Superpoint Transformers use superpoints as the basic processing units, thereby significantly reducing the number of tokens processed by Transformers. However, they have been seldom employed in point cloud roof plane instance segmentation, and existing superpoint Transformers suffer from limited performance due to the use of low-quality superpoints. To address this challenge, we establish a set of criteria that high-quality superpoints for Transformers should satisfy and introduce a corresponding two-stage superpoint generation process. The superpoints generated by our method not only have accurate boundaries, but also exhibit consistent geometric sizes and shapes, which greatly benefit the feature learning of superpoint Transformers. To compensate for the limitations of deep learning features when the training set size is limited, we incorporate multidimensional handcrafted features into the model. Additionally, we design a decoder that combines a Kolmogorov–Arnold Network with a Transformer module to improve instance prediction and mask extraction. Finally, our network’s predictions are refined using traditional algorithm-based post-processing. For evaluation, we annotated a real-world dataset and corrected annotation errors in the existing RoofN3D dataset. Experimental results show that our method achieves state-of-the-art performance on our dataset, as well as both the original and corrected RoofN3D datasets. Our model also shows significant advantages over existing methods when handling data with low point density, large density variations, or low 3D point precision. Moreover, it is not sensitive to plane boundary annotations during training, significantly reducing the annotation burden. We will release our code, trained models, and datasets.

Timely and accurate monitoring of surface water dynamics via remote sensing is critical, given water resources’ importance. However, accurate water body delineation based on high-resolution remotely sensed imagery is still challenging due to the complexity of water bodies’ boundaries and the diversity of their shapes and sizes, which can lead to boundary ambiguity and varying degrees of confusion with near-water vegetation in water body maps. To address this challenge, we introduce an uncertainty-aware hybrid CNN–Transformer model for delineating water bodies using remotely sensed imagery. In our designed network, a multi-scale transformer (MST) module is first designed to effectively model and refine the multi-scale global semantic dependencies of water bodies. Subsequently, an uncertainty-guided multi-scale information fusion (MSIF) module is constructed to extract water body mapping information from these multi-scale features output from the MST module and fuse them adaptively. Across different scales, the extracted features differ in their ability to distinguish water bodies from non-water bodies and in their levels of uncertainty. Consequently, during the adaptive fusion of multi-scale water body information in the MSIF module, the mapping uncertainty is quantified and suppressed to minimize its impact, thus yielding enhanced precision in water body delineation. Ultimately, a comprehensive loss function is designed for model optimization to generate the final water body map. Furthermore, to promote water body segmentation models’ development, this study also presents the HBD_Water water body sample dataset, which contains 44 multispectral, 5000 × 5000-pixel images at 2 m spatial resolution, and will be released on the LuojiaSET platform soon. Finally, to verify the proposed model and its constituent MST and MSIF modules, extensive water mapping experiments were performed on three datasets. The experimental results substantiate their effectiveness. Furthermore, comparative experiment results demonstrate that the proposed model performs better at water body extraction than advanced networks including TransUNet, DeeplabV3+, and ADCNN.

Waterbody extraction is of great significance for water resource investigation and monitoring. In addition to RGB bands, most common satellite images have a near-infrared (NIR) band. By combining these RGB-NIR bands, certain water, vegetation, and shadow indices can be calculated. The near-infrared band and these indices are very similar to the X modality in RGB+X data (common examples include RGB-D and RGB-Thermal). However, at present, no studies have thoroughly examined multimodal feature fusion from the RGB+X perspective in order to extract waterbodies with high precision. As a result, existing algorithms do not fully utilize satellite image information and have limited generalization ability. To overcome this limitation, we propose a dual-complexity backbone for waterbody extraction from the perspective of RGB+X data semantic segmentation. Its complex Transformer branch is used to extract RGB modality features, while its simple CNN branch is used to extract X modality features. This network structure can effectively capture multimodal, global, and local features in remote sensing images. It can also fully leverage the fact that the scale of RGB image datasets in computer vision is significantly larger than that of remote sensing waterbody extraction datasets. If a large pretrained model is used in the RGB branch, it is unnecessary to freeze the weights. Instead, both branches can be trained jointly, allowing the RGB branch to better adapt to the remote sensing waterbody extraction task without raising concerns that fine-tuning might undermine the pretrained model’s strong representation capability. We also propose two X modality configurations with strong generalization performance. To fully fuse multimodal features, we design a hybrid fusion module combining a CNN and a cross-attention mechanism. To integrate the multi-scale features, we employ a multi-scale Transformer structure in the RGB branch and design a multi-scale decoder. Our algorithm achieves state-of-the-art performance on the GID-5 dataset and competitive performance on the S1S2-Water dataset. Furthermore, it significantly outperforms existing methods in cross-dataset zero-shot transfer between the two datasets, with IoU/F1-score gains of 26.08%/27.33% on GID-5 and 38.74%/31.37% on S1S2-Water over previous SOTA methods. Our processing paradigm of modeling RGB-NIR remote sensing images as RGB+X data shows potential for generalization to other multi-modal remote sensing tasks. The dual-complexity backbone we design also has potential to be extended to other tasks that transfer large pretrained RGB models to remote sensing imagery with RGB-NIR four bands or even more spectral bands. We have open-sourced the code and trained models used in this research.