Datasets for time-lapse camera monitoring of insects and their floral environments

Contains the dataset for training and validation of models to estimate flower cover and identify taxa of arthropods in time-lapse camera recordings described in the paper:

Kim Bjerge, Henrik Karstoft, Hjalte M. R. Mann, Toke T. Høye, A deep learning pipeline for time-lapse camera monitoring of insects and their floral environments, 2024, bioRxiv, https://doi.org/10.1101/2024.04.12.589205

The zip files contain the needed files and directory structure to train the models in Python code published at: https://github.com/kimbjerge/insectsFlowers

Content of zip files:
===============

insects.zip: Contains images and labels in YOLO format: https://github.com/ultralytics/yolov5/issues/2293

trainI21m contains the images and labels to train the insect detector with YOLOv5. Contains only the motion-informed enhanced images (MIE).
testI21m contains the images and labels to test the insect detector trained with YOLOv5. Contains only the motion-informed enhanced images (MIE).

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

Flowers.zip: contains the images of plants and flowers with black and white masks to train the DeepLabv3 flower semantic segmentation model.

----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

NI2-19cls.zip: contains images for training and validation of the arthropod classifiers 

Image crops of arthropods are organized in 19 subdirectories one for each class.

A1-Coccinellidae
B2-Coleoptera
C3-Background
D4-Bombus
E5-Syrphidae
F6-Lepidoptera
G7-Aranaeae
H8-Formidicidae
I9-Diptera
J10-Hemiptera
K11-Isopoda
L12-Uspecificerede
N13-Hymenoptera
O14-Orthoptera
P15-Rhagonycha_fulva
Q16-Satyrinae
R17-Aglais_urticea
S18-Odonata
T19-Apis_mellifera