Rethinking Camera Choice: An Empirical Study on Fisheye Camera Properties in Robotic Manipulation
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866917306114768896 |
|---|---|
| author | Xue, Han Min, Nan Liu, Xiaotong Chen, Wendi Fang, Yuan Lv, Jun Lu, Cewu Wen, Chuan |
| author_facet | Xue, Han Min, Nan Liu, Xiaotong Chen, Wendi Fang, Yuan Lv, Jun Lu, Cewu Wen, Chuan |
| contents | The adoption of fisheye cameras in robotic manipulation, driven by their exceptionally wide Field of View (FoV), is rapidly outpacing a systematic understanding of their downstream effects on policy learning. This paper presents the first comprehensive empirical study to bridge this gap, rigorously analyzing the properties of wrist-mounted fisheye cameras for imitation learning. Through extensive experiments in both simulation and the real world, we investigate three critical research questions: spatial localization, scene generalization, and hardware generalization. Our investigation reveals that: (1) The wide FoV significantly enhances spatial localization, but this benefit is critically contingent on the visual complexity of the environment. (2) Fisheye-trained policies, while prone to overfitting in simple scenes, unlock superior scene generalization when trained with sufficient environmental diversity. (3) While naive cross-camera transfer leads to failures, we identify the root cause as scale overfitting and demonstrate that hardware generalization performance can be improved with a simple Random Scale Augmentation (RSA) strategy. Collectively, our findings provide concrete, actionable guidance for the large-scale collection and effective use of fisheye datasets in robotic learning. More results and videos are available on https://robo-fisheye.github.io/ |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_02139 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Rethinking Camera Choice: An Empirical Study on Fisheye Camera Properties in Robotic Manipulation Xue, Han Min, Nan Liu, Xiaotong Chen, Wendi Fang, Yuan Lv, Jun Lu, Cewu Wen, Chuan Robotics Computer Vision and Pattern Recognition I.2.9; I.4.8 The adoption of fisheye cameras in robotic manipulation, driven by their exceptionally wide Field of View (FoV), is rapidly outpacing a systematic understanding of their downstream effects on policy learning. This paper presents the first comprehensive empirical study to bridge this gap, rigorously analyzing the properties of wrist-mounted fisheye cameras for imitation learning. Through extensive experiments in both simulation and the real world, we investigate three critical research questions: spatial localization, scene generalization, and hardware generalization. Our investigation reveals that: (1) The wide FoV significantly enhances spatial localization, but this benefit is critically contingent on the visual complexity of the environment. (2) Fisheye-trained policies, while prone to overfitting in simple scenes, unlock superior scene generalization when trained with sufficient environmental diversity. (3) While naive cross-camera transfer leads to failures, we identify the root cause as scale overfitting and demonstrate that hardware generalization performance can be improved with a simple Random Scale Augmentation (RSA) strategy. Collectively, our findings provide concrete, actionable guidance for the large-scale collection and effective use of fisheye datasets in robotic learning. More results and videos are available on https://robo-fisheye.github.io/ |
| title | Rethinking Camera Choice: An Empirical Study on Fisheye Camera Properties in Robotic Manipulation |
| topic | Robotics Computer Vision and Pattern Recognition I.2.9; I.4.8 |
| url | https://arxiv.org/abs/2603.02139 |