Rescue method and system based on unmanned aerial vehicle, electronic equipment and storage medium

Abstract

The invention discloses a rescue method and system based on an unmanned aerial vehicle, electronic equipment and a storage medium, and relates to the field of unmanned aerial vehicle water lifesaving. The method comprises the steps that a distress signal sent by image processing equipment is received, the image processing equipment is arranged in a water area, and the distress signal is generated when the image processing equipment recognizes a drowning person; calculating a position area to which the drowning person belongs according to the distress signal, planning a path according to the position area, and controlling the unmanned aerial vehicle to travel according to the path; after it is detected that the unmanned aerial vehicle arrives at the position area, state information of each drowning person is collected through the unmanned aerial vehicle; determining the rescue priority of each drowning person and a corresponding rescue tool according to the state information of each drowning person; the unmanned aerial vehicle is controlled to arrive above the position of the drowning person with the highest rescue priority, and the unmanned aerial vehicle is controlled to put a corresponding rescue tool; when it is determined that the drowning person makes contact with the rescue tool, the unmanned aerial vehicle is controlled to put a traction rope for rescue. The rescue efficiency can be improved.

Description

Technical Field

[0001] The present application relates to the field of drone water rescue, and in particular to a drone-based rescue method, system, electronic equipment and storage medium. Background Art

[0002] Drowning deaths have become a serious public health issue worldwide. According to the World Health Organization, approximately 236,000 people die from drowning each year worldwide, with children and adolescents accounting for a disproportionately high proportion. In China, drowning has long been the leading cause of accidental death among children aged 0-14. The drowning rate in rural areas is approximately three times that in urban areas, with over 100 drowning deaths occurring in a single month during peak summer months. Furthermore, the number of adults drowning in accidents such as fishing and water activities is also on the rise, creating an increasingly urgent need for drowning prevention monitoring and rescue efforts.

[0003] In the existing technology, cameras are usually deployed in water areas to obtain abnormal information of the water area, and technicians are relied upon to identify whether the person is drowning. If so, the technicians dispatch drones to carry out rescue in a fixed manner.

[0004] However, once multiple people are drowning, rescuing them in a fixed manner will affect the overall rescue efficiency. Summary of the Invention

[0005] In order to improve rescue efficiency, the present application provides a drone-based rescue method, system, electronic device and storage medium.

[0006] The first aspect of the present application provides a rescue method based on a drone, specifically comprising: receiving a distress signal sent by an image processing device, the image processing device being deployed in a water area, the distress signal being generated by the image processing device when the image processing device identifies a drowning person; Calculating the location of the drowning person based on the distress signal, planning a path based on the location, and controlling the drone to travel along the path; After detecting that the drone has arrived at the location area, the drone is used to collect status information of each drowning person; the rescue priority of each drowning person and the corresponding rescue tool are determined based on the status information of each drowning person; the drone is controlled to arrive above the location of the drowning person with the highest rescue priority, and the drone is controlled to release the corresponding rescue tool; When it is determined that the drowning person has come into contact with the rescue tool, the drone is controlled to release a traction rope for rescue.

[0007] By adopting the above technical solution, when the image processing device identifies a drowning person in the water, it generates a distress signal and transmits it to the control center. After receiving the distress signal, the control center calculates the drowning person's location area and then plans a path based on the location area. The control center controls the drone to travel along the path. After the drone arrives at the location area, it collects the drowning person's status information and transmits this status information to the control center. The control center determines the priority of the drowning person and the appropriate rescue tool based on the status information. The control center controls the drone to fly over the highest-priority drowning person and releases the corresponding rescue tool. After the control center determines that the drowning person has encountered the rescue tool, it controls the drone to release a tow rope, which connects to the rescue tool to complete the rescue. The image processing device automatically processes the abnormal water data to generate a distress signal, shortening the response time of the distress signal. The drone travels along the planned path, improving the drone's arrival efficiency. The control center determines the priority of the drowning person and the appropriate rescue tool based on the status information, maximizing rescue efficiency, thereby improving overall rescue efficiency.

[0008] Optionally, the first data is obtained by performing a superposition calculation on the initial coordinates and the water velocity vector in the distress signal; Combining the first data with a timestamp in the distress signal, and using a drift model to predict, to obtain the location area; The path is planned based on the location area and the current location of the drone, and the drone is controlled to travel along the path.

[0009] By adopting the above technical solution, the control center constructs a dynamic drift prediction model by fusing the initial coordinates, water velocity vector and timestamp data, which significantly improves the tracking accuracy of the real-time location of the drowning person and the response efficiency of the drone rescue. It can dynamically plan the optimal path based on the relative relationship between the real-time location of the drone and the predicted area. By driving according to the planned path, it can reach the location area in a short time and can also efficiently capture the drowning person.

[0010] Optionally, obtaining images taken by the drone in the location area, and obtaining behavioral characteristics of the drowning person through the images; Obtaining a thermal image scanned by the drone in the location area to obtain the body temperature of the drowning person; Based on the high-frequency electromagnetic waves emitted by the millimeter-wave radar carried by the drone, the chest fluctuation characteristics of the drowning person are captured, and the respiratory frequency is obtained according to the chest fluctuation characteristics; The behavioral characteristics, respiratory rate, and body temperature of the drowning person are integrated to obtain the status information of each drowning person.

[0011] By adopting the above technical solutions and through the deep integration of multimodal sensor data, a three-dimensional assessment system for the life status of drowning people was constructed, which significantly improved the comprehensiveness and reliability of status information acquisition.

[0012] Scoring the behavioral characteristics, the respiratory rate, and the body temperature in the status information according to a preset standard range to obtain a behavioral characteristic score, a respiratory rate score, and a body temperature score of the drowning person; Substituting the behavioral characteristic score, the respiratory rate score, and the body temperature score into a priority score calculation formula to calculate a rescue priority score for each drowning person; According to the rescue priority score of each drowning person, the rescue priority score is inversely proportional to the rescue priority, and the rescue priority of each drowning person is obtained; The rescue tool is determined based on the position distribution data of the drowning person, where the position distribution data is obtained by taking an image of the position area by the drone.

[0013] By adopting the above technical solutions and a standardized scoring system, the rescue priority of drowning people can be determined, the decision-making delay can be shortened, and the corresponding rescue tools can be determined through the location distribution data of drowning people, so as to achieve accurate matching of rescue resources and improve the rescue success rate.

[0014] Optionally, the priority score calculation formula is Among them, p i is the rescue priority score of the i-th drowning person, and the indicators include j items, namely the action characteristics, the breathing rate, and the body temperature; w j is the weight of the jth indicator; S ij is the j-th indicator score of the i-th drowning person; is the best score of the jth indicator; i is the state deterioration rate of the i-th person who fell into the water, and the smaller the value, the slower the state change; t is the drowning time.

[0015] By adopting the above technical solution, the scores of the three indicators of behavioral characteristics, respiratory rate, and body temperature in the status information are weighted. This formula takes into account the scores of the three indicators and the weight coefficients of each indicator, ensuring that the weight coefficients can be adjusted according to actual conditions, thereby ensuring that the real-time scoring is more accurate and improving the reliability of rescue.

