Computer Vision. Exercise Session 1. Institute of Visual Computing

Similar documents
What are the camera parameters? Where are the light sources? What is the mapping from radiance to pixel color? Want to solve for 3D geometry

Calibrating a single camera. Odilon Redon, Cyclops, 1914

CS 534: Computer Vision Model Fitting

SLAM Summer School 2006 Practical 2: SLAM using Monocular Vision

Angle-Independent 3D Reconstruction. Ji Zhang Mireille Boutin Daniel Aliaga

Range images. Range image registration. Examples of sampling patterns. Range images and range surfaces

Lobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide

Structure from Motion

Barycentric Coordinates. From: Mean Value Coordinates for Closed Triangular Meshes by Ju et al.

Lecture 4: Principal components

Calibration of an Articulated Camera System with Scale Factor Estimation

12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification

Computer Animation and Visualisation. Lecture 4. Rigging / Skinning

Face Recognition University at Buffalo CSE666 Lecture Slides Resources:

New dynamic zoom calibration technique for a stereo-vision based multi-view 3D modeling system

Support Vector Machines

Solving two-person zero-sum game by Matlab

EVALUATION OF RELATIVE POSE ESTIMATION METHODS FOR MULTI-CAMERA SETUPS

Some Tutorial about the Project. Computer Graphics

Accounting for the Use of Different Length Scale Factors in x, y and z Directions

Gaussian elimination. System of Linear Equations. Gaussian elimination. System of Linear Equations

An Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices

LU Decomposition Method Jamie Trahan, Autar Kaw, Kevin Martin University of South Florida United States of America

AIMS Computer vision. AIMS Computer Vision. Outline. Outline.

A Robust Method for Estimating the Fundamental Matrix

Model-Based Bundle Adjustment to Face Modeling

LLVM passes and Intro to Loop Transformation Frameworks

Loop Transformations, Dependences, and Parallelization

An efficient method to build panoramic image mosaics

Reading. 14. Subdivision curves. Recommended:

Homography Estimation with L Norm Minimization Method

6.1 2D and 3D feature-based alignment 275. similarity. Euclidean

6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour

The Comparison of Calibration Method of Binocular Stereo Vision System Ke Zhang a *, Zhao Gao b

Computer Vision I. Xbox Kinnect: Rectification. The Fundamental matrix. Stereo III. CSE252A Lecture 16. Example: forward motion

Camera calibration. Robotic vision. Ville Kyrki

SENSITIVITY ANALYSIS IN LINEAR PROGRAMMING USING A CALCULATOR

Steps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices

Learning the Kernel Parameters in Kernel Minimum Distance Classifier

Multi-stable Perception. Necker Cube

An Optimal Algorithm for Prufer Codes *

Smoothing Spline ANOVA for variable screening

Parallel Numerics. 1 Preconditioning & Iterative Solvers (From 2016)

Biostatistics 615/815

CS 231A Computer Vision Midterm

3D Modeling Using Multi-View Images. Jinjin Li. A Thesis Presented in Partial Fulfillment of the Requirements for the Degree Master of Science

Finding Intrinsic and Extrinsic Viewing Parameters from a Single Realist Painting

3D Metric Reconstruction with Auto Calibration Method CS 283 Final Project Tarik Adnan Moon

OPTIMIZATION OF PROCESS PARAMETERS USING AHP AND TOPSIS WHEN TURNING AISI 1040 STEEL WITH COATED TOOLS

Programming in Fortran 90 : 2017/2018

Amnon Shashua Shai Avidan Michael Werman. The Hebrew University, objects.

Today Using Fourier-Motzkin elimination for code generation Using Fourier-Motzkin elimination for determining schedule constraints

ScienceDirect. The Influence of Subpixel Corner Detection to Determine the Camera Displacement

Ecient Computation of the Most Probable Motion from Fuzzy. Moshe Ben-Ezra Shmuel Peleg Michael Werman. The Hebrew University of Jerusalem

Comparison of traveltime inversions on a limestone structure

A Comparison and Evaluation of Three Different Pose Estimation Algorithms In Detecting Low Texture Manufactured Objects

Correspondence-free Synchronization and Reconstruction in a Non-rigid Scene

Non-Parametric Structure-Based Calibration of Radially Symmetric Cameras

Lecture #15 Lecture Notes

Calibration of an Articulated Camera System

A 3D Reconstruction System of Indoor Scenes with Rotating Platform

CS246: Mining Massive Datasets Jure Leskovec, Stanford University

IP Camera Configuration Software Instruction Manual

Image Alignment CSC 767

New Extensions of the 3-Simplex for Exterior Orientation

Proposed Simplex Method For Fuzzy Linear Programming With Fuzziness at the Right Hand Side

O n processors in CRCW PRAM

Collaboratively Regularized Nearest Points for Set Based Recognition

LECTURE NOTES Duality Theory, Sensitivity Analysis, and Parametric Programming

ROBOT KINEMATICS. ME Robotics ME Robotics

A Simple and Efficient Goal Programming Model for Computing of Fuzzy Linear Regression Parameters with Considering Outliers

Exercises (Part 4) Introduction to R UCLA/CCPR. John Fox, February 2005

R s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes

Line-based Camera Movement Estimation by Using Parallel Lines in Omnidirectional Video

NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS

ON SOME ENTERTAINING APPLICATIONS OF THE CONCEPT OF SET IN COMPUTER SCIENCE COURSE

User Authentication Based On Behavioral Mouse Dynamics Biometrics

REPRESENTING 2D and 3D data sets with implicit polynomials

Improving Initial Estimations for Structure from Motion Methods

On Some Entertaining Applications of the Concept of Set in Computer Science Course

