S5, the outline of each diversion unit is updated according to the overall movement direction of the crowd, thereby adjusting the inspection path in real time. When an individual leaves the original diversion unit and the individual's distance from the outline of the original diversion unit exceeds a distance threshold, the system will detect the individual leaving the original diversion unit. At that time, this individual is included in the diversion unit it entered, at fixed intervals. The population density was recalculated, and the contours of each diversion unit were fine-tuned using the S2-2 method.

[0014] Compared with the prior art, the beneficial effects achieved by the present invention are as follows: The present invention proposes an air-ground collaborative method for inspecting crowd thinning and trampling risk. The drone constructs a global guidance field above, dynamically dividing the planar distribution of a large crowd into several diversion units, and forming a gap corridor with a minimum safety interval between adjacent diversion units. Multiple drones perform synchronous inspections in the gap corridor, forming a yieldable soft boundary with posture and light signals. The inspection path is updated according to the real-time movement of the crowd, achieving low-disturbance physical segmentation of the crowd without affecting normal personnel movement, and preventing excessive crowd gathering. During the inspection process, the robot dog monitors the gait stability of the crowd in real time. When it judges that there is a risk of trampling based on the aerial view of the drone, it will issue an audible and visual alarm and guide the movement direction of the crowd. At the same time, it will promptly detect vulnerable individuals who have fallen. The robot dog will form a temporary encirclement around them to prevent them from being trampled by the crowd and slowly guide them to the side so that they can be handed over to medical personnel. This achieves physical intervention in dangerous situations. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a schematic diagram of the overall modular structure of the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Please see Figure 1This invention provides a technical solution: a drone and robot dog collaborative inspection system, including a drone inspection module, a robot dog inspection module, and a collaborative control module. The drone inspection module is used to overlook large crowds from the air, calculate crowd density, velocity field, and congestion hotspots in real time, predict congestion evolution, and provide global guidance through sound amplification or projection. The robot dog inspection module is used to conduct close-range inspections in ground-crack corridors, detect gait stability and fallen individuals in real time, form soft boundaries for micro-diversion guidance, and physically intervene in dangerous situations. The collaborative control module is used to integrate the aerial perspective of the drone and the ground detection information of the robot dog, dynamically update the inspection path, and trigger sound and light alarms and linkage guidance when risks are identified, so as to achieve integrated air-ground risk prevention and control. The drone inspection module includes a visual perception unit, a crowd density calculation module, a velocity field calculation module, and a reverse flow analysis module. The visual perception unit is electrically connected to the crowd density calculation module and the velocity field calculation module, and the velocity field calculation module is electrically connected to the reverse flow analysis module. The visual perception unit uses a top-down camera and an infrared sensor to obtain the position of individuals in the crowd. The crowd density calculation module is used to calculate the crowd distribution density in each area based on the perception data. The velocity field calculation module is used to perform field modeling of the overall movement speed and direction of the crowd. The reverse flow analysis module is used to detect and evaluate the proportion of individuals moving in the reverse flow in the crowd to identify potential congestion or disturbance risks. The robot dog inspection module includes a walking control module, a gait analysis module, a hazardous individual detection module, a guidance and intervention module, and a mobile encirclement control module. The gait analysis module is electrically connected to the guidance and intervention module, and the hazardous individual detection module is electrically connected to the mobile encirclement control module. The walking control module is used to control the robot dog to move stably in the crevice corridor and perform inspection tasks. The gait analysis module is used to analyze the gait stability of individuals in real time to identify abnormal movements. The hazardous individual detection module is used to detect individuals in a state of falling or imbalance and mark their positions. The guidance and intervention module is used to form soft boundaries through posture and light signals to guide the crowd to achieve micro-diversion. The mobile encirclement control module is used to form a temporary protective circle around hazardous individuals to prevent secondary trampling. The collaborative control module includes a crowd segmentation module, a gap corridor generation module, a path update module, and a data transmission module. The crowd segmentation module is electrically connected to the gap corridor generation module and the path update module in sequence. The data transmission module is electrically connected to all modules of the drone inspection module and the robot dog inspection module. The crowd segmentation module is used to dynamically divide a large crowd into several relatively independent diversion units. The gap corridor generation module is used to generate corridors with safe intervals between adjacent diversion units for inspection and guidance. The path update module is used to dynamically adjust the robot dog's inspection path according to the real-time movement of the crowd and the congestion situation. The data transmission module realizes information interaction and linkage command transmission between the drone and the robot dog. The system's operating steps include: S1. The drone flies above the crowd and detects the crowd, generates a Cartesian coordinate map of the ground, and collects data on the location, density, velocity field and reverse flow ratio of individuals in the crowd; S2. Divide the crowd into several flow units. After determining the outline and number of people in each flow unit, generate safe intervals between adjacent units along the outline edge. The fissure corridor serves as the inspection path for the robot dog; S3. Multiple robot dogs patrol and inspect the rift corridor. When someone blocks the robot dog's path, they use gestures and light signals to