A Distributed Computing Framework for Satellite Swarms
本文提出一种基于CRDT的分布式计算框架，用于解决大规模卫星群的自主控制和容错问题，减少了地面与卫星间的通信量。
0 citationsRead paper
Selected work
Representative Papers
TriCCOT: Tri-part Convolutional Conformal Transformer for Onboard Space Object Detection
TriCCOT通过结合卷积区域提议网络、保形预测阶段和硬件友好的注意力分类器Aper-GATES，解决了在计算资源有限的情况下进行空间物体检测的问题。
0 citationsRead paper
Can multimodal representation learning by alignment preserve modality-specific information?
This study investigates whether spatial alignment in multimodal representation learning degrades modality-specific information—particularly in remote sensing fusion of heterogeneous sources (e.g., optical and SAR). We first establish a theoretical analysis framework revealing how alignment operations inherently erode modality-unique semantic content. To address this, we propose a self-supervised contrastive learning paradigm that jointly optimizes semantic alignment and modality fidelity. Extensive experiments on real-world remote sensing datasets demonstrate that aggressive spatial alignment improves cross-modal consistency but substantially compromises modality-discriminative feature representation. Our method preserves alignment performance while boosting modality-specific representation capability by 12.7% (average improvement). The work provides an interpretable trade-off principle between alignment and specificity for multimodal remote sensing fusion and releases open-source code and a benchmark dataset.
0 citationsRead paper
Joint attitude estimation and 3D neural reconstruction of non-cooperative space objects
This work addresses the challenging problem of pose estimation and 3D reconstruction for non-cooperative space targets under severe conditions—including monocular imaging, unknown initial pose, limited field-of-view, and non-Lambertian illumination. To this end, we propose an end-to-end framework that jointly optimizes a neural radiance field (NeRF) and camera poses. Our method integrates differentiable rendering with frame-wise pose refinement and introduces a rotation-consistency regularizer to mitigate pose ambiguity and reconstruction degradation caused by sparse observations. Compared to conventional sequential optimization pipelines, our approach achieves higher geometric reconstruction accuracy and more stable pose convergence on synthetic space imagery. It significantly enhances robustness of 3D perception and improves spatial situational awareness under weakly supervised conditions.
0 citationsRead paper
Recent publications
Latest Papers
A Distributed Computing Framework for Satellite Swarms
本文提出一种基于CRDT的分布式计算框架，用于解决大规模卫星群的自主控制和容错问题，减少了地面与卫星间的通信量。
0 citationsRead paper
TriCCOT: Tri-part Convolutional Conformal Transformer for Onboard Space Object Detection
TriCCOT通过结合卷积区域提议网络、保形预测阶段和硬件友好的注意力分类器Aper-GATES，解决了在计算资源有限的情况下进行空间物体检测的问题。
0 citationsRead paper
Can multimodal representation learning by alignment preserve modality-specific information?
This study investigates whether spatial alignment in multimodal representation learning degrades modality-specific information—particularly in remote sensing fusion of heterogeneous sources (e.g., optical and SAR). We first establish a theoretical analysis framework revealing how alignment operations inherently erode modality-unique semantic content. To address this, we propose a self-supervised contrastive learning paradigm that jointly optimizes semantic alignment and modality fidelity. Extensive experiments on real-world remote sensing datasets demonstrate that aggressive spatial alignment improves cross-modal consistency but substantially compromises modality-discriminative feature representation. Our method preserves alignment performance while boosting modality-specific representation capability by 12.7% (average improvement). The work provides an interpretable trade-off principle between alignment and specificity for multimodal remote sensing fusion and releases open-source code and a benchmark dataset.
0 citationsRead paper
Joint attitude estimation and 3D neural reconstruction of non-cooperative space objects
This work addresses the challenging problem of pose estimation and 3D reconstruction for non-cooperative space targets under severe conditions—including monocular imaging, unknown initial pose, limited field-of-view, and non-Lambertian illumination. To this end, we propose an end-to-end framework that jointly optimizes a neural radiance field (NeRF) and camera poses. Our method integrates differentiable rendering with frame-wise pose refinement and introduces a rotation-consistency regularizer to mitigate pose ambiguity and reconstruction degradation caused by sparse observations. Compared to conventional sequential optimization pipelines, our approach achieves higher geometric reconstruction accuracy and more stable pose convergence on synthetic space imagery. It significantly enhances robustness of 3D perception and improves spatial situational awareness under weakly supervised conditions.
0 citationsRead paper