Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[Dataset Title/Name]: UKFields #245

Closed
1 task done
Spiruel opened this issue May 9, 2024 · 8 comments
Closed
1 task done

[Dataset Title/Name]: UKFields #245

Spiruel opened this issue May 9, 2024 · 8 comments
Assignees

Comments

@Spiruel
Copy link

Spiruel commented May 9, 2024

Contact Details

eesjb@leeds.ac.uk

Dataset description

The ukfields dataset is a publicly accessible Earth Engine asset of automatically delineated field boundaries spanning England, Wales, Scotland, and Northern Ireland.

The ukfields dataset contains field boundaries for the United Kingdom, derived from harmonic composites of Sentinel 2 imagery captured in 2021. The delineation process leveraged the Segment Anything Model (SAM) from Meta, ensuring efficient field segmentation at scale. The segmented fields have been masked to a 2021 Dynamic World composite of cropland.

Bancroft, S., & Wilkins, J. (2024). UKFields (1.0.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.11110206

Earth Engine Snippet if dataset already in GEE

var ukfields = ee.ImageCollection("users/spiruel/ukfields")

Enter license information

CC-BY-4.0

Keywords

fields, agriculture, UK, england, scotland, wales, northern-ireland

Code of Conduct

  • I agree to follow this project's Code of Conduct
@samapriya
Copy link
Owner

samapriya commented May 11, 2024

Hi @Spiruel
Couple of things I noticed

  • First of all this is a FeatureCollection not an Image Collection
  • There is some missing geometry features in the collection

Can you share the original files/geojson/shapefiles/CSV or whatever you produced. Also is this part of some research work or does it have any citation?

@samapriya
Copy link
Owner

I have reprocessed your feature collection and removed any feature with missing geometry. You can try the example code here

Let me know if this is part of some research or if it has a citation/DOI

@Spiruel
Copy link
Author

Spiruel commented May 13, 2024

Woops, didn't notice about the imagecollection.

Example code works well, thank you.

DOI for the dataset is 10.5281/zenodo.11110206

New example code with better colours here:

var table = ee.FeatureCollection("projects/sat-io/open-datasets/UK-FIELDS");

Map.centerObject(table.first())
var empty = ee.Image().byte();
var outline = empty.paint({
  featureCollection: table,
  color: 'random',
  width: 3
});
    
Map.addLayer(outline.randomVisualizer(), {opacity:0.5}, 'UK Fields')

Thanks!

@samapriya
Copy link
Owner

Great I wonder if there is a peer reviewed article this work is related to. Totally fine if not but I wanted to check since most datasets in the catalog relate to a peer reviewed article.

I can also add a note that said this is an implementation of SAM but the output field boundaries have not been validated or there isn't a peer reviewed work attached to this ? Let me know if my assumption is wrong. Happy to work with you on this

@Spiruel
Copy link
Author

Spiruel commented May 13, 2024

ukfields_os_compare_mask4

Hi,

Not peer reviewed but we might do something about that after we look at other regions + planet data. UKFields somewhat validated to other UK dataset - see attached. Happy to have a note mentioning this - and if we decided to update it further along the line we can also link it to a piece of peer reviewed work.

Many thanks

@samapriya
Copy link
Owner

This has now been completed and will be available this week. Thank you for contributing. Watch the changelog for links available soon :)

@Spiruel
Copy link
Author

Spiruel commented May 22, 2024

Thank you! 😊

@samapriya
Copy link
Owner

Here is a link to the page

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

No branches or pull requests

2 participants