Cloud detection is a key pre-processing step for optical satellite remote sensing images. Convolutional neural networks (CNNs) are the most widely used and powerful cloud detection techniques in recent years, featuring the fusion of different levels of feature maps. However, existing methods use feature cascading, element addition, or element multiplication to fuse directly, ignoring the similarities and differences between features and their contribution to the detected cloud regions, resulting in a large amount of noise being introduced into the fusion process. To solve this problem, this paper proposes a feature-aware aggregation network (FAANet), which can make full use of valuable information for cloud detection through feature enhancement and feature aggregation. The feature enhancement module (FEM) consists of a high-level feature enhancement module (HFEM) and a low-level feature enhancement module (LFEM). HFEM is used to enhance the advanced semantic features to supplement the cloud features lost during the recovery of high-resolution feature maps; LFEM is used to enhance the image spatial detail features to further refine the predicted cloud maps. In addition, a feature aggregation module (FAM) is designed to selectively aggregate coded block features, decoded block features, and enhanced semantic features through a new aggregation mechanism to compliment the features diluted by up-sampling and improve the integrity of the generated cloud maps. Experimental results on three open-source cloud detection datasets and self-built Landsat-8 cloud detection dataset with higher spatial resolution show that the proposed method achieves competitive performance under different evaluation metrics. The advantages of the method proposed in this paper for saliency image segmentation are verified and provide a reference for other computer vision tasks. Code is available at https://github.com/HaiLei-Fly/FAANet.

Cloud coverage and obstruction significantly affect the usability of remote sensing images, making cloud detection a key prerequisite for optical remote sensing applications. In existing cloud detection methods, using U-shaped convolutional networks alone has limitations in modeling long-range contexts, while Vision Transformers fall short in capturing local spatial features. To address these issues, this study proposes a dual-branch framework, TransCNet, which combines Transformer and CNN architectures to enhance the accuracy and effectiveness of cloud detection. TransCNet addresses this by designing dual encoder branches: a Transformer branch capturing global dependencies and a CNN branch extracting local details. A novel feature aggregation module enables the complementary fusion of multi-level features from both branches at each encoder stage, enhanced by channel attention mechanisms. To mitigate feature dilution during decoding, aggregated features compensate for information loss from sampling operations. Evaluations on 38-Cloud, SPARCS, and a high-resolution Landsat-8 dataset demonstrate TransCNet’s competitive performance across metrics, effectively balancing global semantic understanding and local edge preservation for clearer cloud boundary detection. The approach resolves key limitations in existing cloud detection frameworks through synergistic multi-branch feature integration.