Real-Time Avocado Plant Health and Disease Detection Using UAV Imagery with Faster R-CNN Algorithm

Abstract

Avocado cultivation is a rapidly growing industry known for its creamy, nutritious fruit and economic value In Tamil Nadu the plane bear fruits for a period of 3 to 4 years after which tree becomes highly susceptible to diseases like Anthracnose, Root Rot, Algal Leaf Spot and Scab which poses a severe threat to the crop and in worst cases the tree dies. Therefore, current techniques of disease inspection such as manual inspections and RGB image analysis are imprecise for early detection since RGB datasets only involve channels in the visible spectrum. This is where multispectral imaging shines by covering ‘hidden’ spectral bands such as Orange, Cyan, and Near Infrared bands that are not visible by normal human eye. The research combines UAV systems with multispectral imagery to perform avocado disease identification as a solution for covering extensive monitoring areas spanning high tree heights. The researchers used Faster R-CNN to classify diseases through multispectral datasets training because Near-Infrared scanning proved most successful at detecting infections. The detection capabilities are improved through better feature extraction methods and optimized model training process. Identifying diseases at an early stage enables farmers to intervene in time thus protecting their avocado crops for continuous production. This research establishes new opportunities for UAV-based disease detection through multispectral methods that help maintain agricultural sustainability and avocado industry economic stability.

Plant phenotyping relevance

UAVマルチスペクトル画像とFaster R-CNNによってアボカドの病害状態を直接推定する手法が研究の中心であり、植物病害フェノタイピングに該当する。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.