Parameter estimation for incomplete bivariate longitudinal data in clinical trials

Outline. Self-Organizing Maps (SOM) US Hebbian Learning, Cntd. The learning rule is Hebbian like:

3D Rigid Facial Motion Estimation from Disparity Maps

K-means and Hierarchical Clustering

Programming Assignment Six. Semester Calendar. 1D Excel Worksheet Arrays. Review VBA Arrays from Excel. Programming Assignment Six May 2, 2017

2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements

Inverse-Polar Ray Projection for Recovering Projective Transformations

A Webpage Similarity Measure for Web Sessions Clustering Using Sequence Alignment

Scan Conversion & Shading

Two Dimensional Projective Point Matching

Proper Choice of Data Used for the Estimation of Datum Transformation Parameters

Robust Computation and Parametrization of Multiple View. Relations. Oxford University, OX1 3PJ. Gaussian).

A Scalable Projective Bundle Adjustment Algorithm using the L Norm

Data Modelling and. Multimedia. Databases M. Multimedia. Information Retrieval Part II. Outline

Scan Conversion & Shading

COMPLEX METHODOLOGY FOR STUDY OF INTERCITY RAIL TRANSPORT

Calibration of an Articulated Camera System

Design for Reliability: Case Studies in Manufacturing Process Synthesis

Subspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;

3D vector computer graphics

Transcription:

Computer Vson Exercse Sesson 1

Organzaton Teachng assstant Basten Jacquet CAB G81.2 basten.jacquet@nf.ethz.ch Federco Camposeco CNB D12.2 fede@nf.ethz.ch Lecture webpage http://www.cvg.ethz.ch/teachng/compvs/ndex.php

Organzaton Assgnments wll be part of the fnal grade Assgnment every week (or every two weeks the second part) 1 or 2 weeks tme to solve the assgnment Hand-n Thursday 13: (sharp!) Bonus ponts can be used to compensate for mssng ponts MATLAB Download (www.stud-des.ethz.ch)

Lterature Computer Vson: Algorthms and Applcaton by Rchard Szelsk, avalable onlne ( http://szelsk.org/book/ ) Multple Vew Geometry by Rchard Hartley and Andrew Zsserman Course Notes http://cvg.ethz.ch/teachng/compvs/tutoral.pdf

Camera Calbraton Intrnsc parameters K Radal dstorton coeffcents 2D ponts 3D ponts x x K R t

Camera Calbraton Use your own camera Buld your own calbraton object rnt checkerboard patterns Stch to two orthogonal planes Z Y

Camera Calbraton 4 Tasks: Data normalzaton Drect Lnear Transform (DLT) Gold Standard algorthm Bouguet s Calbraton Toolbox Use the same settngs for all tasks! Good reference: Multple Vew Geometry n computer vson (Rchard Hartley & Andrew Zsserman)

Data normalzaton Shft the centrod of the ponts to the orgn Scale the ponts so that average dstance to the orgn s and, respectvely. Determne usng normalzed ponts. Determne 1 3 3 3 1 2 2 1 1 z D y D x D y D x D c s c s c s c s c s U T 3 U T ˆ 1 2 ˆ

Drect Lnear Transform (DLT) 4 3, 3,3 2 1, 1,1 3 2 1 1 1 y y y y x x x x y w x w z y x z y x z y x z y x T T T T T T A

2n 2n 2n 12 Drect Lnear Transform (DLT) Sngular Value Decomposton 12 2n 12 12 A = U S V

2n 2n 2n 12 Drect Lnear Transform (DLT) Sngular Value Decomposton 12 2n 12 12 A = U S V = 1,1 1, 2 3,3 3, 4 1,1 2,1 3,1 1, 2 2, 2 3, 2 1,3 2,3 3,3 1, 4 2, 4 3, 4

Camera Matrx Decomposton (K and R) R t KR KRC K K s upper trangular R s orthonormal QR decomposton A = QR Q s orthogonal R s upper trangular

Camera Matrx Decomposton (K and R) K R t KR KRC M KR M 1 R 1 K 1 Run QR decomposton on the nverse of the left 3x3 part of Invert both result matrces to get K and R

Camera Matrx Decomposton (C) The camera center s the pont for whch C Ths s the rght null vector of ( SVD)

Gold Standard Algorthm Normalze data Run DLT to get ntal values Compute optmal squared reprojecton errors ˆ by mnmzng the sum of ˆ N mn d( 1 xˆ, ˆ ˆ ) 2 Denormalze ˆ

Mnmzaton n MATLAB Fmnsearch(...) See code framework Lsqnonln(...) nonlnear least-squares Vectorze your parameters

Bouguet s Calbraton Toolbox Download and nstall the toolbox: (http://www.vson.caltech.edu/bouguetj/calb_doc/ndex.html) Go through the tutoral and learn how to calbrate a camera wth that toolbox rnt your own calbraton pattern (avalable on the webste) Use the toolbox for calbraton and compare the result wth the results of your own calbraton algorthm

Hand-n Source code Matlab.mat fle wth hand-clcked 3D-2D correspondences Image used for calbraton Vsualze hand-clcked ponts and reprojected 3D ponts Dscuss and compare values of calbraton obtaned for all methods Dscuss average reprojecton error of all methods.

Some hnts Work wth normalzed homogeneous coordnates always. Camera calbraton K should respect ths conventon. Check that the obtaned orentaton R correspond to the expected world value, otherwse K wll have negatve values n ts dagonal. When reprojectng ponts use average of reprojecton error for comparson and remember to normalze reprojected coordnates to w = 1. Remember to use the same camera wth the same settngs for all tasks!

Hand-n Example reprojecton of the the 3D ponts