Performance of RGB based vegetation indices in crop health and soil nutrient content assessment on a tea plantation

Unmanned Aerial Vehicle industry has exponentially grown in the past decade. This growth has resulted in the manufacture of low-cost and reliable UAVs whose implication is a paradigm shift in the applications of Unmanned Aerial Vehicles (UAVs) from military to a wide range of civilian applications. Among the civilian applications of UAVs is precision agriculture where drones/UAVs are used to assess the growth and health of crops to optimize resource allocation. Consumer-grade Unmanned Aerial Vehicles have proven to be cost-effective, have high spatial resolutions due to low flying heights, and routine observations over specific flight paths are easily implemented as a result of high automation of the entire proces.
Innovations in technology have also led to the development and adoption of numerous RGB-based vegetation indices in agricultural activities. Such indices include but are not limited to Visible Atmospheric Resistant Index, Triangular Greenness Index, Excess Greenness, Colour Index of Vegetation, and Vegetation Index Green. These indices have been developed as alternatives to the Normalised Difference Vegetation Indices as they have proven to be cost-effective due to the associated sensors and platform in data capture.
This paper attempts to assess the sensitivity of the RGB-based indices on a tea plantation using Normalised Difference Green Blue Vegetation Index as a reference index. Normalised Difference Green Blue Vegetation Index is argued to be the best model for Soil Plant Analysis Development of naked barley leaves. Correlation, regression, and T-test are used to evaluate the different levels of performance and relationships of the RGB-based indices to the reference index.