2021 Global Conference on Consumer Electronics (GCCE) Kyoto, Japan

Diseased Crop AR

A drone-and-AR field assistance pipeline for diseased-crop removal. An RTK-GNSS drone surveys the field and pinpoints diseased plants from above; a mobile AR application then overlays in-field guidance to walk the operator straight to each plant.

Read the paper ↗ BibTeX
Drone hovering over a field with a marker and the target diseased crop highlighted on the ground

— Overview

Targeted removal of diseased crops is laborious. Operators traditionally walk the field by eye — looking for visual signs, remembering positions, and hoping nothing was missed. The work scales poorly and is error-prone when plants are far apart or visually similar to healthy ones.

This system pairs an autonomous drone with a mobile AR app. The drone captures geo-referenced aerial imagery and locates diseased plants from above; their positions are then handed to a tablet-based AR overlay that guides the operator to each plant in physical space.

Finding a sick plant in a field of healthy ones is easy from the sky and hard from the ground. AR closes that gap. — Diseased Crop Removal Assistance, GCCE 2021

— Hardware

The drone platform integrates three components: RTK-GNSS for centimeter-level positioning, a gimbal-stabilized camera for aerial imagery, and an onboard Jetson Nano for image processing close to the sensor.

Drone platform annotated with RTK-GPS, Jetson Nano, and gimbal-stabilized camera

— AR Guidance

A mobile AR application tracks the device pose in real time and overlays markers on the diseased plants identified by the drone. The operator no longer needs to memorize positions — they just follow the on-screen guide through the field.

Mobile AR view: image tracker registered to a marker, with diseased crops highlighted in the field

— Findings

Field experiments in outdoor environments verified the accuracy and feasibility of the pipeline. The AR registration stayed stable along realistic walking trajectories, and operators successfully located targeted plants without prior survey of the field.

— Authors

Lingrong Kong, Han Zhang, and Hajime Nobuhara.

University of Tsukuba.

— Links

— BibTeX

@inproceedings{kong2021diseased,
  title     = {Diseased Crop Removal Assistance System Using Augmented Reality and Drones},
  author    = {Kong, Lingrong and Zhang, Han and Nobuhara, Hajime},
  booktitle = {2021 IEEE 10th Global Conference on Consumer Electronics (GCCE)},
  pages     = {522--525},
  year      = {2021},
  publisher = {IEEE}
}