UAV hyperspectral remote sensing for crop nitrogen monitoring: progress, challenges, and perspectives

Abstract

Nitrogen (N) is the most important macro nutrient for crop productivity. However, precise N input balancing productivity and environmental footprint is always challenging due the difficulty of measuring plant N status in a timely manner. Uncrewed aerial vehicles (UAVs) and hyperspectral imaging (HSI) have revolutionized smart vegetation monitoring, offering unprecedented capabilities for mapping plant traits and precision crop nutrient management. This review evaluates recent advancements in crop N monitoring, highlights key challenges in N prediction, and provides recommendations for improving UAV-based HSI to support decision-making and enhance N use efficiency. Our analysis reveals that traditional machine learning (ML) methods have surpassed the vegetation index-based regression methods in N prediction. However, most studies report significant gaps between model training accuracy and validation performance, with independent testing being insufficiently addressed in the literature. Some research highlighted the integration of solar-induced chlorophyll fluorescence (SIF) and plant traits to enhance N prediction by leveraging narrow-band and sub-nanometer resolution hyperspectral imagery. Additionally, we discuss the challenges of disentangling phenological and physiological effects from canopy structural and biochemical properties in hyperspectral analysis. Addressing these confounding factors through advanced modeling techniques like deep neural networks will significantly improve N prediction throughout the growing season. To overcome these challenges, we recommend adopting deep learning and hybrid modeling approaches that combine radiative transfer models, crop simulation models, and advanced machine learning algorithms. Such integrated methods, coupled with transparent data sharing, can enhance plant N prediction accuracy, improve model transferability, and ultimately support precision N management in agriculture.

Prof. Dr. Kang Yu

Professor of Precision Agriculture

My research interests include precision crop farming, hyperspectral remote sensing, and AI in agriculture.

Dr. Anirudh Belwalkar

Postdoctoral Researcher

My research interests include hyperspectral remote sensing of vegetation and precision agriculture.

Wuhua Wang

PhD student

My research focuses on inversion of crop canopy parameters (eg. canopy nitrogen concentration) and drone hyperspectral remote sensing.

PD. Dr. Yuncai Hu

Senior Researcher

My research interests include Remote sensing, Precision N nutrient management, Plant phenotyping for complex traits of abiotic stress tolerance, Agricultural N emissions, and Physiological mechanisms of plant responses to abiotic stresses.