Unmanned aerial vehicle descent

Patent metadata

Field	Value
Publication number	US-10088736-B2
Application number	US-201514863873-A
Country	US
Kind code	B2
Filing date	Sep 24, 2015
Priority date	Sep 24, 2015
Publication date	Oct 2, 2018
Grant date	Oct 2, 2018

Abstract

Described are systems and methods for surveying a destination as an unmanned aerial vehicle (“UAV”) descends toward the destination. To confirm that the destination is clear of objects and includes a safe landing or delivery location, such as a substantially planar surface, the UAV may capture and process images at different altitudes during the descent. Feature points of a first image captured at a first altitude may be paired with feature points of a second image captured at a second, different altitude. A homography may be computed to confirm that the paired feature points lie in the same plane and then the two images may be registered based on the paired feature points. The registered images may then be processed to determine depth information and determine if descent of the UAV is to continue or be aborted.

First claim

What is claimed is:

1. A method to determine a substantially planar surface area at a destination, the method comprising:
providing a user with a delivery location identifier to be placed at a delivery location within the destination;
detecting, by an unmanned aerial vehicle (“UAV”) positioned above the destination, the delivery location identifier;
obtaining with a camera of the UAV, at a first altitude above the destination, a first image of the destination, wherein the first image includes a first representation of the delivery location identifier;
obtaining with the camera of the UAV, at a second altitude above the destination, a second image of the destination, wherein the second image includes a second representation of the delivery location identifier;
determining a first plurality of feature points of the first image, wherein the first plurality of feature points includes a first feature point corresponding to at least a portion of the first representation of the delivery location identifier;
for each of the first plurality of feature points of the first image:
associating a feature point of the first plurality of feature points with a corresponding feature point of the second image to form a paired feature point that includes the feature point of the first image and the corresponding feature point of the second image;

determining a first plurality of paired feature points in which the feature point of the first image and the corresponding feature point of the second image are in a same plane, wherein a first pair of feature points of the first plurality of paired feature points includes the first feature point of the first image corresponding to at least a portion of the delivery location identifier and a second corresponding feature point of the second image corresponding to at least a portion of the delivery location identifier;
registering the first image and the second image based at least in part on the first plurality of paired feature points to generate a registered image set;
generating a depth disparity map of the destination based on the registered image set; and
determining, based at least in part on the depth disparity map, that a substantially planar surface of sufficient size to accept at least one of a package from the UAV or the UAV is available at the destination.

2. The method of claim 1 , further comprising:
determining a second plurality of paired feature points in which the feature point of the first image and the corresponding feature point of the second image are not in a same plane; and
discarding the second plurality of paired feature points.

3. The method of claim 1 , wherein the first image and the second image are obtained as the UAV descends toward the destination; and
the method further comprising:
continuing a descent of the UAV in response to determining that the substantially planar surface is available at the destination.

4. The method of claim 1 , wherein the delivery location identifier is uniquely identifiable.

5. A computer-implemented method, comprising:
under control of one or more computing systems configured with executable instructions:
providing a user with a delivery location identifier to be placed at a delivery location within a destination;
detecting, by an unmanned aerial vehicle (“UAV”) positioned above the destination, the delivery location identifier;
causing the UAV to descend toward the destination;
causing an image capture device of the UAV to obtain a first image of the destination when the UAV is at a first position, wherein the first image includes a first representation of the delivery location identifier;
causing the image capture device of the UAV to obtain a second image of the destination when the UAV is at a second position, wherein the second image includes a second representation of the delivery location identifier;
processing the first image and the second image to generate a registered image set, wherein the registered image set is generated based at least in part on the first representation of the delivery location identifier and the second representation of the delivery location identifier;
determining, based at least in part on the registered image set, that a substantially planar surface is available at the destination; and
causing the UAV to continue to descend toward the destination.

6. The computer-implemented method of claim 5 , wherein the substantially planar surface determined to be available at the destination is of sufficient size to accept at least one of a package from the UAV or the UAV.

7. The computer-implemented method of claim 5 , wherein processing the first image and the second image further includes:
determining a first plurality of feature points of the first image, wherein the first plurality of feature points includes a first feature point corresponding to at least a portion of the first representation of the delivery location identifier;
for each of the first plurality of feature points of the first image, associating a feature point of the first plurality of feature points with a corresponding feature point of the second image to form paired feature points that include the feature point of the first image and the corresponding feature point of the second image;
determining a first plurality of paired feature points in which the feature point of the first image and the corresponding feature point of the second image are in a same plane, wherein a first pair of feature points of the first plurality of paired feature points includes the first feature point of the first image corresponding to at least a portion of the delivery location identifier and a second corresponding feature point of the second image corresponding to at least a portion of the delivery location identifier; and
registering the first image and the second image based at least in part on the first plurality of paired feature points.

8. The computer-implemented method of claim 7 , wherein determining a first plurality of paired feature points in which the feature point of the first image and the corresponding feature point of the second image are in the same plane, further includes:
for each paired feature point:
determining a homography between the feature point of the first image and the corresponding feature point of the second image.

9. The computer-implemented method of claim 5 , wherein processing the first image and the second image to generate a registered feature set includes spatially aligning a first plurality of feature points of the first image with a second plurality of feature points of the second image.

10. The computer-implemented method of claim 5 , further comprising:
generating a depth disparity map of the destination based on the registered image set.

11. The computer-implemented method of claim 10 , wherein generating a depth disparity map includes:
applying a stereo matching algorithm to the registered image set to generate the depth disparity map.

12. The computer-implemented method of claim 10 , further comprising:
defining a plane around the delivery location identifier positioned at the destination; and
determining that a substantially planar surface is within the plane around the delivery location identifier.

13. The computer-implemented method of claim 5 , further comprising:
determining that the UAV has completed the descent toward the destination; and
causing the UAV to release a payload at the destination, wherein the payload includes an item ordered from an electronic commerce website.

14. The computer-implemented method of claim 5 , wherein the delivery location identifier is a visual

Assignees

• Amazon Tech Inc

Inventors

• Harris Scott Raymond
• Safarik Stephen Thomas
• Viola Paul

Classifications

• G06V20/17

  taken from planes or by drones · CPC title

• G03B15/006 Primary

  Apparatus mounted on flying objects · CPC title

• G06V10/40

  Extraction of image or video features · CPC title

• G06V20/13

  Satellite images · CPC title

• G05D1/0676 Primary

  specially adapted for landing · CPC title