Skip to content
DIBULI ↗︎

Dibuli Prism-A4L

Multi-sensor synchronized capture device: 1–4 global-shutter cameras, 6-axis IMU

Prism-A4L cameras around an optional LiDAR, beside the capture box
Photo: DIBULI, product page. LiDAR shown is optional ↗︎
Cameras
1–4 × SC130GS, 1280×1024 MJPEGMaker ↗︎
Shutter
GlobalMaker ↗︎
Field of view
Calibrated 168° × 131°Measured
Baseline
85.3 mmMeasured
IMU
ICM-45686 by default, 6-axisMaker ↗︎
Output
USB 3.x; TF card; ROS bag exportMaker ↗︎

Maker values checked 2026-11-03 · 2 recordings · Data on Hugging Face

Multi-sensor capture box from Dibuli with four global-shutter fisheye cameras and a 6-axis IMU, solved from the maker's public example data at two exposures.

The 3D scene MVIS recovered from the maker's example data, 1,000 µs exposure.
Reprojection median0.18 px
cam0–cam1 baseline85.3 mm
cam0 td+2.06 ms
cam0 lever arm71.6 mm

Reprojection is how far, in pixels, the detected corners lie from where the solve puts them. td is the camera–IMU time offset, and the lever arm is the camera–IMU position offset.

No independent reference on this page. These values show what MVIS solved, not how close they are to the truth.

The path of the rig around the board, as MVIS solved it. To explore it:

  • Press Play or drag the bar to follow the path.
  • Use the view buttons to change the view.
  • Pick a camera card to put that camera on the frustum, the pyramid that shows its view.

Each card draws the board from the solved pose, with the corners MVIS detected. We publish no camera images.

The solved rig in 3D, to scale. Tap a sensor to see its values. In full screen you can rotate and zoom. Grid squares are 20 mm.

To scale, from above and the front

From above
xycam0cam1cam2cam368.1 mm88.9 mmbaseline 85.3 mm20 mm
From the front
yzcam0cam1cam2cam320 mm
CameraLever x, y, z (mm)Distance (mm)log(R) (rad)td (ms)Field of view
cam0+63.9, −26.2, +19.071.6[−1.217, +1.234, −1.195]+2.06168° × 131°
cam1−12.7, −63.7, +20.568.1[−0.005, +2.221, −2.213]+2.06167° × 131°
cam2+64.8, +48.6, +19.883.4[−1.216, +1.217, −1.200]+2.09167° × 130°
cam3−13.1, +85.5, +20.688.9[−1.564, +0.001, +0.006]+2.12168° × 130°

Solved by MVIS. Lever arms (camera–IMU position offsets) are in the IMU frame. Rotation is log(R), camera in the IMU frame.

Published by the maker
CamerasUp to 4 global-shutter, 1280×1024 MJPEG, 1 to 30 fps
SyncPPS
Measured in these recordings
IMU6-axis at 797.7 Hz in these recordings
LensFisheye, EUCM fit
Calibrated field of viewcam0 168° × 131° · cam1 167° × 131° · cam2 167° × 130° · cam3 168° × 130°
Recordings2 camera + IMU bags at 1,000 and 2,000 µs exposure, about 450 MB each
SourceMaker's public download, files.dibuli.com
BoardAprilGrid 8×8, tag 70 mm, spacing 0.3
LicenceCC BY-NC 4.0; commercial use by request
HostedCalibench community, dibuli-prism-a4l/ (access by automatic approval)
AprilGrid 8 × 8, 707 × 707 mm. Standard pattern, to scale
Lens Calibrated field of view

Calibrated field of view, horizontal × vertical. cam0: 168° × 131°, corners reach 93° from the axis · cam1: 167° × 131°, corners reach 97° from the axis · cam2: 167° × 130°, corners reach 91° from the axis · cam3: 168° × 130°, corners reach 97° from the axis

0°15°30°45°60°75°90°0.00.51.01.52.0Ray angle from the optical axisImage radius / f

cam0 cam1 cam2 cam3

We read these from the calibrated lens fit at the image edges, through the principal point (where the optical axis meets the image). ~ marks an axis past where the corners reached, so the fit is extrapolated there. Past the fit means an edge lies beyond where the fit holds, so we give no value.

IMU Gyro–accel time offset t_ga −0.09 ms

Each bar shows a scale as its deviation from 1. IMU model: kalibr. t_ga, the time offset between the gyroscope and accelerometer channels: −0.09 ± 0.02 ms.

Gyro scale x+0.84 %
Gyro scale y+0.94 %
Gyro scale z−0.42 %
Accel scale x+0.46 %
Accel scale y+0.41 %
Accel scale z−0.14 %
−8 %0+8 %
Timing cam0 td +2.06 ms

The time offsets, to scale. td is the camera–IMU time offset, with t_imu = t_cam + td.

-10 ms-5 ms+0 ms+5 ms+10 mscam0 td+2.06cam1 td+2.06cam2 td+2.09cam3 td+2.12t_ga−0.09
Full results Every solved parameter

Rotation is log(R) in rad, camera in the IMU frame.

ParameterMVIS
cam0
fx px462.89
fy px462.45
cx px637.11
cy px523.22
alpha0.65834
beta1.07644
Lever x mm+63.90
Lever y mm−26.16
Lever z mm+19.00
log(R) rad[−1.2169, +1.2344, −1.1950]
td ms+2.063
Reprojection px0.304
cam1
fx px463.25
fy px462.79
cx px622.48
cy px465.02
alpha0.66073
beta1.06531
Lever x mm−12.70
Lever y mm−63.68
Lever z mm+20.52
log(R) rad[−0.0046, +2.2214, −2.2130]
td ms+2.064
Reprojection px0.318
cam2
fx px464.90
fy px464.43
cx px603.71
cy px507.10
alpha0.65709
beta1.08000
Lever x mm+64.79
Lever y mm+48.63
Lever z mm+19.79
log(R) rad[−1.2160, +1.2165, −1.2003]
td ms+2.087
Reprojection px0.261
cam3
fx px463.89
fy px463.91
cx px637.27
cy px488.19
alpha0.65989
beta1.07393
Lever x mm−13.08
Lever y mm+85.54
Lever z mm+20.58
log(R) rad[−1.5643, +0.0005, +0.0058]
td ms+2.122
Reprojection px0.227
IMU
t_ga ms−0.09 ± 0.02
  • No independent reference on this page. These values show what MVIS solved, not how close they are to the truth.
  • td changes with exposure. The exact slope is still open.
  • EUCM camera model, Kalibr IMU model, t_ga estimated, camera intrinsics free, seeded from the camera-only fit. Batch of 2026-11-03. This solve uses physical camera 1 as its base, so cam0 and cam1 on this page are physical cam1 and cam0.