Using UAV flight patterns to enhance machine vision detection of obstacles

Patent metadata

Field	Value
Publication number	US-12436540-B2
Application number	US-202318132288-A
Country	US
Kind code	B2
Filing date	Apr 7, 2023
Priority date	Apr 7, 2023
Publication date	Oct 7, 2025
Grant date	Oct 7, 2025

Abstract

A technique for detection of an obstacle by a UAV includes arriving above a location at a first altitude by the UAV; navigating a descent flight pattern from the first altitude towards the location; acquiring aerial images of the location below the UAV with a camera system disposed onboard the UAV; and analyzing the aerial images with a machine vision system disposed onboard the UAV that is adapted to detect a presence of the obstacle in the aerial images. The descent flight pattern is selected to increase perception by the machine vision system of the obstacle.

First claim

What is claimed is:

1. A method for detection of an obstacle by an unmanned aerial vehicle (UAV) providing an aerial delivery service, the method comprising:
arriving above a delivery location at a first altitude by the UAV;
initiating a descent flight pattern for the UAV that facilitates detection or tracking of the obstacle by a machine vision system disposed onboard the UAV, wherein the descent flight pattern comprises a rectangular spiral descent having linear flight path segments that sequentially descend;
navigating the descent flight pattern from the first altitude towards the delivery location;
acquiring aerial images of the delivery location below the UAV with a camera system disposed onboard the UAV, wherein at least some of the aerial images are acquired while the UAV navigates the descent flight pattern; and
analyzing the aerial images with the machine vision system disposed onboard the UAV to detect or track the obstacle in the aerial images.

2. The method of claim 1 , wherein analyzing the aerial images with the machine vision system comprises:
analyzing the aerial images with an optical flow algorithm, wherein the descent flight pattern includes aerial maneuvers by the UAV selected to increase perception of the obstacle by the optical flow algorithm.

3. The method of claim 2 , further comprising:
measuring an angular velocity of the UAV with a motion sensor disposed onboard the UAV while capturing the aerial images,
wherein analyzing the aerial images with the optical flow algorithm includes:
determining flow velocities of image pixels in the aerial images; and
offsetting rotational components in the flow velocities based upon the angular velocity measured by the motion sensor while capturing the aerial images.

4. The method of claim 1 , wherein the descent flight pattern maintains a constant heading for the UAV while the UAV tracks the spiral descent.

5. The method of claim 1 , wherein the obstacle comprises an object having an aerial perspective shape with a long axis and a short axis, wherein the descent flight pattern is oriented to vacillate along a linear axis that is deemed to be perpendicular to the long axis while the UAV descends towards the delivery location.

6. The method of claim 1 , further comprising:
analyzing the aerial images with a semantic model of the machine vision system that classifies objects detected in the aerial images into an obstacle class or non-obstacle class; and
initiating the descent flight pattern in response to one of the objects in the aerial images being classified into the obstacle class.

7. The method of claim 1 , where the obstacle comprises above ground powerlines or above ground telecommunication lines.

8. The method of claim 1 , wherein the camera system comprises one or more of a frame camera, a stereovision camera, an event camera, or a lidar depth camera.

9. The method of claim 1 , wherein the descent flight pattern is selected to improve estimation of an offset distance between the UAV and the object.

10. The method of claim 1 , wherein selecting the descent flight pattern for the UAV comprises selecting the descent flight pattern to improve tracking of the obstacle by the machine vision system disposed onboard the UAV while the UAV descends towards the delivery location.

11. At least one non-transitory machine-readable storage medium storing instructions that, when executed by a processing system of an unmanned aerial vehicle (UAV) will cause the UAV to perform operations comprising:
arriving above a location at a first altitude;
navigating a descent flight pattern from the first altitude towards the location;
capturing aerial images of the location below the UAV with a camera system disposed onboard the UAV, wherein at least some of the aerial images are captured while the UAV navigates the descent flight pattern;
analyzing the aerial images with an optical flow algorithm of a machine vision system to detect or track an obstacle in the aerial images; and
measuring an angular velocity of the UAV with a motion sensor disposed onboard the UAV while capturing the aerial images,
wherein analyzing the aerial images with the optical flow algorithm includes:
determining flow velocities of image pixels in the aerial images; and
offsetting rotational components in the flow velocities based upon the angular velocity measured by the motion sensor while capturing the aerial images.

12. The at least one non-transitory machine-readable storage medium of claim 11 ,
wherein the descent flight pattern includes aerial maneuvers by the UAV selected to increase perception of the obstacle by the optical flow algorithm.

13. The at least one non-transitory machine-readable storage medium of claim 11 , wherein the descent flight pattern comprises a spiral descent.

14. The at least one non-transitory machine-readable storage medium of claim 13 , wherein the descent flight pattern maintains a constant heading for the UAV while the UAV tracks the spiral descent.

15. The at least one non-transitory machine-readable storage medium of claim 13 , wherein the spiral descent comprises a rectangular spiral descent having linear flight path segments that sequentially descend.

16. The at least one non-transitory machine-readable storage medium of claim 11 , wherein the obstacle comprises an object having an aerial perspective shape with a long axis and a short axis, wherein the descent flight pattern is oriented to vacillate along a linear axis that is deemed to be perpendicular to the long axis while the UAV descends towards the location.

17. The at least one non-transitory machine-readable storage medium of claim 11 , wherein the operations further comprise:
analyzing the aerial images with a semantic model of the machine vision system that classifies objects detected in the aerial images into an obstacle class or non-obstacle class; and
initiating the descent flight pattern in response to one of the objects in the aerial images being classified into the obstacle class.

18. The at least one non-transitory machine-readable storage medium of claim 11 , where the obstacle comprises above ground powerlines or above ground telecommunication lines.

19. The at least one non-transitory machine-readable storage medium of claim 11 , wherein the camera system comprises one or more of a frame camera, a stereovision camera, an event camera, or a lidar depth camera.

20. At least one non-transitory machine-readable storage medium storing instructions that, when executed by a processing system of an unmanned aerial vehicle (UAV) will cause the UAV to perform operations comprising:
arriving above a location at a first altitude;
navigating a descent flight pattern from the first altitude towards the location;
acquiring aerial images of the location below the UAV with a camera system disposed onboard the UAV, wherein at least some of the aerial images are acquired while the UAV navigates the descent flight pattern; and
analyzing the aerial images with a machine vision system that is adapted to detect a presence of an obstacle in the aerial images,
wherein the obstacle comprises an object having an aerial perspective shape with a long axis and a short axis, wherein the descent flight pattern is oriented to vacillate along a linear axis that is deemed to be perpendicular to the long axis while the UAV descends towards the location.

21. At least one non-transitory machine-readable storage medium storing instructions that, when executed by a processing

Assignees

• Wing Aviation Llc

Inventors

• Shoeb Ali
• Hammond Marcus

Classifications

• G05D2111/10

  Optical signals · CPC title

• G05D2109/24

  Convertible aircraft, e.g. tiltrotor aircraft · CPC title

• G05D2107/17

  Spaces with priority for humans, e.g. populated areas, pedestrian ways, parks or beaches · CPC title

• G05D2111/65

  taken successively, e.g. visual odometry or optical flow · CPC title

• G05D2101/20

  using external object recognition · CPC title