Three-Dimensional Computer Vision

HrvatskihrEnglishen

Pristupačnost

Text size:A A

Page layout Normal Simple

Page contrast Normal High contrast Invert high contrast

Clear all

Three-Dimensional Computer Vision

Course Description

The course studies basics of threedimensional computer vision. In particular, we consider pasive stereoscopic reconstruction, reconstruction withh structured light and specifics of reconstruction with LIDAR sensors.

Learning Outcomes

Explain advantages of reconstruction techniques based on pasive stereoscopy.
Explain advantages of reconstruction based on active illumination.
Explain advantages of reconstruction based on LIDAR sensors.

Forms of Teaching

Lectures

The course does not offer lectures in English.

Laboratory

One exercise in each half of the semester.

Grading Method

	Continuous Assessment		Exam
Type	Threshold	Percent of Grade	Threshold	Percent of Grade
Laboratory Exercises	40 %	20 %	50 %	0 %
Mid Term Exam: Written	0 %	40 %	0 %
Final Exam: Written	0 %	40 %
Exam: Written			50 %	80 %
Exam: Oral				20 %

Week by Week Schedule

Image formation, calibration, vanishing points
homography, essential matrix
Feature detection, differential tracking, wide-baseline matching.
Calibrated stereo: rectification, geometry, SGM, metric embeddings
active 3D reconstruction: time of flight, laser scanners, structured light
Structured light: projector/camera calibration, patterns for static scenes, patterns for dynamic scenes
Multiview 3D reconstruction with structured light 1: classic methods of coarse and fine 3D point cloud registration
Midterm exam
Multiview 3D reconstruction with structured light 2: deep models for 3D point cloud registration
Basics of 3D geometry and transformations, rotation and transformation matrices, rotation parameterizations, properties.
Point cloud preprocessing, normal estimation, base plane estimation, 3D features and descriptors, registration (point-to-point, point-to-plane), ICP, RANSAC
Effective point cloud representations, point reduction, projections, octree, voxels, kd-tree
Odometry for autonomous robots and vehicles, odometry from 3D point clouds
Exercises
Final exam

Study Programmes