Autonomous scouting using unmanned aerial vehicle

A system of autonomous scouting, the system comprising:

• an unmanned aerial vehicle having one or more sensors;
• and
• an edge-computing device coupled to the unmanned aerial vehicle, the edge-computing device including one or more processors and memories locally storing (i) at least a first trained machine learning model and (ii) instructions that, when executed, cause the one or more processors to perform operations of:
• determining an initial flight path above a crop field for an unmanned aerial vehicle to fly along, and
• operating the unmanned aerial vehicle to fly at a first height along the initial flight path, wherein operating the unmanned aerial vehicle to fly at the first height along the initial flight path includes:
• operating a first sensor, of the one or more sensors, of the unmanned aerial vehicle, to collect images for the crop field at a first frequency,
• for each image of the collected images, processing a respective image using the first trained machine learning model in response to the respective image being collected by the first sensor, wherein the first trained machine learning model processes the respective image to generate a respective output that indicates whether a respective target region is detected within the crop field,
• determining, based on the respective outputs generated by the first trained machine learning models in processing the collected images, one or more target regions within the crop field, and
• determining, based on the one or more target regions, one or more stops of a reversed flight path for the unmanned aerial vehicle to fly along at a second height, wherein determining the one or more stops includes:
• clustering the one or more target regions, and
• determining the one or more stops of the reversed flight path based on the clustering.

At least one non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, cause the one or more processors to:

• determine an initial flight path above a crop field for an unmanned aerial vehicle to fly along, wherein the unmanned aerial vehicle includes (i) one or more sensors, and (ii) an edge-computing device having one or more processors and a memory locally storing at least a first trained machine learning model;
• and
• operate the unmanned aerial vehicle to fly at a first height along the initial flight path, wherein the instructions to operate the unmanned aerial vehicle to fly at the first height along the initial flight path include instructions to:
• operate a first sensor, of the one or more sensors, of the unmanned aerial vehicle, to collect images for the crop field at a first frequency,
• for each image of the collected images, process a respective image using the first trained machine learning model in response to the respective image being collected by the first sensor, wherein the first trained machine learning model processes the respective image to generate a respective output that indicates whether a respective target region is detected within the crop field,
• determine, based on the respective outputs generated by the first trained machine learning models in processing the collected images, one or more target regions within the crop field, and
• determine, based on the one or more target regions, one or more stops of a reversed flight path for the unmanned aerial vehicle to fly along at a second height, wherein the instructions to determine the one or more stops include instructions to:
• cluster the one or more target regions, and
• determine the one or more stops of the reversed flight path based on the clustering.