[0016] Optionally, the drowning person is comforted by the interactive function of the drone and is guided to use the rescue tool.

[0017] By adopting the above technical solution and the real-time interactive function of the drone, the panic of drowning people can be alleviated and the success rate of rescue can be improved.

[0018] Optionally, based on the built-in rescue sensor, the pressure parameter changes of the rescue tool are obtained, the magnetic interface of the rescue tool is activated, and based on the aerial video stream transmitted back by the drone, it is determined that the drowning person has come into contact with the rescue tool, the drone is controlled to release the traction rope, and the magnetic lock on the traction rope is released to connect the traction rope to the rescue tool.

[0019] By adopting the above technical solution, the pressure sensor on the rescue tool can provide real-time feedback on the drowning person's grasping of the rescue tool. Then, based on the aerial video stream sent back by the drone, it is confirmed again whether the drowning person has grasped the rescue tool. Then the traction rope is connected to the rescue tool and the rescue is carried out. The double inspection makes the rescue more accurate. The entire process is automated, shortening the response time.

[0020] In a second aspect of the present application, a rescue system based on a drone is provided, specifically comprising: a signal receiving module, configured to receive a distress signal sent by an image processing device, the image processing device being deployed in a water area, the distress signal being generated by the image processing device when it identifies a drowning person; A positioning and path planning module is used to calculate the location area of ​​the drowning person based on the distress signal, plan a path based on the location area, and control the drone to travel along the path; A status information collection module is configured to collect status information of each drowning person through the drone after detecting that the drone has arrived at the location area; A rescue decision module is used to determine the rescue priority of each drowning person and the corresponding rescue tools according to the status information of each drowning person; A rescue tool delivery module is used to control the drone to arrive above the location of the drowning person with the highest rescue priority, and control the drone to deliver the corresponding rescue tool; The traction rescue module is used to control the UAV to release the traction rope for rescue when it is determined that the drowning person has come into contact with the rescue tool.

[0021] By adopting the above technical solution, when the image processing device identifies a drowning person in the water, it generates a distress signal and transmits it to the control center. After receiving the distress signal, the control center calculates the location area of ​​the drowning person and then plans a path based on the location area. The control center controls the drone to travel along the path. After the drone arrives at the location area, it collects the status information of the drowning person and transmits the status information to the control center. The control center determines the priority of the drowning person and the appropriate rescue tool based on the status information, controls the drone to fly over the highest-priority drowning person and releases the corresponding rescue tool. After the control center determines that the drowning person has encountered the rescue tool, it controls the drone to release a tow rope, which connects to the rescue tool through the tow rope to complete the rescue. In this solution, the image processing device automatically processes abnormal data in the water area to generate a distress signal, shortening the response time of the distress signal. The drone travels according to the planned path, improving the efficiency of the drone's arrival. The control center determines the priority of the drowning person and the appropriate rescue tool based on the status information, maximizing the rescue efficiency, thereby improving the overall rescue efficiency.

[0022] In the third aspect of the present application, an electronic device is provided, including a processor 701, a memory 705, a user interface 703 and a network interface 704, wherein the memory 705 is used to store instructions, the user interface 703 and the network interface 704 are both used to communicate with other devices, and the processor 701 is used to execute the instructions stored in the memory 705 so that the electronic device executes any of the methods described above.

[0023] In a fourth aspect of the present application, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores instructions, and when the instructions are executed, any one of the methods described above is executed. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 This is a schematic diagram of the architecture of a drone-based rescue method disclosed in an embodiment of the present application; Figure 2 This is a flow chart of a rescue method based on a drone disclosed in an embodiment of the present application; Figure 3 yes Figure 2 A schematic flow chart of a sub-step of step S102; Figure 4 yes Figure 2 A schematic flow chart of a sub-step of step S103; Figure 5 yes Figure 2 A schematic flow chart of a sub-step of step S104; Figure 6 This is a module diagram of a drone-based rescue system provided in an embodiment of the present application; Figure 7 This is a structural diagram of an electronic device disclosed in an embodiment of the present application.

[0025] Explanation of the accompanying symbols: 20. UAV-based rescue system; 21. Signal receiving module; 22. Positioning and path planning module; 23. Status information acquisition module; 24. Rescue decision module; 25. Rescue tool delivery module; 26. Towing rescue module; 701. Processor; 702. Communication path; 703. User interface; 704. Network interface; 705. Memory. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments.

[0027] In the description of the embodiments of this application, words such as "for example" or "for instance" are used to indicate examples, illustrations, or explanations. Any embodiment or design described as "for example" or "for instance" in the embodiments of this application should not be construed as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "for example" or "for instance" is intended to present the relevant concepts in a concrete manner.

[0028] In the description of the embodiments of the present application, the term "multiple" means two or more. For example, multiple systems refer to two or more systems, and multiple screen terminals refer to two or more screen terminals. In addition, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the indicated technical features. Thus, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. The terms "including", "comprising", "having" and their variations all mean "including but not limited to", unless otherwise specifically emphasized.

[0029] Figure 1 An exemplary system architecture 10 is shown to which an embodiment of a drone-based rescue method or a drone-based rescue system of the present application can be applied.

[0030] like Figure 1 As shown, system architecture 10 may include terminal devices 11 and 12, a control center 14, and a network 13. Network 13 is a medium for providing a communication link between terminal devices 11 and 12 and control center 14. Network 13 may include various pathless communication methods, such as long- and short-range pathless communication, wide area communication, and satellite communication.

[0031] Users can use terminal devices 11 and 12 to interact with the control center 14 through the network 13 to receive or send data, etc. Various communication client applications can be installed on the terminal devices 11 and 12, such as model training applications, video recognition applications, etc.

[0032] Terminal device 11 is a drone, which may also be equipped with a video capture device. The video capture device can be any device capable of capturing video, such as a camera, sensor, etc. Users can use the video capture device of terminal 11 to capture video, and terminal 12 can also perform preliminary processing on the captured video.

[0033] The terminal device 12 is hardware, and may be an electronic device with image acquisition and processing capabilities, including but not limited to a video surveillance platform, a camera, a visual detector, etc.

[0034] The control center 14 may be a server that provides various services, such as a background server that processes data displayed on the terminal devices 11 and 12. The background server may analyze and process the received data, and may feed back the processing results (such as recognition results) to the terminal devices.

[0035] It should be noted that the server can be either hardware or software. When the server is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the server is software, it can be implemented as multiple software or software modules (e.g., multiple software or software modules used to provide distributed services), or it can be implemented as a single software or software module. No specific limitations are given here.

[0036] It should be understood that Figure 1 The number of terminal devices, networks, and servers in the above description is merely illustrative. Any number of terminal devices, networks, and servers may be used as needed. In particular, if target data does not need to be acquired remotely, the above system architecture may not include a network, but may instead include only terminal devices or servers.

[0037] This application provides a rescue method based on drones, referring to Figure 2 , Figure 2 : This is a flowchart of a drone-based rescue method provided in an embodiment of the present application, including steps S101 to S106. The above steps are as follows: S101: receiving a distress signal sent by an image processing device, where the image processing device is deployed in a water area and generates the distress signal when the image processing device identifies a drowning person.

