Fabric defect detection is a crucial step in the textile manufacturing process and serves as a key factor in ensuring high-quality fabric production. However, fabric defect images exhibit significant diversity, with defects often complex and irregularly distributed. Existing fabric defect detection algorithms face challenges such as low detection accuracy and slow processing speeds. In this paper, we propose a novel fabric defect detection algorithm based on YOLOv11, named YOLO-FEE. To reduce model parameters and computational overhead, we integrate a FasterNet module into YOLOv11. To enhance the model’s ability to represent discriminative features, we design the EMA module, which leverages a multi-scale attention mechanism to effectively capture detailed information at various scales. Furthermore, to improve the model’s stability and detection accuracy, we adopt the EIoU bounding-box loss function, which refines the IoU calculation method to more precisely evaluate the overlap between predicted and ground-truth bounding boxes. Experimental results demonstrate that even without applying quantization or other acceleration techniques, our approach shows significant potential to enhance the speed of inference and overall performance.

The spatiotemporal prediction of on-orbit satellites is crucial for intention understanding and ensuring the successful completion of missions. Current spatiotemporal prediction methods primarily use convolutional neural networks (CNNs) and recurrent neural networks (RNNs) to process sequential observation images to predict future states. However, these methods often result in poor prediction performance due to the network’s inherent limited ability to express complex details. In this work, a residual spatiotemporal diffusion (RSTD) model is proposed to learn the spatial and temporal characteristics of targets from spaceborne imaging sequences. Leveraging a historical database, the model utilizes the patterns of image feature changes to assist in predicting target shape variations during the next observation period. A spatiotemporal perception module, capable of capturing long-term dependencies, is incorporated into the denoising process, thereby endowing it with forecasting capabilities. Furthermore, by incorporating a residual dual-stream structure, the model separates the prediction of the target’s overall shape and dynamic changes, thus overcoming the issue of overly smooth predicted images. Comprehensive experiments demonstrate that the proposed method achieves a peak signal-to-noise ratio (PSNR) of about 35 dB during uniform and uniformly variable motion. It also outperforms existing methods in subsequent feature extraction and attitude estimation, supporting the spatiotemporal attitude prediction of on-orbit satellites.

In Earth observation, cloud severely affects the interpretation of optical satellites generated high-resolution images. Cloud-free optical images are vital for downstream tasks such as semantic segmentation and object detection. Thus, the elimination of clouds from optical imagery has emerged as a significant topic in remote sensing. Currently, most existing methods are proposed to leverage the texture information from auxiliary synthetic aperture radar (SAR) images to restore cloud-free images via direct channel merging. However, such a unified feature extraction approach often neglects the inherent distribution disparity between SAR and optical images—the result of differing imaging principles-potentially leading to significant feature loss. To this end, we introduce a network by jointing SAR and optical images multimodal and cross-attention network (MMCANet) to effectively extract multiscale contextual features from SAR imagery and integrate them with optical features. Specifically, instead of simple concatenation of the channels of SAR and optical images, we obtain high-dimensional features from them through independent feature extractors. The integration of these features is facilitated by a cross-attention mechanism that provides a more fine-grained amalgamation of information. Meanwhile, an atrous spatial pyramid pooling (ASPP) module is introduced into the integration of high-level features, which captures multiscale contextual information around clouded areas. In addition, we propose four advanced remote sensing image restoration algorithms that approach image restoration as a series of subtasks, gradually eliminating clouds to enhance performance. Comprehensive assessments show that MMCANet performs well on the SEN 12 MS-CR dataset with peak signal-to-noise ratio (PSNR) of 39.8871, structural similarity index (SSIM) of 0.9672, mean absolute error (MAE) of 0.0081, and spectral angle mapper (SAM) of 2.9884.

Measuring target angles is a crucial issue in airborne radar detection. Traditional dynamic programming track-before-detect (DP-TBD) often operates in low signal-to-noise ratio (SNR), posing challenges to angle measurement and localization. This letter presents an angle estimation technique for TBD processing. It interprets target angles across multiple frames using initial frame angle (IF-AG) and inter-frame angular rate of change (IF-ARC). Compensating for the latter and performing noncoherent accumulation to improve SNR before estimating the IF-AG enhance multiframe target angle estimation precision. The method initially formulates the cost function for multiframe angle measurement based on the raw data extracted from the multiframe array by DP-TBD. It subsequently estimates and compensates for IF-ARC, related to velocity, followed by IF-AG estimation, providing target angles for all frames concurrently. Simulation results validate the effectiveness of this method.

