Integrating UAV-based multispectral imaging with ground-truth soil nitrogen content for precision agriculture: A case study on paddy field yield estimation using machine learning and plant height monitoring

This research explores how multispectral UAVs assist plant height monitoring and paddy field yield estimation by combining the aerial imagery with soil nitrogen data. The primary objective of this research is to develop accurate and affordable models for improving farming by linking plant health indicators. A secondary aim is to enhance farming by integrating plant health indicators from UAVs with soil nutrient levels. Multispectral UAV (Phantom 4), which provides five multispectral bands (Blue, Green, Red, Red Edge, Near-Infrared) and one RGB camera, was used to capture images during six stages of the crop growth to calculate vegetation indices like NDVI, GRVI for assessing crop health. Soil samples were taken from nine spots, and nitrogen levels were measured throughout the six growth stages. UAV photogrammetric technique was used to estimate plant height by comparing the Digital Surface Model (DSM) at different growth stages, which was then compared to field measurements. The collected data was used to develop models that predict crop yield by analysing the connection between soil nitrogen, Plant height and vegetation indices. The results obtained concluded the interrelationship between vegetation indices, nitrogen levels and yield, which demonstrated that UAV-based monitoring can accurately predict crop performance. This approach helps farmers to use fertiliser and make more accurate predictions, encouraging precise agriculture. This research emphasizes the significance of evolving technologies like UAVs, in offering valuable information to farmers, agronomists and policymakers for better crop management and data driven decision making.