As spatial resolution increases, the information conveyed by remote sensing images becomes more and more complex. Large-scale variation and highly discrete distribution of objects greatly increase the challenge of the semantic segmentation task for remote sensing images. Mainstream approaches usually use implicit attention mechanisms or Transformer modules to achieve global context for good results. However, these approaches fail to explicitly extract intra-object consistency and inter-object saliency features leading to unclear boundaries and incomplete structures. In this paper, we propose a Category-Guided Global-Local Feature Interaction Network (CGGLNet), which utilizes category information to guide the modeling of global contextual information. To better acquire global information, we proposed a Category-Guided Supervised Transformer module (CGSTM). This module guides the modeling of global contextual information by estimating the potential class information of pixels so that features of the same class are more aggregated and those of different classes are more easily distinguished. To enhance the representation of local detailed features of multi-scale objects, we designed the Adaptive Local Feature Extraction Module (ALFEM). By parallel connection of the CGSTM and the ALFEM, our network can extract rich global and local context information contained in the image. Meanwhile, the designed Feature Refinement Segmentation Head (FRSH) helps to reduce the semantic difference between deep and shallow features and realizes the full integration of different levels of information. Extensive ablation and comparison experiments on two public remote sensing datasets (ISPRS Vaihingen dataset and ISPRS Potsdam dataset) indicate that our proposed CGGLNet achieves superior performance compared to the state-of-the-art methods.

Hyperspectral image (HSI) classification is one of the hotspots in remote sensing, and many methods have been continuously proposed in recent years. However, it is still challenging to achieve high accuracy classification in applications. One of the main reasons is the lack of labeled data. Due to the limitation of spatial resolution, manual labeling of HSI data is time-consuming and costly, so it is difficult to obtain a large amount of labeled data. In such a situation, many researchers turn their attention to the study of HSI classification with small samples. Focusing on this topic, this paper provides a systematic review of the research progress in recent years. Specifically, this paper contains three aspects. First, considering that the taxonomy used in previous review articles is not well-developed and confuses the reader, we propose a novel taxonomy based on the form of data utilization. This taxonomy provides a more accurate and comprehensive framework for categorizing the various approaches. Then, using the proposed taxonomy as a guideline, we analyze and summarize the existing methods, especially the latest research results (both deep and non-deep models) that were not included in the previous reviews, so that readers can understand the latest progress more clearly. Finally, we conduct several sets of experiments and present our opinions on current problems and future directions.

Small objects are widely distributed on remote sensing images (RSIs), and most of them are achieved by super resolution (SR) reconstruction followed by detection. However, due to the independent training of the SR network and the detection network, the lack of interaction between them leads to the limited performance of small object detection (SOD). Furthermore, time accountability is increased since the SR task is performed before the detection task. To address these problems, we develop a new SOD network to improve the SOD performance in RSIs, which embeds the SR task into the SOD task. First, a channel attention weighting module is proposed before the backbone to assign weights to different channels of the input image, allowing the network to selectively focus on different channels. Secondly, a self-attention encoding module is designed between the backbone and neck to add self-attention weight to the features extracted from the backbone and enhance the feature representation ability of small objects. Finally, the SR perceptual branch and perceptual loss are designed so that the SR task and the detection task can be associated through the SR perceptual loss, and the SR perceptual branch can guide the backbone network to learn high-resolution features through joint training, thus improving the detection performance of small objects. In the inference phase, the SR perceptual branch has been removed to improve the speed. Extensive experimental results on VEDAI and DOTA datasets show that the proposed method achieves an accuracy of 82.63% and 78.63%.

The latest development of multi-robot collaborative systems has put forward higher requirements for multi-sensor fusion localization. Current position methods mainly focus on the fusion of the carrier’s own sensor information, and how to fully utilize the information of multiple robots to achieve high-precision positioning is a major challenge. However, due to the comprehensive impact of factors such as poor performance, variety, complex calculations, and accumulation of environmental errors used by commercial robots, the difficulty of high-precision collaborative positioning is further exacerbated. To address this challenge, we propose a low-cost and robust multi-sensor data fusion scheme for heterogeneous multi-robot collaborative navigation in indoor environments, which integrates data from inertial measurement units (IMUs), laser rangefinders, cameras, and so on, into heterogeneous multi-robot navigation. Based on Discrete Kalman Filter (DKF) and Extended Kalman Filter (EKF) principles, a three-step joint filtering model is used to improve the state estimation and the visual data are processed using the YOLO deep learning target detection algorithm before updating the integrated filter. The proposed integration is tested at multiple levels in an open indoor environment following various formation paths. The results show that the three-dimensional root mean square error (RMSE) of indoor cooperative localization is 11.3 mm, the maximum error is less than 21.4 mm, and the motion error in occluded environments is suppressed. The proposed fusion scheme is able to satisfy the localization accuracy requirements for efficient and coordinated motion of autonomous mobile robots.

Due to differences in acquisition, domain shift is prevalent between different hyperspectral scene, and conventional classification models usually have poor generalization performance in cross-scene case. Domain generalization allows models trained in the source domain to be applied directly to the unseen domain, thus it is an important way to address cross-scene classification problems. Most of the existing methods eliminate domain shift by aligning the output space. However, due to the possible loss of information in the output space, enforcing consistency of distributions in the output space alone is not enough to ensure that distributions between different domains are correctly aligned. To alleviate the above problems, we propose a two-stage domain alignment single-source domain generalization network for cross-scene hyperspectral images classification based on generative domain adversarial networks. First, to reduce the negative impact of noise in data cube on pseudo domain generation, a spectral learning branch is introduced to guide generator training. Meanwhile, to enhance the ability of the discriminator to extract joint spatial-spectral features, a pyramid feature fusion network with double projection heads is proposed. Finally, combining the double projection head of the discriminator, we propose a two-stage domain alignment loss function to better eliminate domain shift. Several experiments on three public datasets demonstrate that the proposed method has competitive advantage in classification. The code will be available at: https://github.com/XiaozhenWang-NUAA/TSDAnet .