RECENT POSTS

2021 3D LiDAR Synthetic Data

Posted in Data Repository of ACES Mobility

---

3D LiDAR Synthetic Data

Training deep neural network algorithms for LiDAR based object detection require huge amount of labeled data and various scenario data.

So we use simulator for exporting training and testing data. This approach can generate critical scenarios.

Here are some information explaining how we have labeled the synthetic dataset for perception and formatted the data and sample data.

Data Labels      3D Lidar Labels

We provide 3D bounding box labels in synthetic lidar data for training and testing. The lidar labels are 3D 7-DOF bounding boxes in the vehicle frame.

The bounding boxes have zero pitch and zero roll. 

We comprised the label format based on KITTI labels and stored as .txt file

* Label data format description(Source : http://www.cvlibs.net/datasets/kitti/eval_object.php?obj_benchmark=3d)

#Values    Name      Description
----------------------------------------------------------------------------
   1    type         Describes the type of object: 'Car', 'Van', 'Truck',
                     'Pedestrian', 'Person_sitting', 'Cyclist', 'Tram',
                     'Misc' or 'DontCare'
   1    truncated    Float from 0 (non-truncated) to 1 (truncated), where
                     truncated refers to the object leaving image boundaries
   1    occluded     Integer (0,1,2,3) indicating occlusion state:
                     0 = fully visible, 1 = partly occluded
                     2 = largely occluded, 3 = unknown
   1    alpha        Observation angle of object, ranging [-pi..pi]
   4    bbox         2D bounding box of object in the image (0-based index):
                     contains left, top, right, bottom pixel coordinates
   3    dimensions   3D object dimensions: height, width, length (in meters)
   3    location     3D object location x,y,z in camera coordinates (in meters)
   1    rotation_y   Rotation ry around Y-axis in camera coordinates [-pi..pi]

                                           

Data Format     Coordinate Systems

This section explains the coordinate systems, as well as the format of the synthetic lidar data.

Vehicle frame

The x-axis is positive forwards, y-axis is positive to the left, z-axis is positive upwards.

Sensor frame 

The lidar sensor placed in the center of the vehicle loop and sensor frame has the z-axis pointing upward with the x/y plane depending on the lidar position.

Lidar Data

                                          We use virtual lidar sensor(VLS-128) in simulator and captured and synchronized at 10Hz.

Each point is stored with (x,y,z) coordinate and an additional intensity value and we stored as bin file.

     

Camera Data

One camera image is provied for JPEG format and size is 1280x720.

 * LK_CIL_ST Scenario scene (Image/PointCloud)

* Data List

• LK_CIL_ST : A Lane Keeping _ B Cut in from Left _ Straight Data
  
• LK_CIL_CU : A Lane Keeping _ B Cut in from Left _Curve Data
  
• LK_CIR_CU : A Lane Keeping _ B Cut in from Right _Curve Data
  

• LK_COL_CU : A Lane Keeping _ B Cut out from Left _Curve Data
• LK_COL_ST : A Lane Keeping _ B Cut out from Left _ Straight Data
• LK_COR_CU : A Lane Keeping _ B Cut out from Right _Curve Data
• LK_COR_ST : A Lane Keeping _ B Cut out from Right _ Straight Data
• LK_LF_ST : A Lane Keeping _ B Lane Following _ Straight Data

---

Posted in Data Repository of ACES Mobility by Prof. Bongsob Song

Topics : Engineering, ACES Data Repository of ACES Mobility, 3D Object Detection

 June 14, 2021