remind the crowd to avoid it. They perform gait analysis and dangerous individual detection, and promptly identify and protect individuals who fall or lose balance. S4. When the system determines that there is a risk of congestion and stampede, it triggers an audible and visual alarm and guides the crowd to divert through drone amplification, projection, robot dog gestures, and light signals to reduce local pressure. S5. Based on the data collected in real time by drones and robot dogs, the risk of trampling in each diversion unit is analyzed in real time, and the gap corridor is updated in real time in combination with the changes in the position of the crowd. In S1, the data collected specifically include the location, density, velocity field, and proportion of reverse flow among individuals in the crowd: S1-1. Mark the location of each individual in the crowd, and generate corresponding coordinates for each individual. ,in Assign numbers to individuals in the population, mark each coordinate in the coordinate system, count the density of coordinate points in each area, track and record the position of each individual, and display the change of coordinate position over time in the coordinate system; S1-2. Obtain the velocity vector based on the positional changes of the individual coordinates over time. For any defined region, take the velocity vector of all individuals within that region. The averaged velocity direction is taken as the mainstream flow direction of the population in the area. When the angle between the velocity vector of an individual and the mainstream direction exceeds a set threshold, the population dynamics are considered to be in the mainstream direction. When the time is right, it is determined to be a reverse flow individual, and the proportion of reverse flow individuals to the total number of individuals in the region is counted as the reverse flow proportion; In S2, the crowd is divided into several flow control units as follows: S2-1. If there are obstacles on the ground, use the outline of the obstacles as the boundary line of the diversion unit; if there are no obstacles on the ground, set a threshold for the number of people. As a basis for division, when the total number of people is At that time, it was divided into Each part, and will Round up to get , As the initial number of diversion units, the number of people in each diversion unit is roughly divided. S2-2, The drone scan obtains the overall coverage area of ​​the crowd. Connect the coordinates of the edge individuals within the diversion unit to obtain the initial contour, and then calculate the area of ​​the initial contour of the diversion unit. and the initial number of people To calculate population density, if If the population density in this diversion unit is high, then... To determine the density threshold, the outline size and number of people in this diversion unit need to be reduced, and individuals in the edge region with the highest population density within this diversion unit should be removed. If the population density of this subdivision unit is low, this subdivision unit will accommodate individuals removed from other subdivision units and generate new profiles. This process will continue until the population density of each subdivision unit approaches a certain level. Number of people in each diversion unit and outline area It can be determined; In S3, the specific steps for performing gait analysis and at-risk individual detection are as follows: S3-1. Scan the leg movements of all individuals in the crowd and record the frequency of each individual's leg movements. When a scan detects... When the frequency of leg movements of more than one individual changes abruptly, and due to the difficulty in direct location, the lateral angle between the individual experiencing the frequency change and the current robot dog is calculated. This is then done by scanning all individuals at this lateral angle using a drone from above. When multiple individuals within the group are simultaneously scanned... When all individual coordinates are of individuals flowing in the opposite direction, it is determined that there is a risk of stampede in the current diversion unit. Conventional drone judgment methods can only identify individuals flowing in the opposite direction and their direction, but cannot accurately reflect the chaotic state of the crowd. Changes in the crowd's attention may cause a sudden increase in individuals flowing in the opposite direction. In this case, judging it as a risk of stampede will cause many misjudgments. However, by adding the chaotic leg frequency, the situation of stampede risk can be determined more accurately. Compared with simple aerial view judgment and ground judgment, this judgment method can achieve accurate positioning and high recognition accuracy. S3-2. The robot dog identifies the legs and torso of the human body. When a person's torso is found within the leg height range of a crowd during scanning, it determines that the person has fallen. The robot dog issues a warning and shoos away the crowd ahead, while simultaneously adjusting its patrol path to temporarily deviate from the crevice corridor. The drone then uses this location to pinpoint the coordinates of the fallen person, with a radius of [missing information]. All robot dog inspection paths within the area change to the direction towards this coordinate. When they reach the person who has fallen, the robot dog moves along the radius... Conduct a circular inspection until maintenance personnel arrive. In S4, when a risk of congestion and stampede is determined, the proportion of individuals flowing in the opposite direction within a diversion unit is calculated, i.e., the number of individuals flowing in the opposite direction within a given diversion unit. Per capita proportion The higher the value, and the higher the absolute value of the velocity vector of all individuals in the splitting unit. The larger the value, the greater the risk of a stampede. The larger it is, the more complex the calculation formula is. The decibel level of the drone's amplification and the brightness of the light signals were both related to The current stampede risk value is directly proportional to the risk level. The higher the level, the more prominent the warnings and guidance should be, in order to avoid hindering the normal activities of the crowd; In S5, the outlines of each diversion unit are updated according to the overall movement direction of the crowd, thereby adjusting the inspection path in real time. When an individual leaves the original diversion unit and the individual's distance from the outline of the original diversion unit exceeds a distance threshold, the system will detect the movement of the individual. At that time, this individual is included in the diversion unit it entered, at fixed intervals. The population density was recalculated, and the contours of each diversion unit were fine-tuned using the S2-2 method.

