chengyuanshu98/Grasp_both

Bimanual Storage Task — Aligned Ego + UMI Dataset

Bimanual tabletop manipulation dataset with synchronized ego (head-mounted) and UMI (wrist-mounted) cameras. 50 episodes of a two-handed storage/organization task, each ~32 seconds.

Task

双手收纳 (Bimanual Storage): An operator uses two FastUMI Pro grippers to pick, move, and place objects on a tabletop. A head-mounted ego camera records a continuous third-person overhead view of the entire workspace.

Scene

Property	Value
Table height above floor	72.5 cm (measured)
ArUco tag	DICT_4X4_50 ID=0, 144×144 mm, placed flat on table surface

Hardware

Device	Role	Sensor	Serial No.
XVISIO SeerSense DS80	Ego (head-mounted)	RGB fisheye + IMU + SLAM	250801DR48FP26008004
XVISIO SeerSense DS80	UMI left hand (wrist)	RGB fisheye + IMU + SLAM	250801DR48FP25002692
XVISIO SeerSense DS80	UMI right hand (wrist)	RGB fisheye + IMU + SLAM	250801DR48FP25002313
Robotiq 85	Gripper (x2)	Aperture encoder	—

All three cameras: 1280x1280 resolution, ~200° DFOV fisheye (SEUCM model), 60 fps RGB, 500 Hz SLAM, 500 Hz IMU.

Ground Truth for KILO

This dataset is designed to provide ground truth for validating KILO's height/gauge estimation pipeline. The following quantities are independently measured or computed with high confidence and can serve as ground truth:

Measured (physical measurement)

Quantity	Value	Source	KILO relevance
Table height	72.5 cm	Tape measure	Absolute surface height reference
ArUco tag size	144 mm	Printed at known scale	Scale anchor for vision-based pose estimation

Computed — high confidence (can be treated as ground truth)

Quantity	Value	Method	Precision	KILO relevance
UMI camera height above table	L: 38.9±2.7 cm, R: 35.8±2.4 cm	ArUco tag solvePnP, per-episode	±3 cm	Combined with table height (72.5 cm) → UMI camera absolute height above floor = ~108–111 cm
UMI inter-camera distance	49.5±2.7 cm	ArUco tag relative pose, per-episode	±3 cm	Inter-hand distance ground truth; validates KILO's inter-hand gauge
UMI camera-to-EE offset	~14 cm (Z direction)	Cross-validated: ArUco height − trajectory min Z, consistent across 4 independent sessions at ±0.35 cm	<1 cm	Converts camera-frame measurements to EE-frame; hardware constant for Robotiq 85 non-prismatic mount
Ego height above floor	135.7±8.8 cm, per-episode	Multi-frame triangulation floor plane fit (KILO §6d estimator #3), using ego SLAM poses + ORB features	±1 cm (cross-session)	The z-gauge ground truth; the quantity KILO's estimator ladder tries to recover

Derived (from the above)

Quantity	Formula	Value	KILO relevance
UMI EE height above table	camera_height − 14 cm	L: ~25 cm, R: ~22 cm	EE-space z for retargeting
UMI EE height above floor	EE_above_table + 72.5 cm	L: ~97 cm, R: ~94 cm	Absolute EE z in world frame
Ego height above table	ego_floor_height − 72.5 cm	~63 cm	Head-to-table distance

What KILO can test against this dataset

1. z-estimator validation: The ego height above floor (135.7 cm) is computed via the strongest non-contact estimator (§6d #3). KILO's weaker estimators (mono depth, anthropometric prior) can be benchmarked against it.
2. Contact-event estimator: UMI gripper-close events + merged trajectory Z give per-surface contact heights. With known table height (72.5 cm), these can be checked for absolute accuracy.
3. Inter-hand gauge: The ArUco-derived inter-camera distance (49.5 cm per episode) is the ground truth for KILO's inter-hand relative pose, independent of any SLAM drift.
4. Height conditioning: The per-episode ego height variation (115–158 cm, reflecting operator posture) provides natural height diversity for testing z-conditioned augmentation.
5. Surface-cluster detection: Multiple grasp/release events per episode at known table height (72.5 cm) enable testing the contact-event surface clustering logic.

