Skip to content

Latest commit

 

History

History
62 lines (41 loc) · 1.81 KB

README.md

File metadata and controls

62 lines (41 loc) · 1.81 KB
Raw

Carla Sensor Visualization

Allows to visualize an ego vehicle's sensors using PyGame.

It is dynamic, being able to create new vehicle configurations based on Json files.

To Run the Project

  1. It is recommended to use a virtual environment with python 3.8.
  2. Install the requirements with pip install -r requirements.txt
  3. Run the CARLA server with the desired map
  4. Run the project with python main.py

Modules

  • main.py: Main file, controls the entire process
  • display.py: Contains the methods to display the sensors in a PyGame window
  • vehicle.py: Contains the methods responsible for creating the ego vehicle, attaching sensors to it, destroying it, and in future versions, controlling it
  • sensors.py: Contains classes for each sensor, with methods to attach them to the vehicle, and to get their data through callbacks
  • 'config.py': Contains option parameters for the simulation

These modules are coded to be extremely dynamic, allowing their integration with any other project.

Sensors

Available sensors:

  • RGB Camera
  • LiDAR
  • Radar
  • GNSS
  • IMU
  • Collision
  • Lane Invasion

Future sensors:

  • Semantic Segmentation Camera
  • Instance Segmentation Camera
  • Depth Camera
  • Lidar Semantic Segmentation
  • Obstacle Detection
  • Optical Flow Camera (AKA: Motion Camera)

The collision and lane invasion sensors are special in the way that they are only triggered when the vehicle collides with something or invades a lane, respectively. And their information is not displayed in the PyGame window, but in the terminal.

TODO

  • Maybe make a web client of this that lets you chose different vehicles and verify if they have sensors and all. Cool idea.

  • Add the rest of the available CARLA sensors

  • Make the window dynamic based on the amount of sensors