PointCloud Visualizer & Correction Tool

This tool visualizes point cloud data and bounding boxes from .bin and .txt files using Open3D. It supports features such as zoom controls, frame-by-frame navigation, automatic frame progression, and saving corrected data with RANSAC plane fitting.

Prerequisites

Ensure that you have Conda installed on your machine. You can install Miniconda if you do not have Conda.

Setting Up the Environment

To set up the environment with the required dependencies, follow these steps:

1. Clone the Repository:

   git clone https://github.com/roboticvedant/lidar_ransac_visualize
   cd lidar_ransac_visualize

2. Create the Conda Environment:

   • The environment is defined in the environment.yml file. You can create the environment using:

     conda env create -f environment.yml

3. Activate the Environment:

   conda activate racecarenv

4. Verify Installation:

   • Verify the environment by checking the installed packages:

     conda list

Running the Visualizer

Command-Line Arguments

• -r, --raw_data_dir (default: RACECAR_DATA/data/): Path to the directory containing raw point cloud data.
• -f, --fixed_data_dir (default: RACECAR_DATA/correctedData/): Path to the directory where corrected data will be stored.
• -p, --play_animation: Enable automatic frame progression.
• -g, --generate: Generate corrected data and save it.
• -n, --no_ransac: Disable RANSAC visualization.
• -s, --start_idx (default: 0): The starting index for visualization.
• -i, --frame_interval (default: 200 ms): Time in milliseconds between frames for automatic progression.
• -l, --length_of_car (default: 5.0): Length of the car.
• -w, --width_of_car (default: 2.0): Width of the car.
• -hc, --height_of_car (default: 2.0): Height of the car.
• -a, --axle_from_center (default: 2.0): Distance from the center to the rear axle.
• -hr, --height_of_rear_axle (default: 0.5): Height of the rear axle from the ground.

Example Command

To run the script with default parameters:

python viewer.py -r path/to/raw_data -f path/to/fixed_data -p -g -s 100 -i 100

Key Commands During Visualization

• N: Load the next frame manually.
• I: Zoom in.
• O: Zoom out.
• V: Toggle between side and top views.
• S: Save the corrected bounding box to a text file.

Directory Structure

Ensure your data is structured as follows:

RACECAR_DATA/
  └── data/
      ├── cloud/    # Contains .bin point cloud files
      └── labels/   # Contains .txt label files

Example Usage in Python

visualizer = VisualizerSequence(
    raw_data_dir="RACECAR_DATA/data/",
    fixed_data_dir="RACECAR_DATA/correctedData/",
    start_idx=230,
    play_animation=True,
    generating_data=True,
    ransac_visualize_flag=True,
    frame_interval=200,
    length_of_car=5.0,
    width_of_car=2.0,
    height_of_car=2.0,
    axle_from_center=2.0,
    height_of_rear_axle=0.5
)
visualizer.run()

Special Thanks

• Special thanks to Pragyan Dahal in guiding and providing boiler plate code for visualizer