Field of View (Zoom)

Similar documents
Single View Geometry. Camera model & Orientation + Position estimation. What am I?

CIS 580, Machine Perception, Spring 2015 Homework 1 Due: :59AM

Single View Geometry. Camera model & Orientation + Position estimation. What am I?

Camera Model and Calibration

Augmented Reality II - Camera Calibration - Gudrun Klinker May 11, 2004

Robot Vision: Camera calibration

Camera Model and Calibration. Lecture-12

Homogeneous Coordinates. Lecture18: Camera Models. Representation of Line and Point in 2D. Cross Product. Overall scaling is NOT important.

Image Formation. Antonino Furnari. Image Processing Lab Dipartimento di Matematica e Informatica Università degli Studi di Catania

Camera model and calibration

Rigid Body Motion and Image Formation. Jana Kosecka, CS 482

Camera model and multiple view geometry

Computer Vision: Lecture 3

Camera models and calibration

Computer Vision cmput 428/615

CIS 580, Machine Perception, Spring 2016 Homework 2 Due: :59AM

Introduction to Computer Vision. Introduction CMPSCI 591A/691A CMPSCI 570/670. Image Formation

CS201 Computer Vision Camera Geometry

Lagrange Multipliers. Lagrange Multipliers. Lagrange Multipliers. Lagrange Multipliers. Lagrange Multipliers. Lagrange Multipliers

Projective Geometry and Camera Models

CS535 Fall Department of Computer Science Purdue University

5LSH0 Advanced Topics Video & Analysis

Camera Projection Models We will introduce different camera projection models that relate the location of an image point to the coordinates of the

Module 4F12: Computer Vision and Robotics Solutions to Examples Paper 2

ECE-161C Cameras. Nuno Vasconcelos ECE Department, UCSD

Perspective projection and Transformations

Pin Hole Cameras & Warp Functions

Computer Vision Projective Geometry and Calibration. Pinhole cameras

CS 664 Slides #9 Multi-Camera Geometry. Prof. Dan Huttenlocher Fall 2003

Cameras and Radiometry. Last lecture in a nutshell. Conversion Euclidean -> Homogenous -> Euclidean. Affine Camera Model. Simplified Camera Models

DD2423 Image Analysis and Computer Vision IMAGE FORMATION. Computational Vision and Active Perception School of Computer Science and Communication

Full Field Displacement and Strain Measurement. On a Charpy Specimen. Using Digital Image Correlation.

Geometric camera models and calibration

Jump Stitch Metadata & Depth Maps Version 1.1

Robotics - Single view, Epipolar geometry, Image Features. Simone Ceriani

Computational Photography

Today. Stereo (two view) reconstruction. Multiview geometry. Today. Multiview geometry. Computational Photography

ROBOTICS 01PEEQW. Basilio Bona DAUIN Politecnico di Torino

A Stereo Machine Vision System for. displacements when it is subjected to elasticplastic

Stereo SLAM. Davide Migliore, PhD Department of Electronics and Information, Politecnico di Milano, Italy

Computer Vision Projective Geometry and Calibration. Pinhole cameras

Pin Hole Cameras & Warp Functions

CSE 252B: Computer Vision II

Module 6: Pinhole camera model Lecture 32: Coordinate system conversion, Changing the image/world coordinate system

The Lens. Refraction and The Lens. Figure 1a:

COSC579: Scene Geometry. Jeremy Bolton, PhD Assistant Teaching Professor

Assignment 2 : Projection and Homography

Computer Vision Projective Geometry and Calibration

Information page for written examinations at Linköping University TER2

EM225 Projective Geometry Part 2

Single-view 3D Reconstruction

CS4670: Computer Vision

Introduction to Computer Vision

Introduction to Homogeneous coordinates

Chapters 1-4: Summary

3D Sensing and Reconstruction Readings: Ch 12: , Ch 13: ,

3D Geometry and Camera Calibration

Camera Calibration. Schedule. Jesus J Caban. Note: You have until next Monday to let me know. ! Today:! Camera calibration

calibrated coordinates Linear transformation pixel coordinates

Theory of Stereo vision system

Outline. ETN-FPI Training School on Plenoptic Sensing

Agenda. Perspective projection. Rotations. Camera models

Digital Imaging Study Questions Chapter 8 /100 Total Points Homework Grade

Pinhole Camera Model 10/05/17. Computational Photography Derek Hoiem, University of Illinois

Understanding Variability

CV: 3D to 2D mathematics. Perspective transformation; camera calibration; stereo computation; and more

September 18, B Math Test Chapter 1 Name: x can be expressed as: {y y 0, y R}.

