US 12,282,338 B1
Automated generation of training data including airborne object trajectory data
Harshul Gupta, Kent, WA (US); and Francesco Giuseppe Callari, Seattle, WA (US)
Assigned to Amazon Technologies, Inc., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on Nov. 29, 2022, as Appl. No. 18/059,667.
Int. Cl. G05D 1/10 (2006.01); G01S 19/01 (2010.01); G05D 1/00 (2006.01); G06T 5/50 (2006.01)
CPC G05D 1/101 (2013.01) [G01S 19/01 (2013.01); G06T 5/50 (2013.01); G06T 2207/20221 (2013.01); G06T 2207/30241 (2013.01)] 19 Claims
OG exemplary drawing
 
1. A computer-implemented method to generate training data including airborne object trajectory data, comprising:
receiving, by a control system from an airborne object, an actual trajectory in a first environment;
determining, by the control system, an elevation difference between the first environment and a second environment;
adjusting, by the control system, the actual trajectory based on the elevation difference;
randomly selecting, by the control system, an end point vector for a transformed trajectory in the second environment;
determining, by the control system, a rotation matrix based on a transformation difference between an end point vector of the actual trajectory and the end point vector for the transformed trajectory;
determining, by the control system for each vector of the actual trajectory, a respective delta between a respective vector and the end point vector of the actual trajectory;
determining, by the control system for each vector of the actual trajectory, a respective transformed vector of the transformed trajectory based on the rotation matrix and a respective delta;
generating, by the control system, the transformed trajectory based on the end point vector for the transformed trajectory in the second environment and the respective transformed vectors;
receiving, by the control system from an unmanned aerial vehicle, imaging data of the second environment; and
causing, by the control system, superimposition of the transformed trajectory within the imaging data of the second environment.