Analysis of Vegetation Indices in the Cotton Crop in South Asia region using UAV Imagery

Cotton is one of the most economically significant crops in the South Asia region. Following the decline in cotton production in recent years, it is necessary to study and analyse cotton growth patterns by using modern precision agriculture techniques to enhance crop yield and quality. Multispectral imagery plays a vital role in estimating crop quantification by computing vegetation indices at different phenological stages of the crop. It facilitates in estimating plant health and impact on yield before the visible symptoms appear. In this study, four different vegetation indices maps are computed using two-band multispectral imagery acquired through Unmanned Aerial Vehicle (UAV). The UAV multispectral data is collected from two fields with different sowing dates using Sentera NDVI single sensor. The variation in the vegetation indices map generated is analysed by vegetation index thresholding and k-means clustering. By using Sentinel-2A imagery, an average value of each vegetation index for every phenological stage is validated. The correlation coefficient of each vegetation index conveys a positive progression trend and validates the accuracy of UAV data computed results. In conclusion, vegetation indices computed using multispectral UAV imagery conveys vital information regarding the progression of crop phenology and growth pattern.