Estimating of Maize Yield Using UAV Images and NDVI Index

Abstract

The objective of this research is to study the relationship of the maize growth using an aerial geographic mapping with an unmanned aerial-vehicle (UAV) in order to forecast the actual number of yielded maizes. The work was done using Normalized Difference Vegetation Index (NDVI), acquired from considering the reflection of light waves Red (668 nm), Green (560 nm), NIR (842 nm) and Red Edge (717 nm). The study area was in Lam Sai, Wang Noi District, Phra Nakhon Sri Ayutthaya Province, started from June to November 2021. The experimental area was 12,467 square meters, separated into 198 blocks of 6x10 square meters. In the experiment, the area was divided into area A, prepared without soil improvement and B, with soil improvement, each measured 5,600 square meters in size. The result, after collecting the data for 107 days, showed that the relationship between NDVI and weight of maize seed can be described with linear regression, where the acquired yield forecast is Y = 0.67X – 30.927, with 95% level of confidence and 61% r-squared value. Furthermore, the correlation was calculated to be 78% when comparing the yield approximation of the entire area with the weight of harvested seeds. In addition, seeds at 30% humidity are 3,208 kilograms and 2,999 kilograms for area A and B, respectively. When compared to the areas consisting NDVI 50-59%, 60-69, 70-79%, 80-89% and 90-100%, the error was calculated to be 1.75 and 1.67, respectively.