Robotics - Projective Geometry and Camera model. Marcello Restelli

CS6670: Computer Vision

An introduction to 3D image reconstruction and understanding concepts and ideas

Viewing. Reading: Angel Ch.5

Cameras and Stereo CSE 455. Linda Shapiro

Camera Models and Image Formation. Srikumar Ramalingam School of Computing University of Utah

NAME :... Signature :... Desk no. :... Question Answer

3-D D Euclidean Space - Vectors

CSE328 Fundamentals of Computer Graphics

CS 6320 Computer Vision Homework 2 (Due Date February 15 th )

3D Sensing. 3D Shape from X. Perspective Geometry. Camera Model. Camera Calibration. General Stereo Triangulation.

Review Exercise. 1. Determine vector and parametric equations of the plane that contains the

3. The domain of a function of 2 or 3 variables is a set of pts in the plane or space respectively.

Rendering. Generate an image from geometric primitives II. Rendering III. Modeling IV. Animation. (Michael Bostock, CS426, Fall99)

Specifying Complex Scenes

solidthinking User Interface

Lecture 7 Measurement Using a Single Camera. Lin ZHANG, PhD School of Software Engineering Tongji University Fall 2016

3D Rendering. Course Syllabus. Where Are We Now? Rendering. 3D Rendering Example. Overview. Rendering. I. Image processing II. Rendering III.

Realistic Camera Model

Announcements. Mosaics. How to do it? Image Mosaics

Computer Vision, Laboratory session 1

Vision Review: Image Formation. Course web page:

Computer Vision I - Algorithms and Applications: Multi-View 3D reconstruction

Introduction to 3D Machine Vision

For 3CCD/3CMOS/4CCD Line Scan Cameras. Designed to be suitable for PRISM based 3CCD/CMOS/4CCD line scan cameras

Epipolar geometry. x x

Computer Vision CS 776 Fall 2018

More Mosaic Madness. CS194: Image Manipulation & Computational Photography. Steve Seitz and Rick Szeliski. Jeffrey Martin (jeffrey-martin.

Scene Modeling for a Single View

CS 563 Advanced Topics in Computer Graphics Camera Models. by Kevin Kardian

Face Recognition At-a-Distance Based on Sparse-Stereo Reconstruction

M12VD1240IR M123VD4510IR M125VD3410IRCS M125VD922IRCS

Scene Reconstruction from Uncontrolled Motion using a Low Cost 3D Sensor

Transcription:

Image Projection

Field of View (Zoom)

Large Focal Length compresses depth 400 mm 200 mm 100 mm 50 mm 28 mm 17 mm 1995-2005 Michael Reichmann

FOV depends of Focal Length f f Smaller FOV = larger Focal Length

Field of View (Zoom)

Field of View (Zoom)

http://2blowup.com/fotografiapara-egobloggers-ii/

Fisheye lens distortion

Camera Model

Lens Pixel CCD sensor 3D object

3D Point Projection (Metric Space) 3D point ( X, Y, Z) ( u, v ) f m : Focal length in meter ( u, v ) ( X, YZ, ) X Y ( u, v ) ( fm, fm ) Z Z 2D projection onto CCD plane

3D Point Projection (Metric Space) Projection plane 3D point ( X, Y, Z) ( u, v ) ( u, v ) f m ( u, v ) Focal length in meter ( u, v ) ( X, YZ, ) X Y ( u, v ) ( fm, fm ) Z Z 2D projection onto CCD plane

3D Point Projection (Metric Space) Projection plane 3D point ( X, Y, Z) ( u, v ) ( u, v ) f m Focal length in meter ( X, YZ, ) X Y ( u, v ) ( fm, fm ) Z Z 2D projection onto CCD plane

3D Point Projection (Pixel Space) ( u, v ) (0,0) w h ( u, v ) w (0,0) h ( px, p ) y : Image principal point CCD sensor (mm) Image (pixel) u w u p x w v h v p y h w u u p w x h v v p h y

3D Point Projection (Pixel Space) O ( u, v ) w f m Projection plane h (0,0) ( u, v ) Z CCD sensor (mm) h (0,0) ( X, YZ, ) ( u, v ) w ( p, p ) x Image (pixel) y X Y ( XY,, Z) ( u, v ) ( fm, fm ) Z Z w w X u u px fm p w w Z Focal length in pixel h h Y v v py fm p h h Z Focal length in pixel y x

