Many people in South Sudan rely on international aid to meet basic food needs.  Having accurate estimates of crop land is required for correctly estimating crop yields and addressing food shortages caused by weather variances and political unrest.  Current estimates are limited to visual estimates of crop land, and outdated algorithms.  Data exists to improve these estimates, but AI must be written and trained to correctly identify crop land and to use the large amounts of raw data in an efficient way. This project is using samples of ground cover types to inform the data selection process.  The goal of this work is to improve food security for millions of people in rural sub-Saharan Africa.