A framework for AI-based plant disease detection and autonomous robotic agricultural spraying

Abstract

This study presents a novel framework to detect plant diseases using artificial intelligence (AI) and efficient agricultural spraying using a mobile robot manipulator. The dataset for training the AI model was created by taking photos of plant leaves with and without disease and labeling the dataset according to the YOLO algorithm. A camera with a depth sensor providing point cloud data was used to detect plant disease and its location relative to the end effector was calculated using kinematic methods. The Robot Operating System (ROS) was used for system integration along with Moveit! package for kinematic calculations and motion planning of the robotic arm. The robotic arm is located on a two-wheeled mobile platform that autonomously navigates among the plants using Navigation-Stack of ROS. With the help of the developed spot spraying on the diseased area concept, not only the labor cost for agricultural spraying but also the amount of pesticides used for agricultural spraying can be reduced, lowering the pesticide costs and consumer exposure to the pesticides.

Code and data availability

植物フェノタイピング解析を再現する公開資産であることを、入力本文と直接リンクから確認できなかったため保留しました。

No evidence-backed public reproduction asset is currently recorded.