[0038] Specifically, in the embodiment of the present application, the water area may be a natural open water area, or it may be a city public water area, a special high-risk water area, or other place where drowning may occur.

[0039] It is understood that the image processing device communicates with the control center. The image processing device can be a multi-spectral intelligent monitoring camera, or a millimeter radar array, a surface acoustic wave sensor array, or other electronic device that can acquire water area image data and perform data processing.

[0040] In one feasible implementation, the image processing equipment is deployed on a buoy in the water area, or on a cliff, dam or other suitable location, or the image processing equipment can be dynamically deployed on a drone in a preset water area.

[0041] Specifically, in an embodiment of the present application, image processing devices are deployed at preset intervals in the water. Using a pre-trained deep learning model, they perform real-time posture analysis of surface targets. When abnormal behavior (such as persistent sinking or erratic limb movement) is detected, a distress signal containing the initial coordinates of the drowning person and the water velocity vector is automatically generated from the captured image and transmitted to the control center. The deep learning model is a technical system that automatically identifies characteristic human drowning movements (such as sinking and struggling) by analyzing surface surveillance video.

[0042] For example, multispectral cameras are deployed in water areas every 250 meters. A multispectral camera deployed in a certain water area has longitude and latitude coordinates of 118.70 degrees east longitude and 24.43 degrees north latitude, a vertical height of 15.3 meters, a lens pitch angle of -12 degrees, and a horizontal deflection angle of 93 degrees. When the deep learning model built into the multispectral camera detects a drowning person, and the drowning person is located at the center of the image (x=640, y=320), the system calculates through a geometric coordinate conversion program that the horizontal distance between the drowning person and the multispectral camera is approximately 72.3 meters, and the actual diameter distance is approximately 74 meters. Combined with a horizontal deflection angle of 93 degrees (that is, the lens is facing 3 degrees south of due east), the system calculates that the drowning person has an eastward displacement of approximately 72.2 meters and a southward displacement of 3.8 meters in the southeast direction. Taking into account the influence of the earth's curvature, these displacements are converted into an increase of 0.000725 degrees in longitude and a decrease of 0.000034 degrees in latitude, respectively. After superimposing the original coordinates from the multispectral camera, the final output of the drowning person's initial coordinates was 118.700725 degrees east longitude and 24.429966 degrees north latitude. Furthermore, by capturing dynamic images of the water surface and combining them with time differences and position transformations, the camera calculated the water velocity vector to be 30 degrees north-east, with a water velocity of 0.8 m / s.

[0043] Finally, the camera integrates the initial coordinates of the drowning person and the water velocity vector to generate a distress signal and transmit it to the control center.

[0044] S102: Calculate the location area of ​​the drowning person based on the distress signal, plan a path based on the location area, and control the drone to travel along the path.

[0045] Specifically, in the embodiment of the present application, the control center will calculate the displacement according to the initial coordinates and water velocity vector in the distress signal through Newton's law of motion to correct the initial coordinates.

[0046] Using the corrected coordinates as the center of the ellipse, the major axis is aligned with the direction of the water flow, defining an elliptical location area. The length of the ellipse's major axis is determined by the maximum possible drift under the influence of the water flow. This is calculated using the formula: major axis = water flow velocity × transmission time + initial error + velocity error, taking into account the transmission time, water flow velocity, and possible velocity error. The minor axis is a preset length. This location area represents the range where a drowning person might be located.

[0047] The drone uses its current location as its starting point and the area within it as its destination to automatically generate a radial navigation path based on electronic chart data. The drone will cruise along this path at high speed while simultaneously scanning the preset area ahead in real time using its front-facing camera. If it detects drifting objects such as moving vessels or floating debris, the drone will automatically trigger its dynamic obstacle avoidance program and correct its path.

[0048] For example, a distress signal generated in a certain area had initial coordinates of 118.700725° East longitude and 24.429966° North latitude. The current velocity vector was 30° North-East, the current velocity was 0.8 m / s, and the transmission time was 1.2 seconds. The control center calculated the displacement to be 0.96 meters, including 0.48 meters to the east and 0.831 meters to the north. The corrected initial coordinates were approximately (118.700730° East longitude, 24.429974° North latitude).

[0049] Centered on the corrected coordinates, the main axis is aligned with the direction of the water flow to determine the elliptical positioning area. The initial error can be set to 3 meters, representing the uncertainty of the system's initial positioning. The flow rate error can be estimated based on actual conditions and is set to 1.24 meters in this scenario. Substituting these values ​​into the formula, the major axis length is 0.96 meters + 3 meters + 1.24 meters = 5.2 meters. The major axis is 5.2 meters, and the minor axis is 3 meters.

[0050] The drone automatically generates a diameter navigation path and cruises at high speed to the target point. During the cruise, the drone uses its front-facing camera to scan a 200-meter area in real time. If it detects drifting objects, it automatically triggers a dynamic obstacle avoidance program and corrects its path. When it approaches the target area within 500 meters, the drone automatically slows down to 5 m / s and descends to 10 meters above the water, ultimately approaching the target point.

[0051] Reference Figure 3 , Figure 3 This embodiment of the present application provides Figure 2 A schematic flow chart of a sub-step of step S102 is shown in FIG. 1 , wherein the above steps are as follows: S201: Perform superposition calculation on the initial coordinates and the water velocity vector in the distress signal, and obtain first data in combination with the timestamp.

[0052] Specifically, in this embodiment, the system first decomposes the water flow velocity vector. Using trigonometric calculations, the eastward velocity component is equal to the water flow velocity multiplied by the cosine of the angle between the water flow direction and the due east direction. The northward velocity component is equal to the water flow velocity multiplied by the sine of the angle between the water flow direction and the due north direction. The same applies to other directions. Then, based on the time elapsed since the drowning occurred, the system multiplies the velocity component by time to calculate the displacement of the drowning person after being driven by the water flow. Finally, the displacement is converted into changes in longitude and latitude, superimposed on the initial coordinates, and the new coordinates of the drowning person, i.e., the first data, are obtained as the initial predicted coordinates.

[0053] For example, the initial coordinates of the drowning person received are 118.75° east longitude and 24.45° north latitude, the water flow velocity is 0.8m / s, and the direction is 30° north-east. The system first decomposes the water flow velocity into two components, eastward and northward. Through trigonometric calculations, the eastward velocity component is 0.8m / s multiplied by cos30°, which is approximately equal to 0.4m / s, and the northward velocity component is 0.8m / s multiplied by sin30°, which is equal to 0.4m / s. However, according to more precise calculations, the northward component is actually 0.69m / s. Combined with the 180 seconds that have passed since the drowning occurred, the system calculates that the drowning person was moved 72 meters east and 124 meters north by the water flow. These displacements are converted into changes in longitude and latitude, superimposed on the initial coordinates, and the first data, namely 118.701436° east longitude and 24.431090° north latitude, is obtained as the preliminary predicted coordinates.

[0054] S202: According to the first data, a drift model is used to predict and obtain a location area.

