Development and evaluation of a ROS package to publish stereo vision sensing data as a LiDAR type message.

Resumo

The ability to obtain a representation of the environment is essential in applications that require
knowledge of the arrangement of objects that make up a particular work space. The collection of environmental
data through the use of Light Detection And Ranging (LiDAR), cameras or a combination of the two methods are
widely studied for several purposes. However, sensors of the LiDAR type still present significatively higher
financial costs when compared to alternatives that employ cameras. Currently, these types of sensing have been
widely used in autonomous mobile robots and vehicles due to the ability of these sensors to provide good detailing
of the scene within its angle of view allowing the objects surrounding the autonomous system to be detected, which
permit it to take necessary actions for its safe motion within the environment. The Robot Operating System (ROS)
is an open-source framework that helps researchers and developers to build and reuse code for robotic applications.
This work proposes the development of a ROS package that uses stereo cameras for the publication of messages
characteristic of a LiDAR system and, hence, be able to evaluate the replacement of a LiDAR by stereo cameras
with minimum changes to architecture of the system.