Search papers, labs, and topics across Lattice.
This paper introduces KSG-Net, a novel network designed for accurate 3D ship detection in maritime environments by addressing the challenges of weak feature representation for small vessels and limited global structural modeling for larger vessels. The method employs a Key Sparse Multi-scale Aggregation (KSMA) module to enhance the representation of small vessels and a Global Context Aggregation (GCA) module to capture long-range geometric dependencies, significantly improving detection performance. Experimental results on the Thames River vessel dataset reveal that KSG-Net outperforms existing detection methods, demonstrating robust performance in complex maritime conditions.
KSG-Net achieves unprecedented accuracy in detecting both small and large vessels in cluttered maritime environments by integrating key-sparse and global-context learning techniques.
Accurate 3D ship detection in maritime environments is critical for autonomous navigation, yet remains challenging due to large-scale vessel variations, sparse point clouds of small vessels, and severe sea-clutter interference. Existing methods, primarily based on 2D features or dense representations, struggle to balance detection accuracy and computational efficiency, while sparse 3D detectors designed for road scenes generalize poorly to maritime scenarios. This paper focuses on two key challenges in maritime LiDAR perception: weak feature representation for small and sparse vessels, and insufficient global structural modeling for large vessels due to the limited receptive field of local sparse convolutions. To address these issues, we propose KSG-Net, a Key-Sparse and Global-Context learning network for maritime 3D ship detection. The core idea is to jointly enhance local discriminative features and global structural awareness within a unified fully sparse detection framework. Specifically, a Key Sparse Multi-scale Aggregation (KSMA) module is designed to enhance the representation of small and sparse vessels by selecting informative key voxels and aggregating cross-scale neighborhood features. Furthermore, a Global Context Aggregation (GCA) module is introduced to capture long-range geometric dependencies through scene-level context modeling with gated residual interactions, thereby improving the representation of large vessels. Extensive experiments on the Thames River vessel dataset and simulated datasets demonstrate that KSG-Net consistently outperforms existing methods in multi-scale vessel detection and exhibits strong robustness in complex maritime environments.
