Use of Aerial NDVI Imagery in Determining Plant Water Status of Individual Grape Vines

With incidents of drought increasing in California and other areas around the world, increased water conservation in agriculture is more essential than ever before. As such, the developement of technology focused on understanding the plant water status of key agricultural crops is more important than ever before. In our study we sought to assess the feasibility of using aerial Normalized Difference Vegetative Index imagery to determine the plant water status of individual grape vines within a vineyard. Our research focused on fusing ground data with aerial data by comparing individual grapevine pruning weights to NDVI values extracted from aerial images. The ground and aerial data have different scales of resolution, both in space and time. We aimed to understand how to fuse these types of data. Additionally, our study focused on disambiguating spatial fluctuations into periodic parts (regularly spaced plants) and fluctuations due to irrigation and environment. We began our study by focusing on descriptive statistics (means and variances) and scatterplots (to understand any basic patterns). Next, we moved to more modern techniques, such as random forest and low-rank approximations to further understand patterns. In the conclusion of our study, we did not find NDVI adept in determining in time fluctuations in plant water strain of individual vines. However, our study revealed the potential for NDVI to be used in identification of live vegetation on a scale as fine as the individual grape vine.