The recent flourish of deep learning on various tasks is largely accredited to the rich and high-quality labeled data. Nonetheless, collecting sufficient labeled samples is not very practical for many real applications. Few-shot Learning (FSL) provides a promising solution that allows a model to learn the concept of novel classes with a few labeled samples. However, many existing FSL methods are only designed for computer vision tasks and are not suitable for radio signal recognition. This paper calls for a radically different approach to FSL: in contrast to developing a new FSL model, we should focus on transforming the radio signal to be better processed by the state-of-the-art (SOTA) FSL model. We propose Modulated Signal Pre-transformation (MSP), a parameterized radio signal transformation framework that encourages the signals having the same semantics to have similar representations. MSP currently adapts to various SOTA FSL models for signal modulation recognition and can support the mainstream deep learning backbone. Evaluation results show that MSP improves the performance gains for many SOTA FSL models while maintaining flexibility.

Mode decomposition methods are the current workhorse for the analysis of non-stationary signals. However, current attempts at these methods mainly focus on improving accuracy, leaving computational efficiency untouched. To this end, we leverage the neural mode decomposition technique and propose an open-source Neural Mode Estimation (NME) to deliver a large speedup (at least 50×) while maintaining accuracy. Specifically, we transform the mode decomposition problem into an extremum problem of a functional in the cosine transform domain and train a neural network to approximate the solution. We demonstrate in extensive empirical results that NME can provide an improved trade-off between speed and accuracy, enabling fast, high-quality, stable mode decomposition of non-stationary signals.

We study the problem of object reconstruction in a multi-agent collaboration scenario. Specifically, we focus on the reconstruction of specific goals through several cooperative agents equipped with vision sensors to achieve higher efficiency than single agents. Our main insight is that a complete 3D object can be split into several local 3D models and assigned to different agents. In addition, we can use the salient characteristics of the collaboration agent itself to help realize the integration of local models. We develop a novel pipeline that first restores local 3D models from the images obtained from different agents, then the relative poses between collaborative agents are estimated by aligning intrinsic features. After that, all local models are integrated using the estimated parameters. Extensive experiments show that our proposed method is capable of accurately reconstructing 3D objects in the real world in a multi-agent collaborative manner. The full reconstruction pipeline is released to the public as an open-source project.

In Earth observation activities, cloud severely affects the interpretation of the high-resolution imagery, generated by optical satellites. Therefore, removing clouds from optical imagery becomes a topic of interest in the remote sensing field. Currently, most methods use auxiliary Synthetic Aperture Radar (SAR) images to reconstruct optical images by merging SAR and optical images into a deep learning network. However, the speckle noise of the SAR image is not taken into the consideration during feature fusion processing, leading to the blurry edges in the reconstructed optical images. To get fine-grained optical images, we propose a novel cloud removal framework based on the edge fusion of SAR and optical images. Firstly, the edge feature of SAR images is extracted by the GRHED. As the prior knowledge, it can provide fine-grained edge information for subsequent reconstruction work. Then channels from three modal data are stacked to guide the reconstruction of optical images by exploiting their correlations and interactions. Furthermore, a structural similarity (SSIM) loss function is introduced to optimize the training network and improve the coherence of the image structure. Experimental results confirm its advantages on the SEN12MS-CR dataset.

Retrieval of attitude and geometry of satellite targets from inverse synthetic aperture radar (ISAR) images is an important but difficult task, because of the complex scattering phenomenon and motion-dependent projection mechanism. In this letter, we construct a novel component extraction network (CEN)-based oriented object detection to obtain the projection parameters of the target components from ISAR images, then combine this CEN with particle swarm optimization (PSO) algorithm to retrieve the 3-D attitude and geometry of the target components. This proposed method can be used to accurately estimate the attitude and geometry of the target. The simulation experiments show the effectiveness and superiority of this method.