Dataset Structure

aligned_dataset/
├── README.md
├── dataset_meta.json
│
├── episode_01/
│   ├── alignment.json              # Timing + spatial alignment + ground truth
│   │
│   ├── ego/                        # Head-mounted camera data (sliced per episode)
│   │   ├── video.mp4               # RGB video, 1280x1280, 60 fps
│   │   ├── timestamps.csv          # Per-frame timestamps (unix + steady_clock)
│   │   ├── slam_poses.txt          # 6-DOF pose trajectory, TUM format, ~500 Hz
│   │   ├── imu.txt                 # 6-axis IMU (gyro + accel), ~500 Hz
│   │   ├── transform/
│   │   │   ├── left_hand/          # MediaPipe hand tracking, 21 joints
│   │   │   │   ├── 00_wrist.csv
│   │   │   │   ├── 01_thumb_cmc.csv
│   │   │   │   ├── ...
│   │   │   │   └── 20_pinky_tip.csv
│   │   │   └── right_hand/         # Same 21 joints
│   │   │       └── ...
│   │   ├── calibration/
│   │   │   ├── rgb_intrinsic.json  # SEUCM fisheye camera model
│   │   │   └── rgb2imu_extrinsic.json
│   │   └── metadata.json           # SDK/firmware versions
│   │
│   └── umi -> /path/to/session     # Symlink to UMI session data
│       ├── left_hand_.../
│       │   ├── RGB_Images/
│       │   │   ├── video.mp4       # Wrist RGB, 1280x1280, 60 fps
│       │   │   └── timestamps.csv
│       │   ├── SLAM_Poses/
│       │   │   ├── slam_processed.txt   # Raw SLAM, TUM format, ~500 Hz
│       │   │   └── slam_raw.txt
│       │   ├── Merged_Trajectory/
│       │   │   └── merged_trajectory.txt  # Fused trajectory (EE frame)
│       │   └── Clamp_Data/
│       │       └── clamp_data_tum.txt     # Gripper aperture, ~200 Hz
│       ├── right_hand_.../          # Same structure
│       ├── relative_transforms_left_to_right.txt
│       ├── relative_transforms_right_to_left.txt
│       └── quality_report/
│
├── episode_02/
├── ...
└── episode_50/

Data Channels

Ego (head-mounted)

Channel	Format	Rate	Description
RGB video	MP4, 1280x1280	60 fps	Fisheye overhead view of workspace
SLAM poses	TUM text (t tx ty tz qx qy qz qw)	~500 Hz	6-DOF IMU pose, gravity-aligned frame
IMU	Text (t gx gy gz ax ay az)	~500 Hz	Raw gyroscope + accelerometer
Hand tracking	CSV per joint, 21 joints x 2 hands	Sparse	MediaPipe 3D hand pose in camera frame
Intrinsics	JSON	Static	SEUCM model: fx=389.56, fy=389.63, u0=636.35, v0=641.69, α=0.677, β=0.753
Extrinsics	JSON	Static	RGB camera to IMU transform

UMI (wrist-mounted, per hand)

Channel	Format	Rate	Description
RGB video	MP4, 1280x1280	60 fps	Fisheye wrist view
SLAM poses	TUM text	~500 Hz	Raw SLAM in IMU frame
Merged trajectory	TUM text	~500 Hz	Fused trajectory in EE (gripper tip) frame
Gripper aperture	TUM text (t value)	~200 Hz	Robotiq 85 opening angle (degrees)
Relative transforms	TUM text	~500 Hz	Left-to-right and right-to-left transforms
Quality report	JSON	Per-episode	Kinematic validation results

Alignment

Ego and UMI data are recorded on separate hosts. Alignment uses wall-clock timestamps:

• Ego: record_unix_time_ms in timestamps.csv (UNIX epoch, millisecond precision)
• UMI: collection_summary.txt start time (wall-clock, second precision) + header_stamp (steady_clock) offset

Alignment precision: ~1 second. Each ego clip includes a ±1 second buffer around the UMI session window.

