Remote Sensing (Feb 2023)
Estimating Black Oat Biomass Using Digital Surface Models and a Vegetation Index Derived from RGB-Based Aerial Images
Abstract
Responsible for food production and industry inputs, agriculture needs to adapt to worldwide increasing demands and environmental requirements. In this scenario, black oat has gained environmental and economic importance since it can be used in no-tillage systems, green manure, or animal feed supplementation. Despite its importance, few studies have been conducted to introduce more accurate and technological applications. Plant height (H) correlates with biomass production, which is related to yield. Similarly, productivity status can be estimated from vegetation indices (VIs). The use of unmanned aerial vehicles (UAV) for imaging enables greater spatial and temporal resolutions from which to derive information such as H and VI. However, faster and more accurate methodologies are necessary for the application of this technology. This study intended to obtain high-quality digital surface models (DSMs) and orthoimages from UAV-based RGB images via a direct-to-process means; that is, without the use of ground control points or image pre-processing. DSMs and orthoimages were used to derive H (HDSM) and VIs (VIRGB), which were used for H and dry biomass (DB) modeling. Results showed that HDSM presented a strong correlation with actual plant height (HREF) (R2 = 0.85). Modeling biomass based on HDSM demonstrated better performance for data collected up until and including the grain filling (R2 = 0.84) and flowering (R2 = 0.82) stages. Biomass modeling based on VIRGB performed better for data collected up until and including the booting stage (R2 = 0.80). The best results for biomass estimation were obtained by combining HDSM and VIRGB, with data collected up until and including the grain filling stage (R2 = 0.86). Therefore, the presented methodology has permitted the generation of trustworthy models for estimating the H and DB of black oats.
Keywords