[0055] Specifically, in this embodiment, the control center triggers the drift model according to the first data. The drift model constructs a three-dimensional flow field to simulate the motion trajectory of the target object by integrating real-time hydrological and meteorological data with an adaptive particle tracking algorithm. First, the real-time data streams such as the current water depth distribution, surface water temperature, wind speed and direction are connected, and the drowning person is abstracted as a dynamic particle affected by multiple environmental forces. The algorithm generates thousands of particles per second, and its movement is driven by the mainstream velocity, while superimposing the surface water flow deflection caused by the wind speed (such as the southwest wind changes the direction of the water flow through wind stress) and random turbulent disturbances. Body shape features are converted into fluid resistance parameters after image recognition. For example, a medium body shape corresponds to a projected area of ​​0.7m.2 When the particle's trajectory is predicted to hit an obstacle, the algorithm triggers a water detour mechanism, correcting the particle's motion path in real time.

[0056] After simulating particle diffusion for a preset time, the system converts the spatial distribution of tens of thousands of trajectories into a probability cloud map. Using kernel density analysis, high-probability clusters are identified as location regions. These are derived from the drift model's further analysis of the initial data.

[0057] For example, after the control center receives a drowning alarm at the initial location (118.701436° east longitude, 24.431090° north latitude), the system will immediately retrieve the real-time water depth (assuming it is 5 meters), surface water temperature (28°C) and southwest wind 3.4m / s data for the area. The algorithm generates 5,000 virtual particles per second. Each particle not only drifts northeastward with the mainstream velocity of 0.8m / s, but also has an offset of 0.2m / s to the east due to the wind speed, superimposed with the ±0.1m / s lateral disturbance caused by random turbulence. If image recognition determines that the drowning person is of medium size, the water flow resistance coefficient of the particle will be adjusted from the default 0.6 to 0.75, reducing the lateral diffusion rate by 15%. After 180 seconds of simulation, the system found through kernel density analysis that an elliptical high-probability aggregation area with a long axis of 40 meters and a short axis of 25 meters was formed 400 meters downstream of the initial location (confidence level 85%). If the wind speed suddenly increases to 5m / s at this time, the dynamic feedback mechanism will recalculate the flow field within 3 seconds. The predicted area will be shifted 80 meters to the east and the coverage will be expanded by 20%. The search and rescue priority area will be updated in real time to obtain the location area. The location area is the area obtained by further judgment of the first data based on the drift model.

[0058] S203: Planning a path based on the location area and the current location of the drone, and controlling the drone to travel along the path.

[0059] Specifically, in this embodiment, the control center implements spatial coordinate conversion based on UTM projection technology. UTM is a method of converting spherical coordinates into plane coordinates, mapping the latitude and longitude information of the earth's curved surface into a plane rectangular coordinate system, and constructing a three-dimensional environmental model on this basis. The model uses a vertical layering strategy to divide different height layers (such as 0-10 meters, 10-30 meters, 30-100 meters), integrating multi-source data including wind speed gradients at each layer, real-time ship dynamic trajectories, and static obstacle coordinates such as reefs to form a three-dimensional navigation grid containing spatial constraints and dynamic obstacles. The coordinates of the center of the location area and the coordinates of the drone base station are converted into plane rectangular coordinates to plan the path. The path planning adopts the A* algorithm, and the search path is expanded in eight directions in the neighborhood from the current position as the starting point. Each time, the direction with the smallest sum of the current cumulative distance and the heuristic estimated distance is selected to advance until the target point is reached.

[0060] For example, when the target area is centered at 118.7045°E and 24.4342°N, the coordinates of the drone base station are converted to the target area (X: 672,710 meters, Y: 2,708,550 meters) and the drone base station (X: 325,200 meters, Y: 2,708,200 meters). The reefs are marked as red grid areas on the 3D map, representing impassable, infinite-cost obstacles. The control center generates a northeast-diameter path, but after detecting the distribution of reefs in the middle section, the A* algorithm automatically adjusts the path: first offsetting 200 meters east to avoid the obstacle area, then turning northwest to approach the target point, ultimately generating the shortest feasible path after the detour. This path planning process dynamically integrates real-time ship trajectory changes to ensure that the path avoids moving obstacles throughout.

[0061] S103: After detecting that the drone has arrived at the location area, the status information of each drowning person is collected through the drone.

[0062] Specifically, in an embodiment of the present application, after the control center detects that the GPS carried by the drone is located within the location area, it activates the camera equipment carried by the drone to shoot video and receives the shot video transmitted by the drone.

[0063] For example, when the control center detects that the drone has reached the location area, it activates the drone's onboard optical zoom camera, which shoots 4K ultra-high-definition video at 60 frames per second and transmits the video to the control center.

[0064] The artificial intelligence algorithm built into the control center can analyze human posture frame by frame - by identifying the spatial coordinate changes of 17 key bone points (such as shoulders, elbows, hip joints, etc.), calculating the limb swing frequency and the degree of chaos of the movement trajectory.

[0065] For example, the arm swing frequency of a conscious drowning person is typically between 1-3 Hz and tends to be shoreward, while the limb movement frequency of a comatose drowning person is less than 0.5 Hz and is irregularly distributed. State information includes the drowning person's limb swing frequency and movement trajectory.

[0066] Reference Figure 4 , Figure 4 This embodiment of the present application provides Figure 2 A schematic flow chart of a sub-step of step S103 in FIG. 1 is shown in FIG. 1 , wherein the steps are as follows: S301: Acquire images taken by the drone in the location area, and obtain behavioral characteristics of the drowning person through the images.

[0067] Specifically, in this embodiment, the control center uses a drone-mounted camera to capture images of the water surface, which are then transmitted to the control center via wireless communication. The control center extracts 2D coordinates from a 15-frame (approximately 0.5-second) image sequence, overlaying the time dimension to form a 3D motion trajectory, accurately locating the spatial displacement of key nodes such as the elbow and shoulder.

[0068] For example, the control center uses a high-definition camera mounted on a drone to collect water surface images in real time, and transmits the original images with a resolution of 3840×2160 to the control center via a 5G link. The control center extracts two-dimensional plane coordinates from a sequence of 15 consecutive frames (about 0.5 seconds) in the image. If the target hand has a vertical displacement of more than 30 cm within 0.5 seconds and the movement frequency reaches 2 times per second, the system immediately determines it as regular waving behavior. If the angle between the main axis of the body and the horizontal plane exceeds 45 degrees for more than 1 second, a body imbalance warning is triggered. These quantitative thresholds are trained based on a large number of drowning behavior samples and can effectively distinguish normal swimming movements from distressed struggling postures. By capturing at a high frame rate of 30 frames per second, the system can accurately analyze subtle movements at the finger level. Combined with three-dimensional spatiotemporal analysis technology, it can achieve high-precision behavior recognition in complex water surface environments.

[0069] S302: Obtain a thermal image scanned by the drone in the location area to obtain the body temperature of the drowning person.

