Research Description

This research focused on estimating maize crop height and yield estimation by integrating UAV technology and deep learning models. Data collection was conducted over two agricultural seasons (rainy and dry), using a DJI Phantom IV drone (version 2+) with Ground Control Points (GCPs) to ensure positional accuracy. Weekly drone mapping was carried out using PIX4D Mapper at altitudes of 25m, 30m, 50m, and 70m. UAV imagery was collected at three key intervals: one week, two weeks, and three weeks after sowing, with manual seedling counts conducted on the ground for validation. 

In the analysis, two advanced deep learning models, YOLOv8 and Faster R-CNN  along with the Extra-large variant of YOLOv8, to detect and classify maize seedlings. By integrating UAV imagery with advanced deep learning techniques, this work highlighted innovative approaches to precision agriculture, emphasizing accurate crop monitoring for improved agriculture crop monitoring to provide reliable and precise crop height measurements.