Fig 2 - available via license: Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International
Content may be subject to copyright.
Download
View publication
Illustration of Observation Constraints. Blue triangle represents reference cameras. Red points represents point clouds. Pink lines represent observation between cameras and point clouds. Green triangle represents coarse localization result. (a). Yellow triangle represents visible cone of point cloud p 3D j . (b). We firstly filter reference cameras co-visible with coarse camera pose using distance and angle between normal vectors of coarse camera pose and reference cameras. Camera C 3 is filtered out. (c). We filter out point clouds whose visible cone not covering coarse camera pose. Point cloud p 3D 0 is filtered out. (d). Finally we obtain point clouds that possibly observed by coarse camera pose.

Illustration of Observation Constraints. Blue triangle represents reference cameras. Red points represents point clouds. Pink lines represent observation between cameras and point clouds. Green triangle represents coarse localization result. (a). Yellow triangle represents visible cone of point cloud p 3D j . (b). We firstly filter reference cameras co-visible with coarse camera pose using distance and angle between normal vectors of coarse camera pose and reference cameras. Camera C 3 is filtered out. (c). We filter out point clouds whose visible cone not covering coarse camera pose. Point cloud p 3D 0 is filtered out. (d). Finally we obtain point clouds that possibly observed by coarse camera pose.

Source publication
Fig. 1. System framework overview. The system consists of four parts:...
Fig. 2. Illustration of Observation Constraints. Blue triangle...
Fig. 3. Illustration of pose graph. Pose graph includes VIO edges and...
Fig. 4. Illustration of drift RANSAC. Drift RANSAC randomly samples...
Fig. 5. Illustration of localization result. Yellow line is our result....
A Real-Time Fusion Framework for Long-term Visual Localization
Preprint
Full-text available
  • Oct 2022
Visual localization is a fundamental task that regresses the 6 Degree Of Freedom (6DoF) poses with image features in order to serve the high precision localization requests in many robotics applications. Degenerate conditions like motion blur, illumination changes and environment variations place great challenges in this task. Fusion with additiona...
Cite
Download full-text

Contexts in source publication

Context 1
... C i denotes position of camera i, and point cloud j is observed by N cameras. The illustration of observation constraints is shown in Fig. 2. n j is normalized. · · takes magnitude of a vector. ...
Context 2
... < θ j + ∆θ where p c is coarse camera position and p j is point cloud j position. L j is maximum observation distance of point cloud j. θ j is observation cone angle of point cloud j. ∆L and ∆θ enlarge searching range. An illustration of observation constraint is shown in Fig. ...