Content, progress and pedagogy of the module

Objective:

A robotic system needs an awareness of its context, i.e. objects and people in its vicinity. To this end, this course will teach the students how to use computer vision and pattern recognition methods to estimate the type of objects in the surroundings and the whereabouts of people nearby.

Learning objectives

Knowledge

• Must be able to explain the principles behind robust feature point algorithms
• Must have knowledge of feature selection and reduction methods
• Must have knowledge of motion analysis principles
• Must be able to understand tracking frameworks
• Must be able to understand how advanced perception is integrated into robotic systems (e.g visual servoing, obstacle avoidance)

Skills

• Must be able to apply sliding window approaches based on advanced features to detect objects
• Must be able to apply stereo vision to generate 3D data from two or more cameras
• Must be able to apply model-based approaches to estimate the 3D pose of objects and people
• Must be able to apply pattern recognition methods to classify object types and activities
• Must be able to integrate advanced perception into robotic systems

Competences

• Must be able to analyse a specific problem within mobile robotics and based upon this select, implement and evaluate an appropriate computer vision approach

Type of instruction

See the general description of the types of instruction described in § 17.

Exam

Exams