Sim2Real

CARLA Synchronized Dataset Generation

This repository contains code and instructions to generate a synchronized dataset using the CARLA simulator. The dataset includes data from the CARLA environment, such as sensor data and vehicle information, and is useful for various research and development purposes.

Prerequisites

Before using this code to generate a CARLA synchronized dataset, make sure you have the following prerequisites installed:

• CARLA Simulator: Download and install the CARLA simulator on your system.

Usage

1. Start CARLA Server:

   • Open a terminal and navigate to the CARLA directory where the CarlaUE4.sh script is located.
   • Start the CARLA server with the following command:

     ./CarlaUE4.sh --opengl -quality-level=Epic --sync

   • The -quality-level=Epic flag sets the graphics quality to Epic, --opengl ensures it runs with the OpenGL graphics backend, and sync ensures it runs in synchronous mode.

2. Generate Data:

   • Use the provided data_generator.py to generate data.

   Below are the available controls and their descriptions:

   Class IDs for Semantic Segmentation:

   • Class ID:
     • Description: Define the class IDs used for semantic segmentation of objects in the CARLA environment.
     • Example:

       Class ID : 
       1  == Roads
       2  == Sidewalks
       3  == Buildings
       6  == Street lights
       7  == Traffic lights
       8  == Traffic Signs
       9  == Trees / Grass
       11 == Sky 
       14 == Car (Vehicles)
       15 == Trucks (Vehicles)

   YOLO Class ID:

   • class_id_yolo:
     • Description: Specify the YOLO class ID used for object detection or recognition in the CARLA environment.

   Area Threshold Value:

   • area_val:
     • Description: Set the area threshold value to filter objects based on their size in the recorded images.

   Number of Vehicles to Spawn:

   • vehicle_count:
     • Description: Define the number of vehicles you want to spawn in the CARLA environment.

   Folder to Save Recorded RGB Images:

   • save_folder:
     • Description: Specify the folder where the recorded RGB images will be saved.

   Desired Cropped Image Dimensions:

   • cropped_image_height and cropped_image_width:
     • Description: Set the desired height and width dimensions for cropped images.

   Please refer to the data_generator.py script itself for detailed comments and instructions on how to use these controls effectively. Customize the script according to your project's requirements to generate the desired synchronized dataset.

License

This project is licensed under the MIT License - see the LICENSE file for details.