After successful completion of the course, students are able to classify terms and concepts for camera-based navigation. Students of this course will be able to independently evaluate and read relevant literature in the research area of camera based navigation. In addition, the students gain a first experience in programming and algorithmic implementation of visual odometry algorithms.

We will discuss:

• Camera geometry
  • Target based calibration
  • Numerical optimization
• Stereo Cameras
  • Multi View geometry
  • Stereo rectification
  • Stereo Matching
• Optical flow
  • Optical flow model
  • Dense and sparse methods
• Visuelle odometry
  • Lie groups and rotation matrices
  • Stereo odometry
  • Monocular  odometrie
• SLAM
  • Kalman Filter
  • PTAM
  • DTAM

In order to deepen the understanding of the contents of the lecture, three worksheets will be distributed throughout the course. The worksheets contain 2-3 calculation tasks and 1-2 programming tasks.

IMPORTANT: The course "Visual Perception for Autonomous Navigation/Cars/Robots" will not take place in winter semester 2021/22. The next semester in which you can attend this course will be summer semester 2022 or winter semester 2022/23 at the latest.

Detailed course information will be presented in a preliminary discussion in the first lecture on the 1st Oct. 2020.

ECTS-Breakdown: 75h:

• 15h Lectures (5*3h)
• 45h Programming exercise (3*13h )
• 13h Test preparation
• 2h   Test

The performance evaluation of the lecture is based on the results of the worksheets and an oral examination at the end of the semester.

Requirements:

• basic programming skills
• basic knowledge of linear algebra and analysis
• basic knowledge in image processing