In this paper, we study the semi-supervised semantic segmentation problem via limited labeled samples and a large number of unlabeled samples. We propose a self-learning semi-supervised approach for the semantic segmentation of high-resolution remote sensing images. Our approach uses two networks (UNet and DeepLabV3) to predict the labels of the same unlabeled sample, and the pseudo labels samples with high prediction consistency are added to the training samples to improve the accuracy of semantic segmentation under the condition of limited labeled samples. Our method expands training data samples by using unlabeled data samples with pseudo labels. In order to verify the effectiveness of the proposed method, some experiments were conducted on the improved ISPRS Vaihingen 2D Semantic Labeling dataset using the method that we proposed. We focus on the extraction of forest and vegetation information and focus on the impact of a large number of unlabeled samples on the precision of semantic segmentation, we combine water, surface, buildings, cars, and background into one category and named others, and we call the changed dataset the improved ISPRS Vaihingen dataset. The experimental results show that the proposed method can effectively improve the semantic segmentation accuracy of high-scoring remote sensing images with limited samples than common deep semi-supervised learning.

Classification and segmentation of point clouds have attracted increasing attention in recent years. On the one hand, it is difficult to extract local features with geometric information. On the other hand, how to select more important features correctly also brings challenges to the research. Therefore, the main challenge in classifying and segmenting the point clouds is how to locate the attentional region. To tackle this challenge, we propose a graph-based neural network with an attention pooling strategy (AGNet). In particular, local feature information can be extracted by constructing a topological structure. Compared to existing methods, AGNet can better extract the spatial information with different distances, and the attentional pooling strategy is capable of selecting the most important features of the topological structure. Therefore, our model can aggregate more information to better represent different point cloud features. We conducted extensive experiments on challenging benchmark datasets including ModelNet40 for object classification, as well as ShapeNet Part and S3DIS for segmentation. Both the quantitative and qualitative experiments demonstrated a consistent advantage for the tasks of point set classification and segmentation.

While transformers are powerful neural network architectures for feature learning, current Transformer-based approaches for semantic segmentation of high-resolution remote sensing images (HRRSIs) struggle with the extraction of local semantic features. To address this issue, we incorporate a hypergraph into the Transformer. Hypergraph-based methods are proficient at discovering high-order correlations within limited-scale data, extracting pertinent representations to enhance the Transformer’s learning capabilities. We also propose dual pooling and feature aggregation modules, inspired by the adaptive pooling’s potent local modeling capabilities, to additionally extract fine-grained features from HRRSIs. In particular, we conceive a Hypergraph Biformer (HGBT) based on these three proposed modules along with a Biformer backbone. HGBT has the potential to learn general latent features as well as generate high-order representations of HRRSIs by modeling correlations of multi-scale features and local topology within an entirely nonlinear space, leading to the aggregation of features in a compact and localized manner, enhancing the model’s ability to capture detailed variations within small areas. We validate our approach through extensive experiments on ISPRS Vaihingen and Potsdam datasets, where HGBT attains mIoU of 83.71% and 87.88%, respectively. Both quantitative and qualitative assessments underscore the dominance of HGBT. Our code will be accessible at https://github.com/ZhangIceNight/HGBFormer.

Effective forest tree species classification is critical for various application domains such as forest management, bio-diversity conservation, and ecological research. However, existing studies on tree species classification predominantly rely on high-cost and processing-intensive hyperspectral data, which limits practical applications on large scales. In this work, we focus on investigating the potential of cost-effective UAV RGB images for tree species classification in heterogeneous forests and propose a method that fully leverages the rich spatial, semantic, and visible spectral information of UAV RGB images. We propose an RSVMamba model, which incorporates improved Visual State Space Blocks and an AutoDownsampling module to enhance accuracy and stability, while paying particular attention to small objects in sparse spatial locations. The model achieves linear computational complexity while retaining the global receptive field, making it particularly suitable for processing high spatial-resolution images. Additionally, we collected UAV RGB images covering 40 km2 of subtropical forest in southern China. A meticulous evaluation on this data shows that our method achieves an overall accuracy of 84.28% for eight tree species, dead trees, and other broadleaves. We verify the superiority of our method through a series of comparative experiments on the collected and benchmark datasets. Our results affirm the usefulness of single-temporal UAV RGB images for tree species classification in heterogeneous forest environments. Furthermore, the proposed method bridges the gap between data accessibility and precision in tree species classification, broadening the boundaries of single-temporal UAV RGB images for practical forestry applications and providing a more cost-effective and time-flexible solution for this problem.

With the rapid development of the metaverse, AR/VR technology and consumer electronics have emerged as crucial drivers for future virtual experiences. In this context, realistic 3D scene modeling and rendering has become particularly critical. Neural Radiance Fields (NeRF), as a deep learning-based method, have made significant progress in the reconstruction and rendering of real-world 3D scenes. However, NeRF still faces challenges in efficiently handling high-frequency information within the object, leading to issues such as blurry details and artifacts. To address this issue and provide more immersive virtual experiences, we propose Vivid-NeRF, which offers benefits to both metaverse and AR/VR technologies. Vivid-NeRF extracts information at different frequencies from the image and fully exploits the high-frequency information in the 3D feature generation process to obtain more realistic details. In addition, we propose frequency-based sampling to increase the sampling of high-frequency components. Finally, we merge frequency information with viewpoint features obtained through frequency-based sampling to enhance the model’s capability in expressing scene details. With these improvements, Vivid-NeRF significantly reduces surface blur and accurately captures and reproduces the smooth surface appearance. We conduct experiments on the Blender and Shiny-Blender datasets. Experimental results demonstrate that Vivid-NeRF achieves PSNR of 35.00 and 34.01, SSIM of 0.976 and 0.970, and LPIPS of 0.033 and 0.064. Both quantitative and qualitative assessments demonstrate that our approach outperforms the previous state-of-the-art (SOTA) ABLE-NeRF.