Camera Pose Estimation from Bounding Boxes

Visual localization is an important part of many interesting applications, including robotics. The dominant localization strategy is to estimate the camera pose from 2D-3D matches between 2D pixel positions and 3D points. Yet, such approaches can be quite memory intensive and can lead to privacy risks. An interesting alternative to point-based matches is to use higher-level primitives for pose estimation. Consequently, this work investigates using correspondences between 2D and 3D bounding boxes for camera pose estimation. The resulting scene representation is compact and poses fewer privacy risks. In this setting, there are typically orders of magnitude fewer matches available compared to classical feature-based methods. In addition, the available correspondences are significantly more noisy. We investigate multiple strategies based on converting bounding box correspondences to point correspondences and propose a novel and simple 2-point camera absolute pose solver (DP2P) that exploits the fact that the depths of the objects can be approximated from the sizes of their bounding boxes.

Authors

• Václav Vávra
• Torsten Sattler
• Zuzana Kukelova

Keywords

• Location awareness
• Privacy
• Visualization
• Solid modeling
• Three-dimensional displays
• Pose estimation
• Robot vision systems
• Cameras
• Noise measurement
• Intelligent robots
• Bounding Box
• Camera Pose
• Camera Pose Estimation
• Corresponding Points
• 3D Point
• Visual Localization
• Feature-based Methods
• Scene Representation
• Object Depth
• 3D Bounding Box
• Vertical Direction
• Measurement Noise
• Rotation Angle
• Position Error
• Inertial Measurement Unit
• Depth Estimation
• Objects In The Scene
• Roll Angle
• Box Dimensions
• Center Of The Bounding Box
• Direction Of Gravity
• Performance Of Solver
• Camera Angle
• Reprojection Error
• Human Pose Estimation
• Scene Point
• 3D Scene
• Query Image