[0070] Specifically, in this embodiment of the present application, a control center controls a drone-mounted thermal imaging temperature detection system to accurately measure human body temperature in complex water environments. The system's detector surface receives far-infrared radiation emitted by a drowning person, converts it into a digital signal through analog-to-digital conversion, and then visualizes the digital signal to generate a thermal distribution map, from which the body temperature is determined.

[0071] For example, the thermal imaging temperature detection system used to control drones uses multi-level physical sensing and intelligent algorithms to accurately measure human body temperature in complex aquatic environments. The system's core component is a microbolometer array made of vanadium oxide, a material with a significant change in resistance with temperature. When far-infrared radiation (wavelength 8-14 microns) emitted by the human body is projected onto the detector surface, which consists of 16,384 independent sensing units. Each unit generates a corresponding change in resistance based on the intensity of the received thermal radiation. This change is converted into a voltage signal through a circuit and then into a 16-bit digital signal. These signals are converted into grayscale values, with the minimum value corresponding to 0 (pure black) and the maximum value corresponding to 255 (pure white). This creates a thermal distribution map with a resolution of 640×512 pixels, with each pixel corresponding to a temperature sensitivity of 0.05°C.

[0072] Specifically, to ensure measurement accuracy, a dynamic environmental compensation mechanism is integrated. Every 0.5 seconds, a meteorological sensor acquires parameters such as air velocity (which affects the body's heat dissipation rate), relative humidity (which changes heat conduction efficiency), and ambient temperature (which determines the heat exchange difference), and then corrects the thermal imaging data in real time. For example, at a wind speed of 3m / s, the temperature of the exposed skin area is automatically adjusted upward by 0.8°C to compensate for the wind chill effect.

[0073] S303: Based on the high-frequency electromagnetic waves emitted by the millimeter-wave radar carried by the drone, the chest fluctuation characteristics of the drowning person are captured, and the breathing frequency is obtained based on the chest fluctuation characteristics.

[0074] Specifically, in an embodiment of the present application, the control center controls the radar equipment carried by the drone to emit electromagnetic waves, and captures the displacement of the chest surface through electromagnetic waves. The phase of the reflected wave will change regularly with the rise and fall of the chest, and the breathing frequency is obtained according to the capture time.

[0075] For example, the millimeter-wave radar onboard the drone emits high-frequency electromagnetic waves in the 60-64 GHz frequency band. This frequency band has millimeter-level wavelength characteristics, making it sensitive to tiny displacements of the chest surface. When these electromagnetic waves encounter the drowning person's body, the phase of the reflected wave changes regularly with the rise and fall of the chest. During inhalation, the chest cavity expands, increasing the reflection path by approximately 0.5 mm, while during exhalation, it shortens accordingly. The frequency of this change directly corresponds to the breathing rate.

[0076] S304: Integrate the behavioral characteristics, respiratory rate, and body temperature of the drowning person to obtain status information of each drowning person.

[0077] Specifically, in this embodiment of the present application, the behavioral characteristics of the drowning person are used as reference coordinates, and the respiratory rate and body temperature information are associated through a feature point matching algorithm to obtain the drowning person's state information. The state information is the association data of the behavioral characteristics, respiratory rate, and body temperature.

[0078] For example, using an image captured by a high-definition camera as the reference coordinate system, when the camera detects the hand movement of a drowning person at pixel (x=120, y=80) in the image, the corresponding temperature value of 31.5°C at (x'=118, y'=82) on the thermal imager is automatically correlated with the peak interval of 2.14 seconds (corresponding to 28 breaths / minute) in the radar's respiratory waveform at that spatial coordinate. Status information is a combination of each drowning person's behavioral characteristics, body temperature, and respiratory rate.

[0079] S104: Determine the rescue priority and corresponding rescue tools for each drowning person according to the status information of each drowning person.

[0080] Specifically, in an embodiment of the present application, the control center prioritizes the drowning persons according to the behavioral characteristics in the drowning persons' status information and determines the appropriate rescue tools.

[0081] For example, when the optical zoom camera onboard a drone detects a sudden drop in limb frequency (such as decaying from 2.5Hz to 0.3Hz within 3 seconds), it automatically triggers a critical condition warning and raises the rescue priority of the target to the highest level.

[0082] The drone is controlled to make intelligent selections based on priority levels and movement characteristics: for the highest priority drowning people, a propulsion-type life-saving buoyancy device (such as a life buoy, a self-propelled life-saving float, etc.) is deployed, and for other priority groups, a wearable emergency buoyancy device (automatic inflatable life jacket, life-saving wristband, etc.) is deployed.

[0083] Reference Figure 5 , Figure 5 This embodiment of the present application provides Figure 2 A schematic flow chart of a sub-step of step S104 is shown in FIG. 1 , wherein the above steps are as follows: S401: Score the behavioral characteristics, respiratory rate, and body temperature in the status information according to a preset standard range to obtain a behavioral characteristic score, a respiratory rate score, and a body temperature score of the drowning person.

[0084] Specifically, in the embodiment of the present application, the normal distress behavior benchmark is set as a waving amplitude > 50 cm, a frequency > 1 time / second, and a tilt angle < 30°. If the benchmark value is exceeded, the score is calculated using a linear function. The control center scores each indicator according to the benchmark, and the indicators include behavioral characteristics, respiratory rate, and body temperature.

[0085] For example, when a waving amplitude of 30 cm (60% of the baseline value), a frequency of 0.5 times / second (50% of the baseline value), and an inclination angle of 45° (exceeding 50% of the baseline) are detected, the behavior score = (0.6+0.5) / 2×(1-0.5)×100=27.5 points.

[0086] The control center sets a respiratory rate of 16 breaths per minute as the baseline and establishes a segmented scoring function. A measured value between 12 and 20 breaths per minute (normal range) is scored 100 points, with 25 points deducted for every ±5 breaths outside the range. For example, if the measured value is 28 breaths per minute (12 breaths above the baseline), the respiratory score is calculated as 100 - (28 - 20) / 5 × 25 = 60 points.

[0087] The control center sets a dynamic temperature scoring curve based on a core temperature of 36.5°C, using a sigmoid function: Temperature score = 100 / (1 + e^(-0.5 × (T-34))). For a temperature of 31.5°C, the score = 100 / (1 + e^(1.25)) ≈ 22.3 points. This function ensures a sharp deduction of points below 34°C, strengthening the assessment of hypothermia risk.

[0088] S402: Substitute the behavioral characteristic score, the respiratory rate score, and the body temperature score into the priority score calculation formula to calculate the rescue priority score of each drowning person.

[0089] Specifically, in the embodiment of the present application, the control center converts the behavior, respiration, and body temperature scores into a quantitative rescue sequence through a dynamic weighting algorithm. First, each score is normalized, and the behavior feature score, respiratory rate score, and body temperature score are uniformly mapped to the range of 0-1.

[0090] For example, when a drowning person scores 35 points for behavior, 60 points for breathing, and 22 points for body temperature, the standardized scores correspond to 0.35, 0.6, and 0.22 respectively.

[0091] The priority scoring formula uses a dynamic weighting mechanism, with a preset base weight of 40% for body temperature, 35% for respiration, and 25% for behavior. When any standardized value exceeds the threshold, the weighting is adaptively adjusted.

