Image dataset for training of an insect detection model for the Insect Detect DIY camera trap

This dataset contains images of an artifical flower platform with different insects sitting on it or flying above it. All images were automatically recorded with the Insect Detect DIY camera trap, a hardware combination of the Luxonis OAK-1, Raspberry Pi Zero 2 W and PiJuice Zero pHAT for automated insect monitoring (bioRxiv preprint).
Classes
The following object classes were annotated in this dataset:

wasp (mostly Vespula sp.)
hbee (Apis mellifera)
fly (mostly Brachycera)
hovfly (various Syrphidae, e.g. Episyrphus balteatus)
other (all Arthropods with insufficient occurences, e.g. various Hymenoptera, true bugs, beetles)
shadow (shadows of the recorded insects)
View the Health Check for more info on class balance.
Versions

v4 insect_detect_416_1class

squashed to square (aspect ratio 1:1)
downscaled to 416x416 pixel
all classes merged into one class ("insect")
v5 insect_detect_raw_4K

original images in 4K resolution (3840x2160 pixel)
v7 insect_detect_320_1class

squashed to square (aspect ratio 1:1)
downscaled to 320x320 pixel
all classes merged into one class ("insect")
Deployment
You can use this dataset as starting point to train your own insect detection models. Check the model training instructions for more information.
Open source Python scripts to deploy the trained models can be found at the insect-detect GitHub repo.