Fusing Structure from Motion and Simulation-Augmented Pose Regression from Optical Flow for Challenging Indoor Environments.
Felix OttLucas HeubleinDavid RügamerBernd BischlChristopher MutschlerPublished in: CoRR (2023)
Keyphrases
- structure from motion
- indoor environments
- feature tracking
- optical flow
- image sequences
- image based localization
- relative pose
- camera motion
- camera pose
- mobile robot
- bundle adjustment
- computer vision
- pose estimation
- camera pose estimation
- missing data
- path planning
- point cloud
- camera parameters
- scene structure
- motion estimates
- focal length
- motion estimation
- motion field
- robotic systems
- point features
- visual odometry
- motion parameters
- motion segmentation
- d objects
- line correspondences
- rigid motion
- key ingredient
- flow field
- motion analysis
- d scene
- motion model
- urban scenes
- ego motion
- model selection
- feature points
- moving objects
- three dimensional