videocameraposemegasammapper
September 8, 2026 · View on GitHub
Extract camera poses by leveraging MegaSaM and MoGe-2.
利用 MegaSaM 和 MoGe-2 提取相机位姿。
Type 算子类型: mapper
Tags 标签: gpu, video
🔧 Parameter Configuration 参数配置
| name 参数名 | type 类型 | default 默认值 | desc 说明 |
|---|---|---|---|
tag_field_name | <class 'str'> | 'video_camera_pose_tags' | The field name to store the tags. It's "video_camera_pose_tags" in default. |
frame_field | <class 'str'> | 'video_frames' | The field name where the video frames are stored. |
camera_calibration_field | <class 'str'> | 'camera_calibration' | The field name where the camera calibration info is stored. |
max_frames | <class 'int'> | 1000 | Maximum number of frames to save. |
droid_buffer | <class 'int'> | 1024 | DROID SLAM pre-allocated frame buffer size. Controls GPU memory usage — each buffer slot pre-allocates correlation volumes on GPU. Default 1024, sufficient for clips up to ~100 frames. Reduce for shorter clips to save VRAM, increase for longer videos. |
save_dir | <class 'str'> | None | Directory to save large numpy arrays (depth, cam_c2w) as .npy files instead of storing them inline. When set, tag_dict stores file paths (strings) instead of numpy arrays, which avoids memory limit. |
use_prepare_env | <class 'bool'> | False | Whether to prepare the environment. |
args | `` | extra args | |
kwargs | `` | extra args |