Lecture-11 - Problem ! Lecture-11 Structure from Motion...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
1 Copyright Mubarak Shah 2003 Lecture-11 Structure from Motion Copyright Mubarak Shah 2003 Problem ! Given optical flow or point correspondences, compute 3-D motion (translation and rotation) and shape (depth). Copyright Mubarak Shah 2003 Tomasi and Kanade Orthographic Projection Copyright Mubarak Shah 2003 Assumptions ! The camera model is orthographic. ! The positions of “p” points in “f” frames (f>=3), which are not all coplanar, have been tracked. ! The entire sequence has been acquired before starting (batch mode). ! Camera calibration not needed, if we accept 3D points up to a scale factor. Copyright Mubarak Shah 2003 Tomasi & Kanade ! ! ! " # $ $ $ % & ( V U W W ( u 11 " u 1 P # u F 1 " u FP v 11 " v 1 P # v F 1 " v FP & % $ $ $ $ $ $ $ # " ! ! ! ! ! ! ! * + P p F f v u fp fp , , 1 , , , 1 | ) , ( " " ( ( Image point Copyright Mubarak Shah 2003 Tomasi & Kanade fP fP fP fP fP fP b v v a u u ( ( ~ ~ , ( ( P p p f u P a 1 1 , ( ( P p p f v P b 1 1
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 Copyright Mubarak Shah 2003 ) ( ) ( f P T f fP f P T f fP t s j v t s i u ( ( ) , , ( P P p p Z Y X s ( 3D world point Orthographic projection k f ( i f - j f Copyright Mubarak Shah 2003 Copyright Mubarak Shah 2003 f fP fp a u u ( ~ ) ( 1 ) ( 1 f q P q T f f p T f t s i P t s i ( , ( ! " # $ % & ( , ( P q q P T f s P s i 1 1 P T f s i ( Origin of world is at the centroid of object points Copyright Mubarak Shah 2003 P T f fP P T f fP s j v s i u ( ( ~ ~ ! ! ! " # $ $ $ % & ( V U W ~ ~ ~ Copyright Mubarak Shah 2003 ./ RS s s j j i i W P T f T T f T ( ! ! ! ! ! ! ! ! " # $ $ $ $ $ $ $ $ % & ( " # # 1 1 1 ~ 2FX3 3XP P T f fP P T f fP s j v s i u ( ( ~ ~ ! ! ! " # $ $ $ % & ( V U W ~ ~ ~ Rank of S is 3, because points in 3D space are not Co-planar Copyright Mubarak Shah 2003 Rank Theorem Without noise, the registered measurement matrix is at most of rank three.
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 8

Lecture-11 - Problem ! Lecture-11 Structure from Motion...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online