3 July 2023 in IEEE Transactions on Aerospace and Electronic Systems
Attitude estimation for space targets based on inverse synthetic aperture radar (ISAR) is essential for various on-orbit operations, including spacecraft status analysis and fall forecast. However, traditional attitude estimation methods based on ISAR images suffer from poor feature extraction performance. Furthermore, the present methods can only deal with stabilized spacecraft. In situations where a non-cooperative spacecraft is in a slow-spinning state, the ISAR cross-range scaling is inaccurate, resulting in incorrect attitude estimation. In this study, we propose a cross-range scaling and attitude estimation method that can be applied to slow-spinning spacecraft. The method transforms attitude estimation to a parameter optimization task, where the optimization model aims to recover the cross-range scaling factor and attitude. Firstly, the slow-spinning state is parameterized and the optimization model of cross-range scaling and attitude estimation is established. Secondly, semantic features are extracted from ISAR images by the proposed multi-scale semantic feature extraction network. The optimization model is then solved through semantic features using the proposed robust estimation algorithm. Ultimately, the correct cross-range scaling factor and attitude of corresponding ISAR image are acquired. This paper makes three contributions: 1. A specialized mathematics model to perform cross-range scaling and attitude estimation of slow-spinning model is proposed. 2. By substituting traditional feature points with semantic features, the reliability of ISAR feature extraction is improved. 3. The proposed robust estimation algorithm enhances the accuracy of parameter estimation. Experiments demonstrate that this method has higher cross-range scaling and attitude estimation accuracy for slow-spinning spacecraft than usual methods.
29 January 2024 in IEEE Geoscience and Remote Sensing Letters
Instantaneous attitude estimation for spacecraft through optical/ISAR fusion observation is a hot topic in space situational awareness. However, due to the imaging characteristics and domain gap between the two data types, it is difficult to autonomously estimate instantaneous attitude based on data fusion. In this paper, an automatic optical/ISAR multi-feature fusion instantaneous attitude estimation framework is proposed. We deploy semantic feature extraction networks to extract and associate two kinds of robust features in optical and ISAR images, and an attitude estimation algorithm considering cross-range scaling based on multiple features combination is proposed. Meanwhile, the inherent structure constraints of spacecraft are introduced. Simulation experiments demonstrate that the proposed framework achieves an average attitude estimation accuracy of 1.87°, while also exhibiting excellent robustness.
5 March 2024 in Chinese Journal of Aeronautics
10 July 2025 in IEEE Sensors Journal
Attitude estimation of noncooperative spacecraft based on monocular camera is a crucial technique in on-orbit servicing missions. Most of the existing methods rely on the known 3D model of the target, or require a large number of observation images with ground truth labels, which are not applicable to unseen spacecraft lacking such prior knowledge. In this paper, we present a two-stage framework for semantic feature extraction of spacecraft typical components and on-orbit attitude estimation of unseen targets to solve the above problem, which inverts the three-dimensional attitude information from the two-dimensional axes of spacecraft typical components in the sequential observation images. Firstly, a spacecraft semantic feature extraction network (SSF-Net) is designed, which can learn the common semantic features of typical components in different spacecraft, thereby achieving good generalization for unseen targets and extracting their axes features. Then, we introduce homographic adaptation, the geometric constraint of semantic features, and dynamic constraint in sequential images to optimize false positives or missed detections of the extracted features under extreme observation perspectives. Finally, the axis reconstruction algorithm based on the RANdom Sample Consensus (RANSAC) is proposed to estimate the attitude of unseen on-orbit spacecraft. Simulation results confirm that the proposed method can effectively extract semantic features, with average pixel and angular errors of 6.93 pixels and 1.86 degrees respectively, and estimate the attitude of unseen spacecraft with typical component structures accurately, achieving the average estimation error of 3.25 degrees. Experiments also exhibit significant advantages compared to classical methods and excellent robustness under worse observing conditions.
30 March 2026 in IEEE Transactions on Aerospace and Electronic Systems
3D models of spacecraft can provide more detailed structural parameters than 2D inverse synthetic aperture radar (ISAR) images. Therefore, reconstructing a spacecraft's 3D structure from ISAR images is of great significance in space situational awareness. However, since spacecraft ISAR images are composed of discrete scattering points and exhibit characteristics such as weak texture, high dynamic range, and discontinuity, traditional factorization methods struggle to accurately match scattering points in ISAR images during 3D reconstruction. Although the energy accumulation method proposed in recent years can skip the scattering point matching step, it is still affected by the differences in energy distribution between different components in ISAR images. To tackle the aforementioned issues, this article introduces a spacecraft 3D reconstruction approach relying on component-level energy accumulation. First, an attention-based component segmentation network is designed to automatically segment the ISAR image sequence into cabin regions and solar panel regions. Secondly, a joint energy accumulation model for the cabin and solar panels is established, and this model is optimized using an adaptive particle swarm optimization (APSO) algorithm. Finally, the symmetry of the component structure is utilized to complete the missing regions arising from self-occlusion in the reconstructed point cloud. This method does away with the requirement for extracting and matching scattering points, and effectively solves the problem of uneven component density in the reconstructed point cloud. Experimental results verify the effectiveness, robustness, and superiority of the proposed method, demonstrating that it can provide efficient and reliable 3D model support for spacecraft on-orbit maintenance and fault diagnosis.