The alignment.json per episode contains the full timing correspondence and a ego_steady_to_unix_offset for converting between ego's steady_clock and UNIX timestamps.

Spatial Measurements

Each episode's alignment.json includes initial-state spatial measurements:

UMI Initial State (ArUco tag method)

A DICT_4X4_50 ArUco tag (ID=0, 144mm) is placed on the table surface. Both UMI wrist cameras observe it in the first ~20 frames. Camera-to-tag pose via SEUCM unprojection + solvePnP gives:

Measurement	Mean	Std	Description
Left camera height above table	38.9 cm	±2.7 cm	Vertical distance, camera optical center to table
Right camera height above table	35.8 cm	±2.4 cm	Same
Inter-camera distance	49.5 cm	±2.7 cm	3D Euclidean distance between left and right cameras

Note: camera-to-table height, not EE-to-table. The camera is mounted ~14 cm above the gripper tip (Robotiq 85, non-prismatic configuration).

Ego Initial State (multi-frame triangulation)

Floor height estimated via KILO §6d estimator #3: ORB feature matching across frames with known SLAM poses → midpoint triangulation → gravity-orthogonal floor plane fit.

Measurement	Value	Description
Height above floor	135.7 ± 8.8 cm	Per-episode, varies with operator posture
Floor Z (SLAM frame)	-1.4817 m	Global constant for this recording session

The per-episode variation (115–158 cm) reflects real head-height changes as the operator leans forward over the table vs. stands upright between episodes.

Camera Model

All three cameras use the Extended Unified Camera Model (EUCM/SEUCM):

Projection of point (X, Y, Z):

d = sqrt(β(X² + Y²) + Z²)
u = fx · X / (α·d + (1-α)·Z) + cx
v = fy · Y / (α·d + (1-α)·Z) + cy

Unprojection of pixel (u, v) to ray:

mx = (u - cx) / fx
my = (v - cy) / fy
r² = mx² + my²
mz = (1 - β·α²·r²) / (α·√(1 - (2α-1)·β·r²) + 1 - α)
ray = normalize(mx, my, mz)

Ego intrinsics are per-device calibrated (in calibration/rgb_intrinsic.json). UMI intrinsics use example device parameters (see note in alignment.json).

Coordinate Systems

• SLAM frame: Gravity-aligned (Z up), origin at device boot position. Roll/pitch from IMU; x, y, z, yaw are gauge freedoms.
• UMI EE frame (merged trajectory): Origin at gripper tip. X = forward, Y = left/right, Z = up/down.
• Ego hand tracking: MediaPipe coordinates in camera optical frame (Z = depth from camera).

Statistics

	Value
Episodes	50 (episode_01 through episode_50)
Duration per episode	~32 s (avg)
Total duration	~27 min
Ego RGB frames per episode	~1,947
UMI RGB frames per episode (per hand)	~1,827
SLAM frames per episode	~16,200
IMU frames per episode	~16,270
Dataset size (ego data only)	2.7 GB
UMI data	Symlinked, not duplicated

Known Issues

• UMI intrinsics: ArUco-based measurements use example device parameters, not per-device calibration. Systematic bias estimated at a few cm.
• Alignment precision: Limited to ~1 second by UMI's collection_summary wall-clock resolution. Sub-second alignment possible via visual cross-correlation (ArUco tag appearance timing).
• Ego SLAM tracking gaps: Two periods (~37–43 min and ~82–87 min into the continuous ego recording) show frozen SLAM position, likely tracking loss. Episodes during those windows still have valid RGB/IMU but degraded SLAM.