TEA : Temporal Adaptive Satellite Image Semantic Segmentation

Juyuan Kang, Hao Zhu, Yan Zhu, Wei Zhang, Jianing Chen, Tianxiang Xiao, Yike Ma, Feng Dai · 2026

Read TEA : Temporal Adaptive Satellite Image Semantic Segmentation by Juyuan Kang, Hao Zhu, Yan Zhu, Wei Zhang, Jianing Chen, Tianxiang Xiao, Yike Ma, Feng Dai (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-04956.txt.

Description

Crop maps from satellite image time series help farms and governments track what is planted where. Juyuan Kang and 8 co-authors present TEA, a segmentation method for such series, in a 2026 paper. Existing approaches assume a fixed sequence length of images, and the authors find their accuracy drops sharply when the number of observations varies, as happens with clouds, orbit schedules, or different regions.

TEA is built for temporal adaptivity, staying resilient when sequence length changes. Parcel segmentation, drawing field boundaries, is the essential step it performs. Robust crop mapping across variable image counts matters for yield forecasting and subsidy checks, where a model that only works on one fixed calendar of images is of limited real use.

Catalog record

CREATORJuyuan Kang, Hao Zhu, Yan Zhu, Wei Zhang, Jianing Chen, Tianxiang Xiao, Yike Ma, Feng Dai
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE51 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-04956

Browse the full library · About the Tower of Babel