UAV rescue system and uav-based rescue method

Abstract

A drone rescue system and a drone-based rescue method are provided. The method includes: receiving a location; generating a search path for the drone based on the location and transmitting the search path to the drone; acquiring images captured by the drone and determining whether the images contain a person who has fallen into the water based on a machine learning model; and outputting an alarm message in response to determining that the images contain a person who has fallen into the water.

Claims

1. A drone rescue system, comprising: Transceiver, used for communication with the drone; Storage media, storing multiple modules; and a processor, coupled to the storage medium and the transceiver, and accessing and executing the plurality of modules, wherein the plurality of modules includes: a central control subsystem, and a receiving location; The task management subsystem generates a search path for the drone based on the location and transmits the search path to the drone; and the AI ​​processing subsystem acquires images captured by the drone, determines whether the images include a person who has fallen into the water based on a machine learning model, and outputs an alarm message in response to the determination that the images include the person who has fallen into the water, wherein the central control subsystem generates and outputs an event log, wherein the event log is associated with at least one of the following: the person who has fallen into the water, the flight path or time information of the drone.

2. The drone rescue system as described in claim 1, wherein the central control subsystem receives information from the drone, wherein the information includes drone coordinates, drone altitude, drone field of view and camera resolution, wherein the AI ​​processing subsystem calculates the coordinates of the person who fell into the water based on the information.

3. The drone rescue system as claimed in claim 2, wherein the AI ​​processing subsystem instructs the drone to deploy a rescue device to the coordinates.

4. The drone rescue system as claimed in claim 1, wherein the AI ​​processing subsystem acquires multiple historical images and trains the machine learning model based on the multiple historical images, wherein each of the multiple historical images has a label indicating whether a person who has fallen into the water exists.

5. The drone rescue system as claimed in claim 1, wherein the search path includes a first path, a second path connected to the first path, and a third path connected to the second path, wherein the second path is perpendicular to the first path, the third path is perpendicular to the second path, the third path is parallel to and opposite to the first path, the third path is longer than the first path, and a first time period corresponding to the first path is earlier than a second time period corresponding to the third path.

6. The drone rescue system as claimed in claim 5, wherein the first distance between the first path and the location is shorter than the second distance between the third path and the location.

7. The drone rescue system as claimed in claim 1, wherein the search path includes a first path corresponding to a first time period and a second path corresponding to a second time period later than the first time period, wherein the first path corresponds to a portion of a first virtual circle, wherein the center of the first virtual circle is the location and the radius of the first virtual circle is a first radius, wherein the second path corresponds to a portion of a second virtual circle, wherein the center of the second virtual circle is the location and the radius of the second virtual circle is a second radius, wherein the first radius is shorter than the second radius.

8. The drone rescue system as claimed in claim 1, wherein the search path includes a first path, a second path connected to the first path, and a third path connected to the second path, wherein the second path is perpendicular to the first path and the third path, and the first path is parallel to the third path.

9. A drone-based rescue method, comprising: Receiving location; The system generates a search path for the drone based on the location and transmits the search path to the drone; acquires images captured by the drone and determines whether the images include a person who has fallen into the water based on a machine learning model; outputs an alarm message in response to the determination that the images include the person who has fallen into the water; and generates and outputs an event log, wherein the event log is associated with at least one of the following: the person who has fallen into the water, the flight path of the drone, or time information.

Images