JL1-CUP-2024-Second-Format: Bi-temporal Satellite Patches for Cropland Change Detection

Bi-temporal 256×256 RGB satellite image patches with per-pixel semantic maps and binary change labels. The dataset is aligned with the JL1 cropland change-detection benchmark's second track format and was published by author aodebiao-a on Hugging Face in May 2026.

Use Cases

• Training semantic segmentation models based on the provided per-pixel semantic maps.
• Developing change detection algorithms based on the aligned bi-temporal image patches and binary change map.
• Benchmarking model performance for cropland monitoring using the described JL1 benchmark layout.

Strengths

• Provides aligned bi-temporal image pairs (T1/T2) for change analysis.
• Includes per-pixel semantic maps and a binary change map for supervised learning.
• Follows a defined benchmark format (JL1-Second style) for consistent evaluation.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file formats are unknown, which may limit suitability assessment.

Provenance

Source

JL1 / Jilin-1 ecosystem, as referenced in the contest information and resources portal.

Collection Method

Likely derived from satellite imagery for a specific change detection contest.

Freshness

Last updated 2026-05-22 06:25:12; freshness should be verified.