Abstract
The detection and description of features is one basic technique for many visual robot navigation systems in both indoor and outdoor environments. Matched features from two or more images are used to solve navigation problems, e.g., by establishing spatial relationships between different poses in which the robot captured the images. Feature detection and description is particularly challenging in outdoor environments, and widely used grayscale methods lead to high numbers of outliers. In this paper, we analyze the use of color information for keypoint detection and description. We consider grayscale and color-based detectors and descriptors, as well as combinations of them, and evaluate their matching performance. We demonstrate that the use of color information for feature detection and description markedly increases the matching performance.
Highlights
Autonomous lawn-mowing under visual guidance is a complex task for outdoor robots
We focus on the impact of incorporating color information in feature detection and description measured by an evaluation criterion suitable for a subsequent random sample consensus (RANSAC) [8] step
The right plot shows the most relevant 1000 keypoints that were detected by color interest points (CIP) with the detection done based on color information
Summary
Autonomous lawn-mowing under visual guidance is a complex task for outdoor robots. One of the basic elements is a visual navigation system that enables a systematic covering of the entire working area. In outdoor environments, varying illumination conditions as well as seasonal changes and nonplanar terrain pose multiple challenges to a visual navigation system. For the class of methods that we are interested in, the computation of the spatial relationship between arbitrary views (“home vectors”) is required. We focus on feature matching without restricting the search space by feature tracking and study the effect of incorporation of color information in order to improve feature matching. We concentrate our study on the performance in the context of domestic gardens and lawns near public buildings
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
