New SwissCrop25 benchmark dataset evaluates crop mapping models

A new benchmark dataset called SwissCrop25 has been released, designed to evaluate crop mapping models under realistic operational conditions. This dataset spans seven growing seasons and includes detailed crop taxonomies, land cover data, and environmental factors like temperature. When tested, the TSViT model demonstrated superior performance in crop mapping accuracy compared to U-TAE and Galileo, particularly in distinguishing rare crop classes and handling interannual shifts. AI

IMPACT This benchmark will drive the development of more robust and accurate crop mapping AI systems for operational use.

RANK_REASON The cluster describes a new benchmark dataset and evaluation of models, fitting the research category. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.CV →

• Sentinel-2
• SwissCrop25
• TSViT
• U-TAE

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →