Crop Type Classification Using Optical Remote Sensing

Crop type mapping at the field level is necessary for a variety of applications in agricultural monitoring and food security. In this thesis, the goal is to develop a suitable deep neural network architecture that could detect different crop types in remote sensing images. The crop types typically show varying temporal behaviour in terms of reflective characteristics and thus makes the task highly challenging particularly with a single time-stamp imagery. In this context, the aim in this work is to exploit the temporal characteristics of sequential multispectral SENTINEL 2 observations (images) to perform crop type classification using advanced deep learning methodologies. The outcome of the crop type mapping is useful in precisely forecasting the crop yield productions and retrieving structural parameters which in turn may help decision makers aid in assessing crop health and to devise management policies more effectively.

First degree in Computer Science, Agricultural Technology, with 2:1 or above. MSc degree in the relevant subject areas is desired.