In recent years, weakly supervised semantic segmentation (WSSS) has garnered significant attention in remote sensing image analysis due to its low annotation cost. To address the issues of inaccurate and incomplete seed areas and unreliable pseudo masks in WSSS, we propose a novel WSSS method for remote sensing images based on the Siamese Affinity Network (SAN) and the Segment Anything Model (SAM). First, we design a seed enhancement module for semantic affinity, which strengthens contextual relevance in the feature map by enforcing a unified constraint principle of cross-pixel similarity, thereby capturing semantically similar regions within the image. Second, leveraging the prior notion of cross-view consistency, we employ a Siamese network to regularize the consistency of CAMs from different affine-transformed images, providing additional supervision for weakly supervised learning. Finally, we utilize the SAM segmentation model to generate semantic superpixels, expanding the original CAM seeds to more completely and accurately extract target edges, thereby improving the quality of segmentation pseudo masks. Experimental results on the large-scale remote sensing datasets DRLSD and ISPRS Vaihingen demonstrate that our method achieves segmentation performance close to that of fully supervised semantic segmentation (FSSS) methods on both datasets. Ablation studies further verify the positive optimization effect of each module on segmentation pseudo labels. Our approach exhibits superior localization accuracy and precise visualization effects across different backbone networks, achieving state-of-the-art localization performance.

Due to the special properties of synthetic aperture radar (SAR) images, they are widely used in maritime applications, such as detecting ships at sea. To perform ship detection in SAR images, existing algorithms commonly utilize convolutional neural network (CNN). However, the challenges in acquiring SAR images and the imaging noise hinder CNN in performing SAR ship-detection tasks. In this paper, we revisit the relationship between SAR expert features and network abstract features, and propose an expert-feature-based convolutional neural network (EFCNet). Specifically, we exploit the inherent physical properties of SAR images by manually extracting a range of expert features, including electromagnetic scattering, geometric structure, and grayscale statistics. These expert features are then adaptively integrated with abstract CNN features through a newly designed multi-source features association module, which improves the common CNN’s capability to recognize ship targets. Experiment results on the SSDD demonstrate that EFCNet outperforms general CNN approaches. Furthermore, EFCNet achieves comparable detection performance to baseline methods while utilizing only 70% of the data capacity, highlighting its efficiency. This work aims to reignite interest in leveraging expert features in remote sensing tasks and offers promising avenues for improved SAR image interpretation.