A Framework for Innovation in Earth Observation Applications for Agriculture

Abstract

Agriculture is a top priority both for Romania and the European Union. Agriculture can largely benefit from the Earth Observation data freely available from the Sentinel 2 satellites within the context of the Copernicus program. The validation and correlation of satellite measurements with the in-situ measured data are extremely important for the correct exploitation of the remote sensing data. One way to foster satellite and in-situ data is to use Artificial Intelligence models and tools for extracting useful information for farmers and landowners. In this article, we identify the current needs in the agricultural domain as well as various aspects where innovation can occur in the data processing chain. We focus on convolutional neural networks as this type of deep learning model is perfectly suited for the analysis of images.