Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2603.20778 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915894654926848 |
|---|---|
| author | Cheng, Xiaoya Wang, Long Liu, Yan Liu, Xinyi Tan, Hanlin Liu, Yu Zhang, Maojun Yan, Shen |
| author_facet | Cheng, Xiaoya Wang, Long Liu, Yan Liu, Xinyi Tan, Hanlin Liu, Yu Zhang, Maojun Yan, Shen |
| contents | We present PiLoT, a unified framework that tackles UAV-based ego and target geo-localization. Conventional approaches rely on decoupled pipelines that fuse GNSS and Visual-Inertial Odometry (VIO) for ego-pose estimation, and active sensors like laser rangefinders for target localization. However, these methods are susceptible to failure in GNSS-denied environments and incur substantial hardware costs and complexity. PiLoT breaks this paradigm by directly registering live video stream against a geo-referenced 3D map. To achieve robust, accurate, and real-time performance, we introduce three key contributions: 1) a Dual-Thread Engine that decouples map rendering from core localization thread, ensuring both low latency while maintaining drift-free accuracy; 2) a large-scale synthetic dataset with precise geometric annotations (camera pose, depth maps). This dataset enables the training of a lightweight network that generalizes in a zero-shot manner from simulation to real data; and 3) a Joint Neural-Guided Stochastic-Gradient Optimizer (JNGO) that achieves robust convergence even under aggressive motion. Evaluations on a comprehensive set of public and newly collected benchmarks show that PiLoT outperforms state-of-the-art methods while running over 25 FPS on NVIDIA Jetson Orin platform. Our code and dataset is available at: https://github.com/Choyaa/PiLoT. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_20778 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PiLoT: Neural Pixel-to-3D Registration for UAV-based Ego and Target Geo-localization Cheng, Xiaoya Wang, Long Liu, Yan Liu, Xinyi Tan, Hanlin Liu, Yu Zhang, Maojun Yan, Shen Computer Vision and Pattern Recognition We present PiLoT, a unified framework that tackles UAV-based ego and target geo-localization. Conventional approaches rely on decoupled pipelines that fuse GNSS and Visual-Inertial Odometry (VIO) for ego-pose estimation, and active sensors like laser rangefinders for target localization. However, these methods are susceptible to failure in GNSS-denied environments and incur substantial hardware costs and complexity. PiLoT breaks this paradigm by directly registering live video stream against a geo-referenced 3D map. To achieve robust, accurate, and real-time performance, we introduce three key contributions: 1) a Dual-Thread Engine that decouples map rendering from core localization thread, ensuring both low latency while maintaining drift-free accuracy; 2) a large-scale synthetic dataset with precise geometric annotations (camera pose, depth maps). This dataset enables the training of a lightweight network that generalizes in a zero-shot manner from simulation to real data; and 3) a Joint Neural-Guided Stochastic-Gradient Optimizer (JNGO) that achieves robust convergence even under aggressive motion. Evaluations on a comprehensive set of public and newly collected benchmarks show that PiLoT outperforms state-of-the-art methods while running over 25 FPS on NVIDIA Jetson Orin platform. Our code and dataset is available at: https://github.com/Choyaa/PiLoT. |
| title | PiLoT: Neural Pixel-to-3D Registration for UAV-based Ego and Target Geo-localization |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2603.20778 |