Exploratory approach for automatic detection of vine rows in terrace vineyards

Abstract

The Alto Douro Demarcated Region in Portugal is the oldest and most regulated wine-growing region in the world, formed by an ecosystem of unique value allowing the cultivation of vines on its characteristics terraces vineyards. The detection of vine rows in terrace vineyards constitutes an essential task regarding the achievement of important goals such as multi-temporal crop evaluation and production estimation. Despite the advances and research in this field, most studies are limited to flat vineyards with straight vine rows. In this study an exploratory approach in the precision agriculture for automatic detection of vine rows in terrace vineyards is presented with remote sensing techniques associated with artificial intelligence such as Machine Learning and Deep learning. At the current stage the preliminary results are encouraging for the detection of vine rows in straight and curved lines considering the complexity of the terrain.