The summer of 2018 I joined the NASA DEVELOP Program for a second 10-week contract. I was part of the North Dakota & Georgia Agriculture & Food Security II with my teammates, Jared Belvin and Connor Holzman. The objective of this project was to explore the potential of including synthetic aperture radar (SAR) scenes in with optical imagery for crop classification to improve early-year classification accuracy. I digitized field boundaries with different crop types for training data for a random forest classification model. I did this twice, once for optical only imagery, and once where we included SAR as one of the inputs for the model. I then calculated confusion matrices to compare the accuracy of each classification run, and we found that the version that included SAR had a better accuracy.

As part of the project, I created a video summary that broke down the analysis at an 8th grade level. Additionally, we presented our findings at NASA HQ in D.C. as part of the 2018 Summer Showcase.

For more information on the project, including our abstract and poster, please check out our project website.