Real-time attitude estimation of on-orbit spacecraft is a core task in various space applications. Most of the existing methods are based on long-term observation by high-resolution sensors, such as space-borne cameras and ground-based radars. However, when the observation period is limited, it is difficult to obtain target instantaneous attitude information by these methods. To achieve instantaneous attitude estimation from a single camera image, a U-Linked network (ULNet) is proposed in this work. The prior structural constraints of key points are used to reflect the relationship between three-dimensional (3D) target attitude parameters and two-dimensional (2D) images. In this way, target attitude estimation can be solved through the feature point regression when the large-perspective image dataset can be built. The simulation results confirm the feasibility of the proposed method. Besides, the estimation performance of the proposed method also is investigated under different imaging observation conditions.

Dynamic estimation of spin spacecraft is a challenge and plays a significant role in space situation awareness applications like potential space collision warning. Based on remote sensing technologies of laser and radar sensors, current methods almost adopt a match strategy to estimate the dynamic parameters of a particular target with the long-term measurement collection. These kinds of data-driven methods merely consider the inherent connection between the measured characters and target dynamic patterns, and can hardly be expanded to other spacecraft when the measurement collection is insufficient. Therefore, this article presents a novel approach to interpreting multiple-station inverse synthetic aperture radar (ISAR) images for the dynamic estimation of spin spacecraft. As a unique phenomenon of radar imaging, the imaging plane of ISAR observation not only depends on the change of the relative position between the target and radar, but also changes with the spin of the target. In order to decouple the target dynamic estimation from the determination of the imaging geometry, the angular diversity of multiple-station images is employed. The proposed algorithm deduces an explicit expression of target dynamic parameters under the imaging projection model of the multiple-station observation. By utilizing the chaotic grasshopper optimization algorithm (CGOA), it determines three crucial elements of the target spin motion with a two-step optimization, including instantaneous attitude, rotation shaft and rotation speed. Simulation experiments of a typical spin spacecraft, Tiangong-I (TG-I), illustrate the feasibility of the proposed method under different motion patterns.

As more and more satellites are launched into the space, dynamic estimation of spin satellites has become a critical component of the space situation awareness application. Some explored studies using exterior measurements from different sensors such as optical device and inverse synthetic aperture radar (ISAR) to estimate dynamic parameters of spin satellites. As a single sensor normally provides two-dimensional observation, three-dimensional estimations resulting from these algorithms are strictly related to the prior knowledge of targets characteristics. As a result, it is difficult to expand these methods to other satellites. In order to support the dynamic estimation of most spin satellites, this paper presents a novel dynamic estimation approach which employs synchronized optical-and-radar images. The optical-and-radar fusion strategy has demonstrated its superiority in image analysis field, and breaks down the dynamic estimation of spin satellites into two sub-problems: target attitude estimation and spin parameters estimation. In this work, the proposed algorithm deduces two explicit expressions of target dynamic parameters under the imaging projection model of the joint optical-and-radar observation. Through the particle swarm optimization (PSO), target dynamic parameters are determined in two stages. This paper presents some experiments illustrating the feasibility of the proposed method and subsequent conclusions, which reflect advantages of the joint optical-and-radar observation mode in image interpretation.

This paper proposes a novel approach to interpreting the satellite attitude based on inverse synthetic aperture radar (ISAR) images. In the conventional viewpoint, quadratic and higher order phase terms of ISAR imagery are regarded as negative factors causing the defocusing phenomenon. In this paper, we introduce how to apply quadratic phase coefficients to estimate target attitude from the ISAR imagery. A geometric projection model of ISAR imaging is built according to radar line of sight, and an explicit expression is also derived to connect target attitude parameters and the image defocusing property. With the accommodation of Broyden-Fletcher-Goldfarb-Shanno algorithm, spatial-variant quadratic phase coefficients together with attitude parameters are determined by an image contrast maximization. We also extend the proposed algorithm to multistatic ISAR applications, where the quadratic phase information lying in simultaneous multistatic ISAR images can be mined to enhance the performance of target attitude estimation. Experimental results illustrate the feasibility of the proposed algorithm.