Bark Beetle Grouped Images for AI Classification

Dataset Summary

This dataset comprises high-resolution photographs of bark and ambrosia beetles captured under controlled laboratory conditions. Each image contains multiple beetle specimens arranged on a uniform white background while submerged in 70% ethanol. This approach speeds up data collection and ensures reproducible imaging conditions. Individual beetle images can later be extracted from these grouped photographs for further analysis.

Dataset Details

Image Type: Grouped macro photography images.
Resolution: 3456 x 5184 pixels, 72 dpi, 24-bit color.
Equipment: Captured using a Canon EOS REBEL with a 60mm lens and ring flash.
Capture Conditions:
- Beetles are arranged in a petri dish on a white ceramic tile.
- Specimens are fully submerged in 70% ethanol.
- Standardized lighting and fixed camera setup minimize shadows.
Total Images: 2,860 images.
Average Beetles per Image: Approximately 12.9 (varies by species).

What Are Grouped Images?

In this dataset, each photograph is taken to include several beetle specimens simultaneously. These grouped images are designed to maximize throughput during image acquisition while maintaining consistency in background, lighting, and camera setup. Later, individual beetle images can be segmented from these grouped photographs using provided scripts or custom preprocessing methods.

Species Included

The dataset includes images for the following 12 bark beetle species:

Coccotrypes dactyliperda
Hylesinus varius
Monarthrum fasciatum
Phloeosinus dentatus
Pityophthorus juglandis
Platypus cylindrus
Pycnarthrum hispidium
Scolytodes glaber
Xyleborinus saxesenii
Xyleborus affinis
Xylosandrus compactus
Xylosandrus crassiusculus

Data Generation Process

Specimen Preparation:
- Pre-classified beetle specimens were stored in 70% ethanol at -80°C.
- Each species was processed separately to ensure accurate digital labeling.
Photography Setup:
- Beetles were arranged in a petri dish, ensuring minimal overlap and clear separation (a small ball bearing served as a reference).
- The camera was mounted approximately 11 cm above a ceramic tile, with manual focus and controlled settings (ISO 160, shutter speed 1/50, f-stop 5.6).
- A ring flash provided uniform illumination and minimized shadows.
Image Acquisition:
- Ten images were captured per batch with slight repositioning between shots to obtain varied angles.
- Each photograph captures several beetle specimens at once, facilitating efficient data collection.
Data Handling:
- Only the grouped images are included in this dataset.
- Users can further process these images to extract individual beetle specimens if desired.

Intended Use

This dataset is ideal for:

Training and evaluating deep learning models for insect classification.
Developing automated identification tools for forest health management.
Conducting ecological and biosecurity research involving bark beetle species.

Caveats and Limitations

Grouped Images Only: The dataset provides images containing multiple beetle specimens per shot. Users may need to apply segmentation techniques to extract individual images.
Controlled Conditions: Images were captured under laboratory conditions, which may differ from field environments.
Species Coverage: Only 12 specific species are represented; additional species will be necessary for broader generalization.

Citation

If you use this dataset in your research, please cite:

Marais, G. C., Stratton, I. C., Johnson, A. J., & Hulcr, J. (Year). Progress in Developing a Bark Beetle Identification Tool. University of Florida, Gainesville, FL, USA.

Also, acknowledge the photography protocol:

Stratton, I. C., Johnson, A. J., Marais, C., & Hulcr, J. (Year). Bark and Ambrosia Beetle Macro Photography for an AI Training Dataset. University of Florida, UF Forest Entomology Lab.

License

This dataset is made available under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.

image imagewidth (px) 3.46k 6k