Tar spot is a foliar disease of corn characterized by raised black spots that may or may not be surrounded by a tan or brown halo called a fisheye. Severe infection can lead to a 10-50% yield loss in corn. Timely detection of early symptoms is essential for implementing management tactics to reduce the disease. This study aims to propose a machine learning pipeline to estimate disease severity of tar spot of corn using unmanned aircraft systems (UAS) data. The overall process comprises structure from motion (SfM), canopy attributes extraction, dimensionality reduction, and regression. The proposed method was applied to UAS data collected from research subplots located at Pinney Purdue Agricultural Center (PPAC), Indiana, USA. The study contributes to the expansion of UAS technology in agriculture by providing reliable disease severity information of tar spot of corn.