Robust techniques for monocular simultaneous localization and mapping / von MSc-EE. Mohammad Hossein Mirabdollah ; Referentin: Prof. Dr.-Ing. Bärbel Mertsching, Korreferent: Prof. Dr.-Ing. Reinhold Häb-Umbach. Paderborn, 2016
Inhalt
Introduction
Visual Feature Extraction and Tracking
Visual Odometry
Data Fusion
Contributions of this Work
Thesis Outline
Related Literature
Visual Odometry
Feature Tracking
Feature Tracking Using Feature Matching Techniques
Feature Tracking Based on Lucas-Kanade Method
Comparison of Methods
Inconsistencies of N-point Methods for Camera Motion Estimation
Modified N-Point Methods
Considering Measurement Noise in the Estimation of E
Including Regularization Term
Modified Cheirality Test
Simulation
Experimental Results
2D Monocular SLAM
A Particle Filter Based Method
A Subspace Method with Minimal Number of Samples
Intersection of Two Line Segments Under the Uncertainties of Parameters
Posterior Distributions of f,t and 2 Given the Landmark Positions
Optimal Trimming of Line Segments
Resampling
Complexity Analysis
Simulation
Experimental Results
Conclusion and Outlook
Bibliography
List of Notations
List of Abbreviations
List of Tables
List of Figures