Title
Unsupervised Learning of Terrain Appearance for Automated Coral Reef Exploration
Abstract
We describe a navigation and coverage system based on unsupervised learning driven by visual input. Our objectiveis to allow a robot to remain continuously moving above a terrain of interest using visual feedback to avoid leavingthis region. As a particular application domain, we are interested in doing this in open water, but the approach makes few domain-specific assumptions. Specifically, our system employed an unsupervised learning technique to train a k-Nearest Neighbor classifier to distinguish between images of different terrain types through image segmentation. A simple random exploration strategy was used with this classifier to allow the robot to collect data while remaining confined above a coral reef, without the need to maintain pose estimates. We tested the technique in simulation, and a live deployment was conducted in open water. During the latter, the robot successfully navigated autonomously above acoral reef during a 20 minutes period.
Year
DOI
Venue
2009
10.1109/CRV.2009.41
CRV
Keywords
Field
DocType
open water,unsupervised learning,visual input,automated coral reef exploration,k-nearest neighbor classifier,unsupervised learning technique,domain-specific assumption,coral reef,different terrain type,terrain appearance,acoral reef,visual feedback,coverage system,feedback,visual servoing,navigation,machine learning,robot,pose estimation,image segmentation,robots,path planning,testing,pixel,visualization,computer vision,k nearest neighbor,underwater acoustics,data mining,robotics
Motion planning,Computer vision,Computer science,Visualization,Terrain,Image segmentation,Unsupervised learning,Artificial intelligence,Visual servoing,Classifier (linguistics),Robot,Machine learning
Conference
Citations 
PageRank 
References 
6
0.52
12
Authors
5
Name
Order
Citations
PageRank
Philippe Giguere1454.51
Gregory Dudek22163255.48
Christopher Prahacs360.52
Nicolas Plamondon460.52
Katrine Turgeon5201.44