Rosemary Biomass Estimation from UAV Multispectral Camera

Abstract

Salvia rosmarinus Spenn. (Lamiaceae), commonly known as “rosemary,” is a fragrant evergreen shrub native to the Mediterranean region. It has been utilized for centuries for its culinary, medicinal, and aromatic qualities. The extracts of S. rosmarinus contain various bioactive compounds, such as essential oils, phenolic compounds, flavonoids, and other secondary metabolites, which exhibit antibacterial properties. Similarly, the essential oil (EO) derived from rosemary showcases antioxidant, antimicrobial, anti-inflammatory, and insecticidal activities. The development of Unmanned Aerial Vehicles (UAV) and sensor technology makes it feasible to acquire crop remote sensing images with high temporal and spatial resolution, for timely and accurate crop monitoring. The study area is located in Caccamo (Italy), Palermo District. The aim of this study is to evaluate the efficacy of UAV-based remote sensing data and multimodal data fusion using RGB and multispectral sensors in estimating rosemary biomass production. The methodology permits the determination of rosemary canopy height providing crucial insights for informed decision-making in agriculture. Rosemary RGB and multispectral images at fool blooming were collected using a UAV (DJI P4M, Phantom4 Multispectral). Geographic Information System software was utilized to process multispectral images, extracting spectral information for each cardinal direction’s exposure. The high-resolution multispectral images were then seamlessly combined (mosaicked), leading to the creation of thematic maps showcasing vegetation indices (VIs), including the widely used NDVI (Normalized Difference Vegetation Index). The findings highlight the substantial applicability of Unmanned Aerial Vehicles (UAVs) equipped with multispectral cameras in acquiring valuable data related to vegetation biomass. These capabilities position UAVs as effective tools within Decision Support Systems (DSS). The study proves that the proposed approach can provide a relatively precise estimations of crop harvest properties.