Abstract
This work focuses on national-scale land-use/land-cover (LULC) semantic segmentation using ALOS-2 single-polarization (HH) SAR data over Japan, together with a companion binary water detection task. Building on SAR-W-MixMAE self-supervised pretraining [1], we address common SAR dense-prediction failure modes, boundary over-smoothing, missed thin/slender structures, and rare-class degradation under long-tailed labels, without increasing pipeline complexity. We introduce three lightweight refinements: (i) injecting high-resolution features into multi-scale decoding, (ii) a progressive refine-up head that alternates convolutional refinement and stepwise upsampling, and (iii) an $\alpha$-scale factor that tempers class reweighting within a focal+dice objective. The resulting model yields consistent improvements on the Japan-wide ALOS-2 LULC benchmark, particularly for under-represented classes, and improves water detection across standard evaluation metrics.
Abstract (translated)
这项工作专注于使用ALOS-2单极化（HH）SAR数据在日本进行国家尺度的土地利用/土地覆盖（LULC）语义分割，同时包括一个伴生的二元水体检测任务。基于SAR-W-MixMAE自监督预训练[1]，我们解决了常见的SAR密集预测失败模式，如边界过度平滑、遗漏细长结构以及在长尾标签下罕见类别的性能下降问题，而无需增加管道复杂度。我们引入了三项轻量级改进：(i) 将高分辨率特征注入多尺度解码中；(ii) 一种逐步细化和上采样的交替进行的渐进式细化头部，以及 (iii) $\alpha$-缩放因子，用于调节在焦点+dice目标下的类重新加权。最终模型在日本全境ALOS-2 LULC基准测试中取得了持续性的改进，特别是在代表性不足的类别中，并且在标准评估指标下提高了水体检测性能。
URL
https://arxiv.org/abs/2601.15705
