Book Image

ROS Programming: Building Powerful Robots

By : Anil Mahtani, Aaron Martinez, Enrique Fernandez Perdomo, Luis Sánchez, Lentin Joseph
Book Image

ROS Programming: Building Powerful Robots

By: Anil Mahtani, Aaron Martinez, Enrique Fernandez Perdomo, Luis Sánchez, Lentin Joseph

Overview of this book

This learning path is designed to help you program and build your robots using open source ROS libraries and tools. We start with the installation and basic concepts, then continue with the more complex modules available in ROS, such as sensor and actuator integration (drivers), navigation and mapping (so you can create an autonomous mobile robot), manipulation, computer vision, perception in 3D with PCL, and more. We then discuss advanced concepts in robotics and how to program using ROS. You'll get a deep overview of the ROS framework, which will give you a clear idea of how ROS really works. During the course of the book, you will learn how to build models of complex robots, and simulate and interface the robot using the ROS MoveIt motion planning library and ROS navigation stacks. We'll go through great projects such as building a self-driving car, an autonomous mobile robot, and image recognition using deep learning and ROS. You can find beginner, intermediate, and expert ROS robotics applications inside! It includes content from the following Packt products: ? Effective Robotics Programming with ROS - Third Edition ? Mastering ROS for Robotics Programming ? ROS Robotics Projects
Table of Contents (37 chapters)
Title page
Copyright and Credits
Packt Upsell
Preface
Bibliography
Index

ROS images


ROS provides the sensor_msgs::Image message to send images between nodes. However, we usually need a data type or object to manipulate the images in order to do some useful work. The most common library for that is OpenCV, so ROS offers a bridge class to transform ROS images back and forth from OpenCV.

If we have an OpenCV image, that is, cv::Mat image, we need the cv_bridge library to convert it into a ROS image message and publish it. We have the option to share or copy the image with CvShare or CvCopy, respectively. However, if possible, it is easier to use the OpenCV image field inside the CvImage class provided by cv_bridge. That is exactly what we do in the camera driver as a pointer:

cv_bridge::CvImagePtr frame; 

Being a pointer, we initialize it in the following way:

frame = boost::make_shared<cv_bridge::CvImage>(); 

If we know the image encoding beforehand, we can use the following code:

frame->encoding = sensor_msgs::image_encodings::BGR8; 

Later, we set the OpenCV...