Potatoes furrow lines detection using aerial RGB image processing

Abstract

Navigation in crops has been one of the most important challenges of precision agriculture, especially in countries with diverse geographies, where the variety of the terrain is a problem for autonomous devices that can carry out different activities on plants, such as harvesting, fumigation, and supervision. For this reason, a methodology was developed in this paper for the identification of lines in the furrows of potato crops using RGB images taken by a UAV (Unmanned Aerial Vehicle) at different heights. These results will be used as the basis for the furrow modeling and later as a guide in drone navigation and robots set to explore the crops. In this paper, a methodology for processing aerial images taken with a UAV (a drone) was developed to automatically obtain the lines that form the furrows in a potato crop. The results were successful after using indexes adapted over images of potato crops.