Feature based omnidirectional sparse visual path following

Vision sensors are attractive for autonomous robots because they are a rich source of environment information. The main challenge in using images for mobile robots is managing this wealth of information. A relatively recent approach is the use of fast wide baseline local features, which we developed and used in the novel approach to sparse visual path following described in this paper. These local features have the great advantage that they can be recognized even if the viewpoint differs significantly. This opens the door to a memory efficient description of a path by descriptors of sparse images. We propose a method for re-execution of these paths by a series of visual homing operations which yield a navigation method with unique properties: it is accurate, robust, fast, and without odometry error build-up.

Authors

• Toon Goedemé
• Tinne Tuytelaars
• Luc Van Gool
• Gerolf Vanacker
• Marnix Nuttin

Keywords

• Robot sensing systems
• Mobile robots
• Biosensors
• Robot vision systems
• Insects
• Robot kinematics
• Navigation
• Image sensors
• Sequences
• Mechanical engineering
• Visual Path
• Local Features
• Mobile Robot
• Odometry
• Series Of Operations
• Baseline Fasting
• Sparse Imaging
• Scaling Factor
• Visual Features
• Wheelchair
• Target Image
• Element Of Vector
• Corresponding Points
• Local Map
• Scale-invariant
• Test Platform
• Polar Coordinates
• Extended Kalman Filter
• Powerful Algorithms
• Feature Tracking
• SIFT Features
• Robot Localization
• Random Sample Consensus
• Measurement Equation
• Target Pose
• Computer vision
• robot navigation
• path following
• omnidirectional images