Guardado en:
| Autor principal: | |
|---|---|
| Formato: | Recurso digital |
| Lenguaje: | |
| Publicado: |
Zenodo
2026
|
| Acceso en línea: | https://doi.org/10.5281/zenodo.17600480 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866901661792862208 |
|---|---|
| author | Christiansen, Ditte Marie |
| author_facet | Christiansen, Ditte Marie |
| contents | <p><span lang="EN-GB">Codes used from scanned maps from printed pages to shapefile maps.</span></p> <p><span lang="EN-GB">1) We converted pdf scanned map pages to png format keeping the resolution to 600 dpi.</span></p> <p><span lang="EN-GB">2) align_panels and extract_panels are macros written for the program Fiji. Maps are extracted from the page and are aligned in angle and size based on a reference map using the written macros in Fiji. From Fiji, all map images are saved as jpeg files.</span></p> <p><span lang="EN-GB">3) Georeferencing of the reference map with 33 points located all around the Danish coastline. The georeferenced map coordinates are used to automatise the georeferencing for all other maps, R code named georeferencing_14nov.R</span></p> <p><span lang="EN-GB">4) Division of georeferenced maps into 10 </span><span lang="EN-GB">× </span><span lang="EN-GB">10 km grid cells and annotation of each grid cell image of 50 random maps after distribution category (species absence, species presence 1 and species presence 2). Annotated data are divided into train (60%), test (20%) and validation (20 %) and a pre-trained image classification algorithm is fine-tuned for predicting the three image categories. Prediction of all grid cell categories and estimation of precision of algorithm. Output is shapefile formatted maps with three distribution categories at a 10 </span><span lang="EN-GB">× </span><span lang="EN-GB">10 km resolution. All necessary code for this step is found in the zipfile map_project_final.</span></p> <p><span lang="EN-GB">5) As the island Bornholm was placed in an inserted frame near the rest of the country, we moved the grid cells of Bornholm to the right placement of Bornholm, using the code movebornholmmaps_14nov.R.</span></p> <p><span lang="EN-GB">6) We named all maps after the species name using the R code mapnaming_14nov.R and saved them as separate shapefiles.</span></p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_17600480 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Code repository for From paper to pixels - digitized maps of vascular plant distributions in Denmark in the early 20th century Christiansen, Ditte Marie <p><span lang="EN-GB">Codes used from scanned maps from printed pages to shapefile maps.</span></p> <p><span lang="EN-GB">1) We converted pdf scanned map pages to png format keeping the resolution to 600 dpi.</span></p> <p><span lang="EN-GB">2) align_panels and extract_panels are macros written for the program Fiji. Maps are extracted from the page and are aligned in angle and size based on a reference map using the written macros in Fiji. From Fiji, all map images are saved as jpeg files.</span></p> <p><span lang="EN-GB">3) Georeferencing of the reference map with 33 points located all around the Danish coastline. The georeferenced map coordinates are used to automatise the georeferencing for all other maps, R code named georeferencing_14nov.R</span></p> <p><span lang="EN-GB">4) Division of georeferenced maps into 10 </span><span lang="EN-GB">× </span><span lang="EN-GB">10 km grid cells and annotation of each grid cell image of 50 random maps after distribution category (species absence, species presence 1 and species presence 2). Annotated data are divided into train (60%), test (20%) and validation (20 %) and a pre-trained image classification algorithm is fine-tuned for predicting the three image categories. Prediction of all grid cell categories and estimation of precision of algorithm. Output is shapefile formatted maps with three distribution categories at a 10 </span><span lang="EN-GB">× </span><span lang="EN-GB">10 km resolution. All necessary code for this step is found in the zipfile map_project_final.</span></p> <p><span lang="EN-GB">5) As the island Bornholm was placed in an inserted frame near the rest of the country, we moved the grid cells of Bornholm to the right placement of Bornholm, using the code movebornholmmaps_14nov.R.</span></p> <p><span lang="EN-GB">6) We named all maps after the species name using the R code mapnaming_14nov.R and saved them as separate shapefiles.</span></p> |
| title | Code repository for From paper to pixels - digitized maps of vascular plant distributions in Denmark in the early 20th century |
| url | https://doi.org/10.5281/zenodo.17600480 |