| CPC G01V 3/38 (2013.01) [G01V 3/081 (2013.01); G01V 3/16 (2013.01); G01V 11/00 (2013.01); G06T 7/70 (2017.01); G06V 10/774 (2022.01); G06V 20/17 (2022.01); G06T 2200/24 (2013.01); G06T 2207/10032 (2013.01); G06T 2207/20081 (2013.01)] | 20 Claims |

|
1. A computer-implemented method comprising:
training a machine-learning (ML) model to identify locations of buried pipes, the training based on a training set comprising locations of known buried pipes, images of a surface above the known buried pipes, and magnetic measurements taken over the known buried pipes;
programming a drone to fly over a geographical area with a first buried pipe, the programming based on an approximate location of the first buried pipe;
capturing, during a flight of the drone, geophysical data with geophysical equipment in the drone;
capturing, during the flight of the drone, images with a camera in the drone;
utilizing the machine-learning (ML) model to identify the location of the first buried pipe based on the captured geophysical data and the captured images; and
presenting the identified location of the first buried pipe in a map of the geographical area.
|