AI Transcribe labels from the Vascular plant hebarium

[rukayaj]

From BP:

In the vascular plant collection, the Nordic part, we have about 4800 labels that are not transcribed or partly transcribed. When this is transcribed almost all old herbarium sheets from this collection is transcribed.

I wish help if you can transcribe with AI from handwritten and typed labels. And then I can proofread the results before I import it to the data base.

Topics wished to be transcribed are, see the attached file Nordisk_uten_lokalitet):

• Locality (column E)
• Coordinate, this is MGRS almost always with to letters and four digit, the model wanted shown as example UH 68,14 (on labels the , is often missing)
• LegDate (column K) - the challenge is often that century is missing - model for the date dd.mm.yyyy., but if century is missing you can use dd.mm.yyyy.
• Leg (collector) (column M) - see about attached file further down.

In the attached file you find the record number, taxon, and for some the Leg date and Leg (collector). I have approved this material to day so for the moment the information with images are not on GBIF. Or the images can be found here (model). https://unimus.no/photos/image/jpeg/O-V-xxxxxx-01, and replace the xxxxxx with the actual record number. But nice if you don't overwrite the partly described field LegDate and Leg and for some of the records.

I have also attached an excel file with person names from the vascular plant database (Persons_VascularPlants). If AI find some new names it should be very few. Use column E for search for the collector name and I would be nice if you could translate the name to surname, given name (see column F).

[rukayaj]

@MichalTorma is working on this

[rukayaj]

Closing as complete