In this chapter, we have given you an overview of the Computer Vision tools provided by ROS. We started by showing how we can connect and run several types of cameras, particularly FireWire and USB ones. The basic functionality to change their parameters was presented, so now you can adjust certain parameters to obtain images of good quality. Additionally, we provided a complete USB camera driver example.

Then, we showed how you can calibrate the camera. The importance of calibration is the ability to correct the distortion of wide-angle cameras, particularly cheap ones. The calibration matrix also allows you to perform many Computer Vision tasks, such as visual odometry and perception.

We showed how you can work with stereo vision in ROS and how you can set up an easy solution with two inexpensive webcams. We also explained the image pipeline, several APIs that work with Computer Vision in ROS, such as cv_bridge and image_transport, and the integration of OpenCV within ROS packages.

Finally, we enumerated useful tasks and topics in Computer Vision that are supported by tools developed in ROS. In particular, we illustrated the example of visual odometry using the viso2 and fovis libraries. We showed you an example of data recorded with a high-quality camera and also with the inexpensive stereo pair we proposed. Finally, feature detection, descriptor extraction, and matching was shown in order to illustrate how you can obtain homography between two images. All in all, after reading and running the code in this chapter, you will have seen the basics required to perform Computer Vision in ROS.

In the next chapter, you will learn how to work with point clouds using PCL, which allows you to work with RGBD cameras.