[0018] This invention proposes an air-ground collaborative method for inspecting crowds to reduce trampling risks. A drone constructs a global guidance field above, dynamically dividing the planar distribution of a large crowd into several diversion units, and forming a gap corridor with a minimum safety interval between adjacent diversion units. Multiple drones perform synchronous inspections within the gap corridor, forming yieldable soft boundaries with gestures and light signals. The inspection path is updated according to the real-time movement of the crowd, achieving low-disturbance physical segmentation of the crowd without affecting normal personnel movement, thus preventing excessive crowd gathering. During the inspection process, the robot dog monitors the gait stability of the crowd in real time. When it judges that there is a risk of trampling based on the aerial view of the drone, it will issue an audible and visual alarm and guide the movement direction of the crowd. At the same time, it will promptly detect vulnerable individuals who have fallen. The robot dog will form a temporary encirclement around them to prevent them from being trampled by the crowd and slowly guide them to the side so that they can be handed over to medical personnel. This achieves physical intervention in dangerous situations.

[0019] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0020] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A system for cooperative inspection of unmanned aerial vehicles and robot dogs, characterized in that: The system includes a drone inspection module, a robot dog inspection module, and a collaborative control module. The drone inspection module is used to survey large crowds from the air, calculate crowd density and velocity fields in real time, predict congestion evolution, and provide global guidance through sound amplification or projection. The robot dog inspection module is used to conduct close-range inspections in ground-crack corridors, detect gait stability and fallen individuals in real time, form soft boundaries for micro-diversion guidance, and physically intervene in dangerous situations. The collaborative control module is used to integrate the aerial perspective of the drone and the ground detection information of the robot dog, dynamically update the inspection path, and trigger audible and visual alarms and linkage guidance when risks are identified. 2.The UAV and robot dog cooperative inspection system of claim 1, wherein: The drone inspection module includes a visual perception unit, a crowd density calculation module, a velocity field calculation module, and a reverse flow analysis module. The visual perception unit uses a top-down camera and an infrared sensor to acquire the position of individuals in the crowd. The crowd density calculation module is used to calculate the crowd distribution density in each area based on the perception data. The velocity field calculation module is used to perform field-based modeling of the overall movement speed and direction of the crowd. The reverse flow analysis module is used to detect and evaluate the proportion of individuals moving in the reverse flow in the crowd to identify potential congestion or disturbance risks. The robot dog inspection module includes a walking control module, a gait analysis module, a dangerous individual detection module, a guidance and intervention module, and a movement and encirclement control module. The walking control module is used to control the robot dog to move stably in the gap corridor and perform inspection tasks. The gait analysis module is used to analyze the gait stability of individuals in real time to identify abnormal movements. The dangerous individual detection module is used to detect individuals who are in a state of falling or imbalance and mark their positions. The guidance and intervention module is used to form a soft boundary through posture and light signals to guide the crowd to achieve micro-diversion. The movement and encirclement control module is used to form a temporary protective circle around dangerous individuals to prevent secondary trampling. The collaborative control module includes a crowd segmentation module, a gap corridor generation module, a path update module, and a data transmission module. The crowd segmentation module is used to dynamically divide a large crowd into several relatively independent diversion units. The gap corridor generation module is used to generate corridors with safe intervals between adjacent diversion units for inspection and guidance. The path update module is used to dynamically adjust the robot dog's inspection path according to the real-time movement of the crowd and the congestion situation. The data transmission module realizes information interaction and linkage command transmission between the drone and the robot dog. 3.The UAV and robot dog cooperative inspection system of claim 2, wherein: The system's operating steps include: S1. The drone flies above the crowd and detects the crowd, generates a Cartesian coordinate map of the ground, and collects data on the location, density, velocity field and reverse flow ratio of individuals in the crowd; S2, dividing the crowd into several diversion units, after determining the contour and the number of each diversion unit, generating a fissure corridor with a safety interval between adjacent units along the contour edge as the inspection path of the robot dog; a fissure corridor with a safety interval between adjacent units along the contour edge as the inspection path of the robot dog; S3. Multiple robot dogs patrol and inspect the rift corridor. When someone blocks the robot dog's path, they use gestures and light signals to remind the crowd to avoid it. They perform gait analysis and dangerous individual detection, and promptly identify and protect individuals who fall or lose balance. S4. When the system determines that there is a risk of congestion and stampede, it triggers an audible and visual alarm and guides the crowd to divert through drone amplification, projection, robot dog gestures, and light signals to reduce local pressure. S5. Based on the data collected in real time by drones and robot dogs, the risk of trampling in each diversion unit is analyzed in real time, and the gap corridor is updated in real time in combination with the changes in the position of the crowd. 4.The UAV and robot dog cooperative inspection system of claim 3, wherein: In step S1, the data collection of individual location, density, velocity field, and reverse flow ratio within the crowd specifically includes: S1-1, marking the position of each individual in the crowd, and generating corresponding coordinates for each individual wherein the label of the individual in the crowd, marking each coordinate in the coordinate system, counting the coordinate point density of each area, tracking the position of each individual, and displaying the change of coordinate position with time in the coordinate system; S1-2. Obtain the velocity vector based on the positional changes of the individual coordinates over time. For any defined region, take the velocity vector of all individuals within that region. The averaged velocity direction is taken as the mainstream flow direction of the population in the area. When the angle between the velocity vector of an individual and the mainstream direction exceeds a set threshold, the population dynamics are considered to be in the mainstream direction. When an individual is identified as a reverse-flowing individual, the proportion of reverse-flowing individuals to the total number of individuals in the region is calculated as the reverse-flowing proportion.

