Cotton yield prediction using drone derived LAI and chlorophyll content

Abstract

The unmanned aerial vehicles (UAV) have become a better solution for agricultural growers due to advanced features such as minimal maintenance costs, quick set-up time, low acquisition costs, and live data capturing. Near-ground remote sensing (drone) has opened up new agronomic opportunities for better crop management. This study predicted the seed cotton yield for a cotton field area located at Tamil Nadu Agricultural University, Coimbatore. Pearson correlation analysis and regression analysis were done for ground truth data and vegetation indices for validation and accuracy and also to find the best-performing indices. It was concluded that the Wide Dynamic Range Vegetation Index (WDRVI) showed a better correlation coefficient (R=0.959) with LAI ground truth data (R2=0.919). In contrast, the Modified Chlorophyll Absorption Ratio Index (MCARI) showed a better correlation coefficient (R=0.919) with SPAD chlorophyll ground truth data (R2=0.845). Then the best performing indices WDRVI and MCARI were further used for generating the yield model. High spatial resolution drone imageries for determining LAI and chlorophyll are reliable and rapid, as per the study. It helps to determine the LAI and chlorophyll at a spatial scale and their influence on yield production. This yield prediction was technical support for the widespread adoption and application of unmanned aerial vehicle (UAV) remote sensing in large-scale precision agriculture.

Keywords

• Drone
• LAI
• SPAD chlorophyll
• Multispectral imageries
• Vegetation indices
• Yield