The rapid growth in satellite and aerial remote sensing platforms has created a growing need for distributed remote sensing scene classification. Conventional centralized scene classification methods, which involve transmitting remote sensing data to a ground station for processing, encounter limitations in both transmission efficiency and data privacy. Federated learning (FL) has emerged as a promising approach by enabling terminals to collaboratively train models without exchanging raw data. However, the non-independent and identically distributed (Non-IID) nature of remote sensing data significantly impedes FL performance. To address these challenges, a federated framework with feature distillation (FD) (FedFD) is proposed for FL-based remote sensing scene classification. Specifically, FedFD facilitates collaborative training by aggregating model parameters from multiple terminals to the cloud, thereby optimizing a global model. To further alleviate the impact of Non-IID data, an innovative partial feature-sharing strategy based on FD is designed, which divides features into globally shared essential features and locally maintained supplementary features. Moreover, to cope with object and scene scale variation, the squeeze and excitation module and the pyramid pooling module are incorporated into the scene classification network to enhance multiscale feature extraction. Extensive experiments on the Northwestern Polytechical University Remote Sensing Image Scene Classification 45 (NWPU-RESISC45) dataset and University of California, Merced Land Use (UC-Merced) dataset, under varying numbers of terminals and Non-IID levels, validate the effectiveness and scalability of FedFD, and demonstrate its superior performance in FL-based remote sensing scene classification.
Azimuth ambiguity significantly degrades the quality of synthetic aperture radar images. Sub-look spectral analysis (SSA) is a common ambiguity-detection method, but its performance is limited by threshold sensitivity and the high correlation of specific ambiguities across sub-looks. To overcome these specific limitations, this paper proposes an improved detection method. It first increases the number of sub-looks and constructs a high-dimensional multi-look matrix to enrich the coherence differences between targets and ambiguities. Non-negative matrix factorization is then employed to decompose this matrix, effectively separating the coherent target components from the variably coherent ambiguity components without relying on predefined thresholds. Experimental results on real data demonstrate that the proposed improvements achieve superior azimuth-ambiguity-detection performance compared with conventional SSA methods.
Cross-domain feature fusion offers an approach to weak target recognition in complex sea environments. This paper proposes a distance metric learning-based method for weak target classification. The method first extracts three time-domain features and three frequency-domain features from radar echo signals. Then, the features are partitioned and mapped to low-dimensional subspaces using linear projection matrices. The squared Euclidean distance is used as a metric function to measure the similarity between samples, and supervised optimization is performed by introducing information from similar and dissimilar sample pairs. Next, the projection matrices of each group are jointly updated iteratively using the gradient descent method to achieve supervised feature fusion. Finally, the fused feature is input into an ensemble one-class support vector machine (EOCSVM) for classification. Verified by IPIX measured data, the proposed method can effectively improve the separability of targets and sea clutter and improve the classification ability of sea clutter and weak targets under short-time observation. The proposed method enhances the features correlation from different domains through metric learning and EOCSVM, which can effectively alleviate the sample imbalance problem between sea clutter and targets.
The rapid proliferation of unmanned aerial vehicles (UAVs) has increasingly posed significant challenges for airspace security, particularly under long-range and visually degraded conditions. Effective UAV recognition is thus critical, yet current methodologies typically depend on single-sensor inputs, such as infrared (IR) imaging and radio frequency (RF) analysis, which suffer inherent limitations in complex environments. Although multimodal sensing has been explored in UAV detection, the joint exploitation of IR imagery and RF signals for UAV type recognition remains largely underexplored. The structural heterogeneity between IR and RF features presents challenges for joint representation and decision-making, which remains underexplored in previous work. To address this gap, this paper proposes RF-IRSynNet, a multimodal UAV classification framework that integrates IR imagery and in-flight RF emissions to enhance recognition performance. In RF-IRSynNet, IR images are processed using YOLOv11 to detect UAV candidates and extract structured semantic features. Meanwhile, RF signals are modeled using reservoir computing, which efficiently encodes temporal and spectral dynamics via feature sequences. These modalities are fused through an adaptive confidence-weighted soft-voting strategy, dynamically balancing their contributions based on specific tasks. Experimental results demonstrate that RF-IRSynNet outperforms both unimodal baselines and existing multimodal approaches, achieving robust classification at long ranges. The framework maintains high accuracy even with reduced training data, indicating high efficiency for real-world UAV monitoring.
Extracting infrared small targets from heterogeneous backgrounds remains a challenging task, as these targets lack salient texture and morphological features while the backgrounds are cluttered with noise. Therefore, effectively extracting discriminative features is essential for achieving complete and accurate detection. To address this issue, this paper proposes an algorithmic framework based on robust principal component analysis (RPCA), specifically designed for infrared small target detection in complex backgrounds. First, infrared small target detection is formulated as a generalized RPCA problem. A discriminative and reconstructive dictionary is constructed using supervised learning. Next, by introducing an ideal regularization term, the infrared image is reconstructed without losing structural information, yielding a discriminative principal component representation with respect to the learned dictionary. Finally, the structural features of infrared small targets are reconstructed via sparse coding, thereby enabling the extraction of infrared small targets. Extensive experimental results demonstrate the effectiveness of the proposed method.
Accurately sensing the channel state of heterogeneous networks is key to matching users’ diverse service communication demands with the channel state, and is an effective way to improve the utilization efficiency of network resource. However, existing channel state perception methods are not suitable for heterogeneous network, and their perception performance is easily affected by interference uncertainty. In order to achieve channel state perception of heterogeneous networks, this paper adopts a centralized collaborative perception model, where each node obtains local channel state perception results based on statistical pulse parameters at the physical layer. In order to reduce the impact of interference on perception performance, this paper uses the Jousselme distance to quantify the degree of difference among nodes caused by interference. Using the average credibility as a threshold, nodes in the sensing area are classified. On this basis, the local perception results of each node are performed classification-based correction to improve the accuracy and reliability of channel state perception. Simulation results indicate that the proposed method has good adaptability for channel state perception in complex electromagnetic environments. The perception results can accurately reflect the actual channel state, which is conducive to improving the network throughput.
Aiming at the issues where traditional direction-of-arrival (DOA) estimation algorithms experience substantial performance degradation in low signal-to-noise ratio environments, and deep learning-based DOA estimation methods rely on massive training data with prolonged model training cycles, this paper proposes two efficient and high-precision DOA estimation methods based on ensemble learning. By formulating DOA estimation as a multi-label classification problem and leveraging the classification chain paradigm, data-driven models, classification chain-random forest (CC-RF) and classification chain-eXtreme gradient boosting (CC-XGBoost), are constructed, which are capable of handling multi-label classification tasks. To verify the effectiveness of the proposed methods, a multi-dimensional comparative experiment is designed to benchmark their performance against the traditional multiple signal classification (MUSIC) algorithm and a convolutional neural network (CNN) model. Experimental results indicate that in both single-source and multi-source scenarios, the proposed CC-RF algorithm exhibits excellent performance, achieving DOA estimation accuracy comparable to the MUSIC algorithm; in multi-source estimation scenarios, both proposed models demonstrate strong noise adaptability. Compared with the traditional MUSIC and CNN algorithms, the estimation error of the CC-XGBoost and CC-RF models is reduced by up to nearly 30 times while maintaining low time complexity, with the single estimation time reduced by approximately 90% compared to traditional methods. This study provides a technical pathway for DOA estimation in complex environments and holds significant application value in fields such as radar detection and wireless communication.
