Agricultural Vegetation Monitoring Based on Aerial Data Using Convolutional Neural Networks

In the present paper we discuss a problem of recognition of a state of agricultural vegetation using aerial data of different spatial resolutions. To solve this problem, we develop a classifier allowing us to divide the input images into three classes, which are “healthy vegetation”, “diseased vegetation”, and “soil”. The proposed classifier is based on two convolutional neural networks allowing us to perform classification into two classes, namely “healthy vegetation” and “diseased vegetation” and “vegetation’ and “soil”.