[0092] For example, if the body temperature is less than 0.3, the weighting of body temperature increases to 65%, the weighting of respiration decreases to 25%, and the weighting of behavior remains at 10%. The calculation formula is: Priority score = 0.65 × body temperature + 0.25 × respiration + 0.1 × behavior. Priority score = 0.65 × 0.22 + 0.25 × 0.6 + 0.1 × 0.35 = 0.292. When this value exceeds the preset threshold of 0.28, a Level 1 response is triggered.

[0093] The priority score calculation formula is: Among them, p i is the rescue priority score of the i-th drowning person, and the indicators include j items, namely the action characteristics, the breathing rate, and the body temperature; w j is the weight of the jth indicator; S ijis the j-th indicator score of the i-th drowning person; is the best score of the jth indicator; i is the state deterioration rate of the i-th person who fell into the water, and the smaller the value, the slower the state change; t is the drowning time.

[0094] Specifically, in this embodiment of the application, the preset basic weights are body temperature 40%, respiration 35%, and behavior 25%. When any standardized value exceeds the threshold, the weight is triggered to adjust adaptively. λi is determined by pre-stored data such as age and physical condition, and the drowning time is obtained by the time stamp.

[0095] For example, when the body temperature score is less than 0.3, the body temperature weight is increased to 65%, the respiration weight is reduced to 25%, and the behavior weight is retained at 10%. For example, λ for children is set to 0.15 / min, and λ for young adults is 0.05 / min. Two drowning victims are detected after 10 minutes of drowning: a 35-year-old male (labeled A) and an 8-year-old girl (labeled B). The control center scores the victim based on the status information collected by the drone: A's body movements show intermittent paddling (motion score 72), a respiratory rate of 27 breaths per minute (respiration score 62), and a body temperature of 35.7°C (temperature score 70). B's body movements show stillness (motion score 25), a respiratory rate of 32 breaths per minute (respiration score 38), a body temperature of 33.6°C (temperature score 45). When the drowning time reaches 12 minutes, A's priority score is calculated to be 0.464, and B's priority score is 0.105.

[0096] S403: According to the rescue priority score of each drowning person, the rescue priority score is inversely proportional to the rescue priority, and the rescue priority of each drowning person is obtained.

[0097] Specifically, in an embodiment of the present application, the control center divides priorities according to a priority score, and the priority score is inversely proportional to the priority.

[0098] For example, the priority score of drowning person A is 0.464, and the priority score of drowning person B is 0.105. The priority score is inversely proportional to the priority, so the priority of drowning person A is lower than that of drowning person B.

[0099] S404: Determine a rescue tool based on the location distribution data of the drowning person, where the location distribution data is obtained by taking an image of the location area from a drone.

[0100] Specifically, in this embodiment of the present application, a drone's onboard camera captures images of the water surface and transmits them to a control center. The control center integrates the drone's real-time GPS positioning, the flight altitude measured by a barometer, the camera's optical parameters, and three-dimensional terrain data from the water area in a geographic information system to construct a mapping relationship from image pixels to geographic coordinates. The control center converts the coordinates of the drowning person's location into a density heat map and determines the appropriate rescue tool based on the density.

[0101] For example, the control center receives images of the water surface continuously captured by the drone's high-definition camera. By integrating the drone's real-time GPS positioning (accuracy down to the centimeter level), the flight altitude measured by the barometer (e.g., 100 meters), the camera's optical parameters (e.g., focal length 35mm, viewing angle 78 degrees), and the three-dimensional terrain data of the water area in the geographic information system (including shore shape and water depth changes), a mapping relationship from image pixels to geographic coordinates is constructed. Specifically, when the drone shoots at a 45-degree angle of depression, the calculation of the ground position corresponding to each pixel involves the principle of triangulation. A spatial triangle is formed by the drone's position, shooting angle, and target point, combined with the correction for the earth's curvature (for targets more than 500 meters away), and the actual position of the drowning person is finally determined. The target at the 300th pixel in the upper left corner of the picture can be determined to be 173 meters from the shore after calculation, and the geographic coordinate error does not exceed 1.5 meters.

[0102] The control center converts the coordinates of the drowning person's location into a density heat map. The control center sets the core radius ε = 5 meters and the minimum number of neighbors MinPts = 2, and automatically identifies high-density gathering areas (such as 4 people within a radius of 3 meters) and isolated individuals. For a detected group of 5 people, the system gives priority to matching an automatic inflatable raft with a carrying capacity of 6 people rather than deploying a lifebuoy alone. In the equipment selection logic, the coverage radius of the life raft (10 meters) must be larger than the group distribution diameter (such as 7 meters), and a 20% safety redundancy is automatically added considering the wave height factor.

[0103] S105: Control the drone to arrive above the location of the drowning person with the highest rescue priority, and control the drone to release the corresponding rescue tool.

[0104] Specifically, in an embodiment of the present application, the drone is controlled to arrive above the location of the highest priority drowning person and release corresponding rescue tools.

[0105] For example, control the drone to arrive above the location of the highest priority drowning person, and stop the drone above the highest priority drowning person A. Drowning person A is an isolated individual, and choose to control the drone to drop a lifebuoy.

[0106] Furthermore, the drone in the embodiment of the present application also has a human-machine interaction function. When the control center detects that the drone has arrived over the highest-priority drowning person, it activates the drone's audio module, which plays a preset soothing voice. After the drone releases the rescue tool, the module provides voice guidance.

[0107] For example, when the control center detects a drone approaching a high-priority drowning person, it activates the drone's onboard two-way audio module, playing a short message such as "Please remain calm, rescue equipment has arrived." Simultaneously, the drone's belly-mounted circular LED array flashes alternating blue and green at a 0.5Hz frequency. This color combination is 40% more visible in open water than conventional red, providing visual stimulation to help the drowning person locate rescue tools.

[0108] After the drone drops the lifebuoy, a voice prompt will say something like “Please hold the lifebuoy with both hands” and other guiding sentences.

[0109] S106: When it is determined that the drowning person has come into contact with the rescue tool, the drone is controlled to release a traction rope for rescue.

[0110] Specifically, in an embodiment of the present application, when the rescue tool enters the water, the drone activates the traction rope release mechanism after detecting a contact signal based on the rescue tool pressure sensor, releases the traction rope, and completes the rescue.

[0111] For example, when the lifebuoy enters the water, the drone activates the traction rope release mechanism after detecting the contact signal based on the rescue tool pressure sensor. The carbon fiber rope unfolds at a uniform speed of 3m / s. At the same time, the LED strobe light on the side of the lifebuoy starts optical positioning. The control center controls the drone to fine-tune the position to facilitate the binding of the traction rope to the lifebuoy, thereby completing the rescue.

[0112] Specifically, in other embodiments, the rescue tool detects through a built-in sensor that the pressure value exceeds a threshold and automatically activates the magnetic interface of the rescue tool. At the same time, the camera carried by the drone takes a video and transmits it to the control center. The control center analyzes the video to determine that the drowning person has effectively contacted the rescue tool, and then releases a traction rope with a magnetic module to complete the rescue.

