https://www.youtube.com/watch?v=8nZy1cZFoZM

Other Lecture Slides and Videos / Coding / Pen & Paper

https://uni-tuebingen.de/fakultaeten/mathematisch-naturwissenschaftliche-fakultaet/fachbereiche/informatik/lehrstuehle/autonomous-vision/lectures/self-driving-cars/

이전까지의 강의는 수강하지 않아 초반 부분에 대한 내용은 잘 모름.

Indirect(feature-based) vs. Direct SLAM

EKF SLAM, BUndle adjustment, Windowed BA, …

이 강의에서는 bundle adjustment를 이용한 feature-based SLAM을 다루고자 함.

Untitled

http://wavelab.uwaterloo.ca/
2017-SLAM/Lecture14-Direct_visual_inertial_odometry_and_SLAM

http://wavelab.uwaterloo.ca/ 2017-SLAM/Lecture14-Direct_visual_inertial_odometry_and_SLAM

Feature-based SLAM

Goal: Optimized reprojection errors wrt. camera parameters and 3D Point cloud

Bundle Adjustment

minimizes the reprojection error of all observations: