4. MorBot Package - arieldo/MorBot GitHub Wiki
Welcome to the MorBot packages Wiki!
MorBot Packages are a collection of customized ROS (Robot Operating System) packages designed to enhance robotics applications with advanced computer vision and deep learning capabilities. Each package within the MorBot suite leverages NVIDIA's powerful Jetson platform, specifically tailored to process and interpret visual data from robotics sensors. Below is a detailed overview of each package included in the MorBot offerings:
Jetbot-Camera-ROS is a ROS node that interfaces with the camera on NVIDIA's JetBot, a small-scale robot built on the Jetson Nano. This package captures live images from the JetBot's onboard camera and publishes them to a designated ROS topic, facilitating real-time image processing and interaction in ROS environments.
- Real-time image capture from JetBot camera.
- Seamless integration with ROS.
- Publication of image data to ROS topics for further processing.
- Follow the instructions in the
Jetbot-Camera-ROSto install Jetbot-Camera-ROS.
Posenet-ROS integrates the capabilities of the PoseNet model, a body pose estimation algorithm from NVIDIA's jetson_inference library, into ROS. This package allows for real-time detection and tracking of human body joints, making it ideal for applications involving human-robot interaction and activity recognition.
- Real-time body pose estimation.
- Tracking of multiple human joints.
- Utilization of NVIDIA's efficient PoseNet model.
- Follow the instructions in the
Posenet-ROSto install Posenet-ROS.
Segnet-ROS employs the SegNet deep learning model for semantic segmentation within ROS environments. Semantic segmentation involves labeling parts of images with corresponding class categories (such as roads, people, vehicles) at the pixel level. This package is particularly useful in autonomous navigation and environment understanding.
- Pixel-level image segmentation.
- Supports multiple class categories for detailed environment mapping.
- Integrates SegNet, a robust deep learning model for semantic segmentation.
- Follow the instructions in the
Segnet-ROSto install Segnet-ROS.
Detectnet-ROS utilizes object detection algorithms, specifically DetectNet and SSD MobileNet V2, within ROS. This package provides capabilities to detect and identify various objects in real-time, enhancing situational awareness for robots in diverse environments.
- Real-time object detection and identification.
- Supports multiple detection models including DetectNet and SSD MobileNet V2.
- Integration with ROS for extensive robotic applications.
- Follow the instructions in the
Detectnet-ROSto install Detectnet-ROS.
The MorBot packages collectively provide a comprehensive suite for enhancing robotic systems with advanced visual processing capabilities, making them invaluable tools in fields such as autonomous navigation, interactive robotics, and surveillance. Each package is meticulously designed to integrate seamlessly with ROS, ensuring that users can easily implement state-of-the-art computer vision technologies in their robotic projects.
The ROS Workshop is an hands-on training session focused on JetBot Motion Libraries.
Follow the instructions in Ros-Workshop to Build your first Moving MorBot Demo.