Abstract Glass Building

LightningBug Workflows

LightningBug workflow can be viewed as six parts.

Image by Shapelined

1

Curate specimens before and after  imaging (i.e., pre-curation and post-curation).

Image by Shapelined

OCR/HCR  virtual labels and connect to structured text extraction services

4

Image by Shapelined

Develop  hardware and software platform to capture multi-view imagery of both labels and specimens.

2

Image by Shapelined

Apply photogrammetric analysis to assemble the 3D shape and structure of specimens.

5

Image by Shapelined

Build robust algorithms to automatically process fragmentary views of multiple labels into separate integrated “virtual labels".

3

Image by Shapelined

6

Provide research-ready data.  Labels become records  and shared with GBIF. Images stored & served from MorphoSource

Imaging Systems.jpg
Image by Silvio Kundt
Handwritten Labels
(100 Million)
Notes for Nature Transcription
Label Babel
Typed & Handwritten Labels
(40 Million)
OCR Machine Learning
Typed Labels
(120 Million)
Notes for Nature Vetting
Specimen enters staging area
Computer Vision Identification
Specimen enters staging area
Specimen Imaged
(basic to high resolution)
Trait Metrics (body, length, area)
Labels Imaged
Labels Reconstructed
LightningBug Robotics
LightningBug Software Processes
Not part of LightningBug Project