New study proposes AI model for date palm detection
A new study by a team of scientists from Loughborough University, the UK, Keele University, the UK, and the International Center for Biosaline Agriculture (ICBA) shows that the latest version of a deep learning-based object detection model called You Only Look Once (YOLO-V5) is one of the best methods to detect date palm trees using drones.
Published in Computers and Electronics in Agriculture, the study aimed to identify a suitable detection model for an automated date palm plantation management system that ultimately helps end users, including farmers, researchers, and agribusinesses, forecast date production and tree condition.
The study used 125 images of 5,472 x 3,648 pixels captured by a camera onboard a fixed-wing drone flying at 122 meters above farmlands in the Northern Emirates of the UAE, including Sharjah, Ras Al Khaimah, and Ajman.
Based on the results, YOLO-V5 recorded the highest accuracy with a mean average precision of 92.34% when compared with other popular convolutional neural network-based (CNN) object detection methods such as YOLO-V3, YOLO-V4, and Single Shot Detector (SSD) 300. The model was also able to detect and localize date palm trees of different sizes in crowded, overlapped environments and areas where date palm tree distribution was sparse.
CNN is one of the deep learning algorithms capable of extracting millions of high-level features of objects that can then be used effectively for object detection and classification.
Dr. Ali Elbattay, Senior Scientist in Remote Sensing and Drone Technology at ICBA and a co-author of the study, said: “The study is a very important milestone for us to understand how to better manage date palm fields using high-level data generated by drones in a short time. The approach serves both research and agribusiness purposes, helps improve yield and income and mitigate crop loss due to bad crop management practices.”
Date palm trees have great economic importance, especially in the Arabian Peninsula and the wider Middle East and North Africa region. Not only do they provide delicious and nutritious fruits but are also used for other purposes such as construction, specifically in rural communities.
Date palm is also one of the main research priorities for ICBA. Since 2002 the center has conducted different experiments at its research station in Dubai to determine the long-term effect of saline water irrigation on date palm growth, productivity, and fruit quality.
In line with the center’s research focus on data-driven agriculture, scientists have also been testing and implementing a system of the internet of things (IoT) and drone-based data collection integrated into a GIS-based artificial intelligence analytical platform for monitoring the date palm plantation at ICBA. Moreover, they have been evaluating the effectiveness of various techniques for controlling red palm weevil (Rhynchophorous ferrugineus), the most dangerous date palm pest in the Middle East and North Africa region.