Color learning and illumination invariance on mobile robots: A survey

Mohan Sridharan,Peter Stone
DOI: https://doi.org/10.1016/j.robot.2009.01.004
IF: 3.7
2009-06-01
Robotics and Autonomous Systems
Abstract:Recent developments in sensor technology have made it feasible to use mobile robots in several fields, but robots still lack the ability to accurately sense the environment. A major challenge to the widespread deployment of mobile robots is the ability to function autonomously, learning useful models of environmental features, recognizing environmental changes, and adapting the learned models in response to such changes. This article focuses on such learning and adaptation in the context of color segmentation on mobile robots in the presence of illumination changes. The main contribution of this article is a survey of vision algorithms that are potentially applicable to color-based mobile robot vision. We therefore look at algorithms for color segmentation, color learning and illumination invariance on mobile robot platforms, including approaches that tackle just the underlying vision problems. Furthermore, we investigate how the inter-dependencies between these modules and high-level action planning can be exploited to achieve autonomous learning and adaptation. The goal is to determine the suitability of the state-of-the-art vision algorithms for mobile robot domains, and to identify the challenges that still need to be addressed to enable mobile robots to learn and adapt models for color, so as to operate autonomously in natural conditions.
robotics,automation & control systems,computer science, artificial intelligence
What problem does this paper attempt to address?