SynesthesiaX

C++/OpenCV module for multimodal LiDAR-camera fusion, segmenting point clouds semantically based on camera labels using EfficientViT-SAM. ‍Synesthesiax projects LiDAR point clouds onto semantic camera images and publishes colored clouds + overlays.

📦 Important Dependencies

• ROS 2 Humble (rclcpp, sensor_msgs, cv_bridge)
• PCL (pcl_ros, pcl_conversions)
• OpenCV (>=4.0)
• message_filters

---

🛠️ Installation

# 1. Clone into your ROS 2 workspace
cd ~/ros2_ws/src
git clone https://github.com/Vaive-Logistics/synesthesiaX

# 2. Build the package
cd ~/ros2_ws
colcon build --packages-select synesthesiax

---

▶️ Usage

# Source your workspace
cd ~/ros2_ws
source /install/setup.bash

# Launch the node with default parameters
ros2 launch synesthesiax synesthesiax.launch.py

📁 Calibration file is loaded from install/share/synesthesiax/config/pinhole_model.yaml

---

👍 Enjoy! If you hit any issues, feel free to open an issue on the repo.