5. The UAV and robot dog collaborative inspection system according to claim 4, characterized in that: In S2, dividing the crowd into several diversion units specifically involves: S2-1. If there are obstacles on the ground, use the outline of the obstacles as the boundary line of the diversion unit; if there are no obstacles on the ground, set a threshold for the number of people. As a basis for division, when the total number of people is At that time, it was divided into Each part, and will Round up to get , As the initial number of diversion units, the number of people in each diversion unit is roughly divided. S2-2, The drone scan obtains the overall coverage area of ​​the crowd. Connect the coordinates of the edge individuals within the diversion unit to obtain the initial contour, and then calculate the area of ​​the initial contour of the diversion unit. and the initial number of people To calculate population density, if If the population density in this diversion unit is high, then... To determine the density threshold, the outline size and number of people in this diversion unit need to be reduced, and individuals in the edge region with the highest population density within this diversion unit should be removed. If the population density of this subdivision unit is low, this subdivision unit will accommodate individuals removed from other subdivision units and generate new profiles. This process will continue until the population density of each subdivision unit approaches a certain level. Number of people in each diversion unit and outline area It can be determined.

6. The UAV and robot dog collaborative inspection system according to claim 5, characterized in that: In S3, the specific steps for performing gait analysis and high-risk individual detection are as follows: S3-1. Scan the leg movements of all individuals in the crowd and record the frequency of each individual's leg movements. When a scan detects... When the frequency of leg movements of more than one individual changes abruptly, and due to the difficulty in direct location, the lateral angle between the individual experiencing the frequency change and the current robot dog is calculated. This is then done by scanning all individuals at this lateral angle using a drone from above. When multiple individuals within the group are simultaneously scanned... When all individual coordinates are of individuals flowing in the opposite direction, it is determined that there is a risk of stampede in the current diversion unit. S3-2. The robot dog identifies the legs and torso of the human body. When a person's torso is found within the leg height range of a crowd during scanning, it determines that the person has fallen. The robot dog issues a warning and shoos away the crowd ahead, while simultaneously adjusting its patrol path to temporarily deviate from the crevice corridor. The drone then uses this location to pinpoint the coordinates of the fallen person, with a radius of [missing information]. All robot dog inspection paths within the area change to the direction towards this coordinate. When they reach the person who has fallen, the robot dog moves along the radius... Conduct a circular inspection until maintenance personnel arrive.

7. The UAV and robot dog collaborative inspection system according to claim 6, characterized in that: In step S4, when a risk of congestion and stampede is determined, the proportion of reverse-flow individuals in a diversion unit is calculated, i.e., the number of reverse-flow individuals in a certain diversion unit. Per capita proportion The higher the value, and the higher the absolute value of the velocity vector of all individuals in the splitting unit. The larger the value, the greater the risk of a stampede. The larger it is, the more complex the calculation formula is. The decibel level of the drone's amplification and the brightness of the light signals were both related to It is directly proportional.

8. The UAV and robot dog collaborative inspection system according to claim 7, characterized in that: In step S5, the outline of each diversion unit is updated according to the overall movement direction of the crowd, thereby adjusting the inspection path in real time. When an individual leaves the original diversion unit and the individual's distance from the outline of the original diversion unit exceeds a distance threshold, the system will detect the individual leaving the original diversion unit. At that time, this individual is included in the diversion unit it entered, at fixed intervals. The population density was recalculated, and the contours of each diversion unit were fine-tuned using the S2-2 method.

Images