[0113] For example, when the lifebuoy detects a pressure exceeding 200 Newtons (equivalent to 1 / 4 of an adult's body weight concentrated in the grip area) for more than three seconds through its built-in multi-point pressure sensor, the magnetic interface on the edge of the lifebuoy is automatically activated. This magnetic interface is composed of 16 sets of neodymium iron boron permanent magnets and is normally kept closed by an electromagnetic locking device. When the pressure triggers the threshold, the locking ring is immediately de-energized, and the attraction between the magnets causes the interface to quickly close. At this point, the interface contact resistance drops below 0.02Ω, forming a stable circuit path.

[0114] At the same time, the drone's 4K optical zoom camera (focal length 24-480mm) transmits a stream of aerial video at 60 frames per second to the control center. Using a video analysis system and an improved YOLOv5 algorithm, the control center uses a training set of 200,000 images of drowning victims' movements to identify three key contact features in real time: the contact area ratio between the palm and the lifebuoy handle (must be >75%), the change in the lifebuoy's draft diameter (sinking amount >8cm), and the angle between the body's longitudinal axis and the lifebuoy's center diameter (<30°). When these three conditions are met simultaneously for more than two seconds, the system determines that contact is valid.

[0115] After confirming effective contact, the drone releases the carbon fiber traction rope (6mm diameter, 5000N breaking strength) inside the device. The magnetic lock at the end of the traction rope adopts a dual-redundant design: the main lock is fixed by a 3mm diameter electromagnetic pin, and the backup lock is locked by a shape memory alloy (Nitinol) spring. When the release command is received, a 5A reverse current is passed through the electromagnetic diameter ring of the main lock to demagnetize it, and the electromagnetic pin retracts within 0.1 seconds; the backup lock is heated to 60°C through a resistor, triggering the deformation of the shape memory alloy to unlock. The dual unlocking mechanism ensures reliable release even in harsh conditions such as waves as high as 2 meters and wind speeds of 12m / s.

[0116] After the rescue is completed, the camera carried by the drone will be activated for inspection during the process of controlling the drone to return.

[0117] By adopting the above technical solution, when the image processing device identifies a drowning person in the water, it generates a distress signal and transmits it to the control center. After receiving the distress signal, the control center calculates the location area of ​​the drowning person and then plans a path based on the location area. The control center controls the drone to travel along the path. After the drone arrives at the location area, it collects the status information of the drowning person and transmits the status information to the control center. The control center determines the priority of the drowning person and the appropriate rescue tool based on the status information, controls the drone to fly over the highest-priority drowning person and releases the corresponding rescue tool. After the control center determines that the drowning person has encountered the rescue tool, it controls the drone to release a tow rope, which connects to the rescue tool through the tow rope to complete the rescue. In this solution, the image processing device automatically processes abnormal data in the water area to generate a distress signal, shortening the response time of the distress signal. The drone travels according to the planned path, improving the efficiency of the drone's arrival. The control center determines the priority of the drowning person and the appropriate rescue tool based on the status information, maximizing the rescue efficiency, thereby improving the overall rescue efficiency.

[0118] refer to Figure 6 This application also provides a drone-based rescue system 20. Specifically, it includes: a signal receiving module 21 for receiving a distress signal sent by an image processing device, the image processing device being deployed in a water area, the distress signal being generated by the image processing device when it identifies a drowning person; The positioning and path planning module 22 is used to calculate the location area of ​​the drowning person according to the distress signal, plan a path according to the location area, and control the UAV to travel along the path; A status information collection module 23 is configured to collect status information of each drowning person through the drone after detecting that the drone has arrived at the location area; The rescue decision module 24 determines the rescue priority and corresponding rescue tools for each drowning person according to the status information of each drowning person; The rescue tool delivery module 25 is used to control the drone to arrive above the location of the drowning person with the highest rescue priority, and control the drone to deliver the corresponding rescue tool; The traction rescue module 26 is used to control the drone to release a traction rope for rescue when it is determined that the drowning person has come into contact with the rescue tool.

[0119] Optionally, the positioning and path planning module 22 is also used to perform superposition calculations on the initial coordinates and water velocity vector in the distress signal, and obtain first data in combination with the timestamp; based on the first data, a drift model is used to predict and obtain the position area; combining the position area with the current position of the drone, the path is planned, and the drone is controlled to travel along the path.

[0120] Optionally, the status information acquisition module 23 is further used to obtain images taken by the drone in the location area, and obtain behavioral characteristics of the drowning person through the images; obtain thermal images scanned by the drone in the location area, and obtain the body temperature of the drowning person; based on the high-frequency electromagnetic waves emitted by the millimeter-wave radar carried by the drone, capture the chest fluctuation characteristics of the drowning person, and obtain the respiratory frequency based on the chest fluctuation characteristics; integrate the behavioral characteristics of the drowning person, the respiratory frequency and the body temperature to obtain the status information of each drowning person.

[0121] Optionally, the rescue decision module 24 is further used to score the behavioral characteristics, the respiratory rate, and the body temperature in the status information according to a preset standard range to obtain the behavioral characteristic score, respiratory rate score, and body temperature score of the drowning person; substitute the behavioral characteristic score, the respiratory rate score, and the body temperature score into the priority score calculation formula to calculate the rescue priority score of each drowning person; according to the rescue priority score of each drowning person, the rescue priority score is inversely proportional to the rescue priority, to obtain the rescue priority of each drowning person; determine the rescue tool based on the location distribution data of the drowning person, and the location distribution data is obtained by taking images of the location area by the drone.

[0122] Optionally, the rescue decision module 24 is further configured to calculate the priority score using the following formula: Among them, p i is the rescue priority score of the i-th drowning person, and the indicators include j items, namely the action characteristics, the breathing rate, and the body temperature; w j is the weight of the jth indicator; S ij is the j-th indicator score of the i-th drowning person; is the best score of the jth indicator; i is the rate of deterioration of the state of the i-th drowning person, and the smaller the value, the slower the state change; t is the drowning duration.

[0123] The rescue tool delivery module 25 is further configured to comfort the drowning person through the interactive function of the drone and guide the drowning person to use the rescue tool.

[0124] The traction rescue module 26 is also used to obtain the changes in the pressure parameters of the rescue tool based on the built-in rescue sensor, activate the magnetic interface of the rescue tool, determine that the drowning person has come into contact with the rescue tool based on the aerial video stream transmitted back by the drone, control the drone to release the traction rope, and release the magnetic lock on the traction rope to connect the traction rope to the rescue tool.

[0125] It should be noted that the above embodiments provide devices that implement their functions using only the division of the above functional modules as examples. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the device and method embodiments provided in the above embodiments are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0126] This embodiment also discloses an electronic device, referring to Figure 7 The electronic device 700 may include: at least one processor 701 , at least one communication path 702 , a user interface 703 , a network interface 704 , and at least one memory 705 .

[0127] The communication path 702 is used to implement the connection and communication between these components.

[0128] The user interface 703 may include a display screen (Display) and a camera (Camera). The optional user interface may also include a standard path interface and a pathless interface.

[0129] The network interface 704 may optionally include a standard pathed interface or a pathless interface (such as a WI-FI interface).

