BioAuxdataset: a dataset of ground-dwelling nocturnal fauna for object detection and classification

Description

The study of nocturnal fauna living in the soil is a challenging research task due to its multiple implications in different fields such as pesticide use control, crop yield prediction or plant disease identification. BioAuxdataset is the synthesis of hundreds of gigabytes of images taken in the field of insects and plant auxiliaries over several years. 7 common species of ground-dwelling nocturnal fauna are represented: ground beetles, mice, opilion, slugs, shrews, small slugs and worms. BioAuxdataset consists of a directory of 7470 jpg images and their corresponding 7470 xml annotation files. BioAuxdatset is available in the rar compressed format and its images are splitted into two different directories: non-augmented images and augmented images. The datasets “2016-6-30-Bushnell-Chrysope.rar”, “2016-7-13au15-Carnage.rar” and “2016-7-27-Bushnell-SpidersAnchomenusSlugs.rar” contain raw images acquired in the field, some of which are manually annotated. These datasets also contain a sub-directory called automaticannotation, which contains annotations obtained automatically by a learning model. These directories can be used both for learning new models and/or for validation. The objectives of the dataset range from monitoring the number of organisms and their characteristics (size, development stage) to early detection of plant diseases for more environmentally-friendly precision agriculture.

Categories

Funding

Licence