Home // ICONS 2016, The Eleventh International Conference on Systems // View article

Safe Operation of Autonomous Machines

Authors:
Raj Aggarwal
Haoyuan Lin

Keywords: machine vision, autonomous operation, safe operation, stereo configuration, disparity

Abstract:
Reliable detection of foreign objects is a key requirement for safe operation of autonomous machines. In farming applications, an object is considered a foreign object if it can damage the machine or be damaged by it. The traditional machine vision approaches rely on detecting and classifying each type as a separate class of thereby increasing the computational load and compounding the machine vision problem since these machines have to operate in real-time and the foreign objects can appear in any orientation thereby increasing complexity. Moreover, it is an over kill since the safe operation of an autonomous machine does not require that we classify these objects as long as we can reliably detect them and direct the machine to take appropriate maneuvering action. In our application, the object of interest is bales and everything else is a foreign object. The foreign objects include humans, animals, vehicles and standing crop. We use disparity information from the two cameras in a stereo configuration and use the camera model to calculate the distance to the object. This object detection framework based on distance and size has proved to be more efficient and robust compared to traditional machine vision approaches.

Pages: 16 to 19

Copyright: Copyright (c) IARIA, 2016

Publication date: February 21, 2016

Published in: conference

ISSN: 2308-4243

ISBN: 978-1-61208-451-0

Location: Lisbon, Portugal

Dates: from February 21, 2016 to February 25, 2016