Assessment of the Optimal Flight Time of RGB Image Based Unmanned Aerial Vehicles for Crop Monitoring

Abstract

Unmanned Aerial Vehicles (UAVs) have been developed as a feasible tool for agricultural surveillance. Despite the fact that many researchers have focused on UAVs' ability to offer information on crop growth and development, study on the efficacy of day time period for images is extremely uncommon. As a result, the purpose of this research was to assess the best flying duration for RGB-based UAV technology for field crop monitoring and to develop a procedure for monitoring sugarcane using UAVs in Sri Lanka. The study was conducted on a five-month-old sugarcane field (1 hectare) in Ampara, Sri Lanka. All flights were missioned using a DJI Mavic pro drone (RGB) at flying heights, speeds, frontal overlap, and lateral overlap of 50 m, 4 m/s, 75%, and 70%, respectively. During the experiment day, images were captured during three flying time periods: T1 (07:00 – 09:00 h), T2 (10:00 – 12:00 h), and T3 (13:00 – 15:00 h), with three replicates per flight, and plant density (PD) data were manually recorded for 19 plots (5m×5m). The orthomosaic images were processed using Agisoft PhotoScan software, and the classification and accuracy assessments were carried out using Arc GIS to generate vegetation fraction (VF) and Green-red vegetation index (GRVI) values. To determine the optimal flying time, a relationship between UAV-based VF and plant density (PD) was generated. T2 performed better in vegetation mapping, with an overall accuracy of 88.37% and a Kappa coefficient of 0.75, because more shadowing regions were identified on the other two flights. At T2, the most significant correlation between VF and manual plant density was detected (R2 = 82.9%, SE = 2.20, P<0.05). T2 demonstrated a very strong relation between GRVI and PD (R2 = 82.1%, SE = 2.25, P<0.05). Overall, the ideal flight time can give more accurate and accurate crop monitoring results. The study concludes that the time range 10:00 – 12:00h might be used to acquire UAV images for crop monitoring. Keywords Plant Density, Vegetation Index, Flying time, image classification, vegetation analysis *Corresponding Author Email:pranith.piyumal@ageng.ruh.ac.lk 1. Introduction Agricultural monitoring on a regular basis is essential for addressing field constraints such as field gap detection, pest and disease concerns, weed management, and water stress issues in crop cultivations, resulting in enhanced production [1]. Traditional visual inspection is less productive since it involves more effort, money, and time. Precision farming technology has lately emerged as one of the most potential substitutes for manual crop monitoring [2].Satellite remote sensing imagery outperforms conventional crop monitoring methods [1,3]. However, it has certain disadvantages, such as lower spatial resolution, cloud cover,