Recently, remote sensing (RS) imagery acquired by UAVs has been investigated for counting objects such as plants, because high temporal and spatial resolution data can be acquired over large fields. Manual counting in the large orthophotos is similar to field-based manual counting in the sense that it is a subjective, tedious task and should be performed by experts because of the highly overlapped and complex shape of plants as they develop. Therefore, automatic and accurate techniques based on deep learning have gained particular attention.
We have developed significant algorithms based on few-shot learning for plant counting and flowering date estimation.