Dataset Open Access
Images and Annotations
This dataset contains microscopic images of pollen captured in Graz, Austria. Additionally, bounding box annotations are provided for each image. The images were captured with an automated pollen detection system as described in the paper Automated and Continuous Pollen Detection with an Affordable Technology. The annotations have been produced manually. This dataset contains data of two days (23.05.2019 and 26.05.2019).
The dataset can be used to train an object classifier with the code from the following repository:
In addition, this dataset contains two pre-trained weight files. One weight file is pre-trained on ImageNet (provided by Joseph Redmon) and the other is pre-trained for pollen detection on the above mentioned image/annotation combination.
The dataset (and related aspects) has partly been used and is described in more detail in the following publication:
N. Cao, M. Meyer, L. Thiele, O. Saukh. Automated Pollen Detection with an Affordable Technology. Proc. of the International Conference on Embedded Wireless Systems and Networks (EWSN), February 2020.