[0130] The processor 701 may include one or more processing cores. The processor 701 utilizes various interfaces and paths to connect various components within the server. By running or executing instructions, programs, code sets, or instruction sets stored in the memory 705, and by accessing data stored in the memory 705, the processor 701 performs various server functions and processes data. Optionally, the processor 701 may be implemented using at least one hardware form selected from the group consisting of digital signal processing (DSP), field-programmable gate array (FPGA), and programmable logic array (PLA). The processor 701 may integrate one or a combination of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily processes the operating system, user interface, and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; and the modem is used to handle pathless communication. It is understood that the modem may not be integrated into the processor 701 and may be implemented separately on a single chip.

[0131] Among them, the memory 705 may include a random access memory (RAM) or a read-only memory (ROM). Optionally, the memory 705 includes a non-transitory computer-readable storage medium. The memory 705 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 705 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 705 may also optionally be at least one storage device located away from the aforementioned processor. As shown in the figure, the memory as a computer storage medium may include an operating system, a network communication module, a user interface module and an application based on drone rescue.

[0132] exist Figure 7 In the electronic device shown, the user interface 703 is mainly used to provide an input interface for the user and obtain data input by the user; and the processor 701 can be used to call an application based on drone rescue stored in the memory 705. When executed by one or more processors, the electronic device executes one or more methods in the above embodiments.

[0133] It should be noted that for the aforementioned method embodiments, for simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that this application is not limited by the order of the actions described, because according to this application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required for this application.

[0134] In the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0135] In the several embodiments provided in this application, it should be understood that the disclosed devices can be implemented in other ways. For example, the device embodiments described above are merely schematic, such as the division of units, which is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some service interface, and the indirect coupling or communication connection of devices or units can be electrical or other forms.

[0136] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0137] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0138] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a memory and includes several instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the various embodiments of the present application. The aforementioned memory includes various media that can store program codes, such as USB flash drives, mobile hard drives, magnetic disks or optical disks.

[0139] The above is only an exemplary embodiment of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification, those skilled in the art will easily think of other embodiments of the present disclosure. This application is intended to cover any variations, uses or adaptive changes of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary technical means in the technical field that are not recorded in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A rescue method based on drone, characterized in that: Applied to the control center, the method The law includes: receiving a distress signal sent by an image processing device, the image processing device being deployed in a water area, the distress signal being generated by the image processing device when the image processing device identifies a drowning person; Calculating the location of the drowning person based on the distress signal, planning a path based on the location, and controlling the drone to travel along the path; After detecting that the drone has arrived at the location area, the drone is used to collect status information of each drowning person; the rescue priority of each drowning person and the corresponding rescue tool are determined based on the status information of each drowning person; the drone is controlled to arrive above the location of the drowning person with the highest rescue priority, and the drone is controlled to release the corresponding rescue tool; When it is determined that the drowning person has come into contact with the rescue tool, the drone is controlled to release a traction rope for rescue.

2. The method according to claim 1, characterized in that The method of calculating the location area of ​​the drowning person according to the distress signal, planning a path according to the location area, and controlling the drone to travel along the path specifically includes: Performing a superposition calculation on the initial coordinates and the water velocity vector in the distress signal and combining them with a timestamp to obtain first data; and obtaining the location area by using a drift model prediction based on the first data; The path is planned based on the location area and the current location of the drone, and the drone is controlled to travel along the path.

3. The method according to claim 1, characterized in that The collecting of status information of each drowning person by the drone specifically includes: Obtaining images captured by the drone in the location area, and obtaining behavioral characteristics of the drowning person through the images; obtaining thermal images scanned by the drone in the location area, and obtaining the body temperature of the drowning person; Based on the high-frequency electromagnetic waves emitted by the millimeter-wave radar carried by the drone, the chest fluctuation characteristics of the drowning person are captured, and the respiratory frequency is obtained according to the chest fluctuation characteristics; The behavioral characteristics, respiratory rate, and body temperature of the drowning person are integrated to obtain the status information of each drowning person.

4. The method according to claim 3, characterized in that The determining of the rescue priority and corresponding rescue tools for each drowning person according to the status information of each drowning person specifically includes: Scoring the behavioral characteristics, the respiratory rate, and the body temperature in the status information according to a preset standard range to obtain a behavioral characteristic score, a respiratory rate score, and a body temperature score of the drowning person; Substituting the behavioral characteristic score, the respiratory rate score, and the body temperature score into a priority score calculation formula to calculate a rescue priority score for each drowning person; According to the rescue priority score of each drowning person, the rescue priority score is inversely proportional to the rescue priority, and the rescue priority of each drowning person is obtained; The rescue tool is determined based on the position distribution data of the drowning person, where the position distribution data is obtained by taking an image of the position area by the drone.

5. The method according to claim 4, characterized in that The priority score calculation formula is: Among them, p i is the rescue priority score of the i-th drowning person, and the indicators include j items, namely the action characteristics, the breathing rate, and the body temperature; w j is the weight of the jth indicator; S ij is the j-th indicator score of the i-th drowning person; is the best score of the jth indicator; i is the rate of deterioration of the state of the i-th drowning person, and the smaller the value, the slower the state change; t is the drowning duration.

6. The method according to claim 1, characterized in that After controlling the drone to release the corresponding rescue tool, the method further includes: The drowning person is comforted by the interactive function of the drone and is guided to use the rescue tool.

7. The method according to claim 1, characterized in that When it is determined that the drowning person has come into contact with the rescue tool, controlling the drone to release a traction rope for rescue specifically includes: Based on the built-in rescue sensor, the pressure parameter changes of the rescue tool are obtained, the magnetic interface of the rescue tool is activated, and based on the aerial video stream sent back by the drone, it is determined that the drowning person has touched the rescue tool. The drone is controlled to release the traction rope and release the magnetic lock on the traction rope to connect the traction rope to the rescue tool.

8. A rescue system based on drones, characterized in that: include: a signal receiving module, configured to receive a distress signal sent by an image processing device, the image processing device being deployed in a water area, the distress signal being generated by the image processing device when it identifies a drowning person; A positioning and path planning module is used to calculate the location area of ​​the drowning person based on the distress signal, plan a path based on the location area, and control the drone to travel along the path; A status information collection module is configured to collect status information of each drowning person through the drone after detecting that the drone has arrived at the location area; A rescue decision module determines the rescue priority and corresponding rescue tools for each drowning person according to the status information of each drowning person; A rescue tool delivery module is used to control the drone to arrive above the location of the drowning person with the highest rescue priority, and control the drone to deliver the corresponding rescue tool; The traction rescue module is used to control the UAV to release the traction rope for rescue when it is determined that the drowning person has come into contact with the rescue tool.

9. An electronic device, characterized in that: The electronic device includes a processor 701, a memory 705, a user interface 703, and a network interface 704, wherein the memory 705 is used to store instructions, the user interface 703 and the network interface 704 are both used to communicate with other devices, and the processor 701 is used to execute the instructions stored in the memory 705 so that the electronic device executes the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, and when the instructions are executed, the method according to any one of claims 1 to 7 is executed.

Images