UAV Paddy Rice Panicle Dataset

Accurate panicle segmentation is a key step in rice field phenotyping.  Deep learning methods based on high spatial resolution images provide a potential solution to increase the throughput as well as the accuracy of panicle identification.  The quality and volume of the dataset are crucial to training an accurate and robust deep learning model.  Panicle segmentation tasks require particularly costly annotations.  Here we open a paddy rice panicle dataset, acquired by DJI Mavic Pro in 2018, to public use for rice panicle phenotyping.