Mango (Mangifera indica) tree detection and counting in mango orchard with satellite images using deep learning model YOLO: A comparative analysis

Abstract

Mango (Mangifera indica L.) is a widely cultivated horticultural cash crop in tropical and subtropical regions, valued for its exceptional taste, aroma, nutritional benefits, and medicinal properties. Tree counting is a crucial aspect of orchard inventory management, enabling efficient resource allocation, yield estimation, and precision agriculture applications. However, traditional methods often rely on manual efforts or expensive feature engineering, leading to errors, inefficiencies, and limited scalability. Recent advancements in deep learning-based approaches have demonstrated state-of-the-art performance in automated tree counting, offering improved accuracy, robustness, and computational efficiency. The study was carried out during 2023–24 presenting a comparative evaluation of YOLO architectures for mango tree detection and counting. The research analyzes YOLOv5, YOLOv6, YOLOv7, and YOLOv8 using satellite remote sensing imagery from the Bulandshahr district of Uttar Pradesh. Performance evaluation is conducted using precision, Recall, F1-score, and mean average precision (mAP). Experimental results reveal that YOLOv8 exhibits superior performance, achieving a well-balanced trade-off between detection accuracy, processing speed, and generalization. These findings highlight the potential of deep learning models for scalable orchard monitoring, precision agriculture, and sustainable fruit production.