Abstract

With the decline of rural populations across the globe, much hope is vested in the use of robots in agriculture as a means to sustain food production. This is particularly relevant for high-value crops, such as strawberries, where harvesting is currently very labour-intensive. As part of a larger project to build a robot that is capable of harvesting strawberries, we have studied the identification of the picking point of strawberries — the point that a robot hand should grasp the strawberry — from images of strawberries. We present a novel approach to identify the picking point and evaluate this approach.