Primary competition visual

GEOAI Challenge for Cropland Mapping in Dry Environments

Helping Uzbekistan, Russian Federation
1 000 CHF
Completed (6 months ago)
Classification
Earth Observation
617 joined
184 active
Starti
Jul 02, 25
Closei
Sep 29, 25
Reveali
Sep 29, 25
False Negatives in Training Data
Help · 14 Aug 2025, 18:08 · 4

Hola Amigos

While looking through the data I found some potentially mis-labeled data in the training set. Check out the IDs 451, 464, 473 and 477 from Orenburg_training_samples. The file labels them as 0 but from my view they are cropland. I crosschecked using ESAs WorldCover 10m v200 maps (for 2021).

I just checked a handfull of samples but this makes me wonder if there are also mislabeled datapoints in the test set.

Here an example:

ID: 451

Cropland: 0

POINT: (54.83862 52.03939)

Google Maps:

WorldCover cropland mask:

Discussion 4 answers
User avatar
CodeJoe

Doesn't really look like a cropland to me😅. How are you able to identify a cropland?

18 Aug 2025, 15:17
Upvotes 0

In ChatGPT we trust:

What’s Visible in the Image

  • Agricultural Field: The large, centrally placed area looks like a cultivated field—perhaps recently plowed soil. Its irregular brown tones and area patterns suggest farmland.
  • Surrounding Vegetation and Settlement: To the left side, there’s a built-up or residential area (visible as small clustered structures and roads). On the right, a band of greener landscape appears—a mixture of vegetation, possibly along a watercourse or riverbank.
  • Linear Features: A thin, perhaps vegetated strip or pathway (maybe a hedgerow, irrigation channel, or small tree line) divides the central field from the greener zone to the right.
  • Natural Terrain Variation: The field shows subtle tonal variations—lighter and darker patches—likely reflecting minor changes in soil moisture, texture, or tillage patterns.
User avatar
CodeJoe

I haven't done any form eda yet but are there several images that look like the mislabelled image?

Which IDs then if that's the case?

User avatar
CodeJoe

In ChatGPT we trust🤣