TerraCodec用AI压缩卫星影像，速度超传统方法10倍。
TerraCodec: Compressing Optical Earth Observation Data
- 用Transformer模型捕捉多时相影像的时空依赖关系
- 在相同画质下比传统编码高3-10倍压缩率
- 可零样本修复云层遮挡，适合遥感数据处理
地球观测（EO）卫星产生海量多光谱影像时间序列，给存储与传输带来严峻挑战。现有学习型压缩方法分散且缺乏公开的大规模预训练编码器，且多数研究聚焦于图像压缩，对时间冗余及EO视频编码关注不足。为此，我们提出TerraCodec（TEC），一个基于哨兵-2数据预训练的神经编码器家族，包含高效的多光谱图像变体与利用时序依赖的时序变换器模型（TEC-TT）。为突破当前神经编码器固定码率限制，我们提出潜空间重排（Latent Repacking）方法，实现灵活码率下训练的变压器模型。TerraCodec在同等图像质量下实现3-10倍压缩率，优于经典编码器。此外，TEC-TT可实现零样本云层修复，在AllClear基准上超越现有最优方法。结果表明神经编码器是地球观测领域的有力方向。代码与模型已开源：https://github.com/IBM/TerraCodec。
原文摘要 · Abstract (English)
Earth observation (EO) satellites produce massive streams of multispectral image time series, posing pressing challenges for storage and transmission. Yet, learned EO compression remains fragmented and lacks publicly available, large-scale pretrained codecs. Moreover, prior work has largely focused on image compression, leaving temporal redundancy and EO video codecs underexplored. To address these gaps, we introduce TerraCodec (TEC), a family of learned codecs pretrained on Sentinel-2 EO data. TEC includes efficient multispectral image variants and a Temporal Transformer model (TEC-TT) that leverages dependencies across time. To overcome the fixed-rate setting of today's neural codecs, we present Latent Repacking, a novel method for training flexible-rate transformer models that operate on varying rate-distortion settings. TerraCodec outperforms classical codecs, achieving 3-10x higher compression at equivalent image quality. Beyond compression, TEC-TT enables zero-shot cloud inpainting, surpassing state-of-the-art methods on the AllClear benchmark. Our results establish neural codecs as a promising direction for Earth observation. Our code and models are publically available at https://github.com/IBM/TerraCodec.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。