3D Point Projection (Pixel Space) O ( u, v ) w f m Projection plane h (0,0) ( u, v ) Z CCD sensor (mm) h (0,0) ( X, YZ, ) ( u, v ) w ( p, p ) x Image (pixel) y X Y ( XY,, Z) ( u, v ) ( fm, fm ) Z Z w w X u u px f f m x p w w Z Focal length in pixel h h Y v v py f f m y p h h Z Focal length in pixel y x

3D Point Projection (Pixel Space) Projection plane 3D point ( X, Y, Z) ( u, v ) ( u, v ) f m Focal length in meter ( X, YZ, ) w X h Y ( u, v ) ( fm, fm ) w Z h Z

Homogeneous Coordinate Projection plane ( X, YZ, ) 2D point =: 3D ray λ( xy,,1) O ( fx, fy, f ) f 2 ( xy, ) ( xy,,1) : A point in Euclidean space ( ) can be represented by fxy (,,1) ( xy,,1) 2 a homogeneous representation in Projective space ( P ) (3 numbers).

Homogeneous Coordinate 2D point =: 3D ray Projection plane ( X, YZ, ) λ( xy,,1) O ( fx, fy, f ) f λ( xy,,1) ( XYZ,, ) Homogeneous coordinate : 3D point lies in the 3D ray passing 2D image point.

3D Point Projection (Metric Space) 2D point =: 3D ray Projection plane ( X, YZ, ) λ( xy,,1) O f m ( fx, fy, f ) Z X Y ( xy,,1) ( fxfyf m, m, m) ( fm, fm, fm) Z Z

3D Point Projection (Pixel Space) O ( u, v ) w f m Projection plane h (0,0) ( u, v ) Z CCD sensor (mm) h (0,0) ( X, YZ, ) ( u, v ) w ( p, p ) x Image (pixel) y X Y ( XY,, Z) ( u, v ) ( fm, fm ) Z Z X Y u fx px v fy p Z Z u fx px X v fy py Y 1 1 Z Homogeneous representation y

Camera Intrinsic Parameter Pixel space Metric space O f m ( u, v ) Z ( X, YZ, ) u fx px X v fk y py Y 1 1 Z + Projection plane Camera intrinsic parameter : metric space to pixel space

2D Inverse Projection O f m ( u, v ) Z Projection plane ( X, YZ, ) 2D point == 3D ray K -1 u v 1 Note: arrow direction Pixel space Metric space u fx px X v fk y py Y 1 1 Z K -1 u X v Y 1 Z 3D ray The 3D point must lie in the 3D ray passing through the origin and 2D image point.

3D Point Projection (Pixel Space) Projection plane 3D point ( X, Y, Z) ( u, v ) ( u, v ) f m Focal length in meter ( X, YZ, ) w X h Y ( u, v ) ( fm, fm ) w Z h Z

Exercise What f to make the height of Eifel tower appear 960 pixel distance? 960 pix 21.8 mm 1280 pix 324 m size fm? Y h Y y f f m Z h Z 1500 m 1280 324 960 fm fm 0.0757m 0.0218 1500

Exercise What f to make the height of Eifel tower appear 960 pixel distance? 960 pix 21.8 mm 1280 pix 324 m size fm = 50 mm Y h Y y f f m Z h Z Z? 1280 324 960 0.05 Z 990.826m 0.0218 Z

Exercise What Zp to make the height of Eifel tower appear twice of the person? h e h p 324 m fm = 50 mm Z p 1500 1.7 m Y f h e Y Z p f Y Z f p 2 Y Z h p f Z p p s.t. h p h e 2 1.7 Z p 2 1500 157.41m 234

Where Was I? 324 m 670 pix fm = 50 mm 0.9 m Z 1 250 pix Circa 1984 Z 2 Y h Y h 1 Y1 1280 0.9 y1 f fm Z1 fm 0.05 8.03m Z h Z h y 0.0218 250 1 1 Y h h 2 Y2 Y2 1280 324 y2 f fm Z2 fm 0.05 1079m Z h Z h y 0.0218 670 2 2 2

Where Was I? Y h h 2 Y2 Y2 1280 324 y2 f fm Z2 fm 0.05 1079m Z h Z h y 0.0218 670 2 2 2 670 pix 250 pix Circa 1984 200m 400m 600m 800m 1000m

Where Was I? 800 m

f Focal Length

f Focal Length

f Focal Length

Focal Length

Dolly Zoom (Vertigo Effect) (Jaws 1975)

Dolly Zoom Given focal length (fm=100mm), what Z100 to make the height of the person remain the same as fm=50mm? h p H p f100 = 100 mm Z 100? Z 50 =157.41m

Dolly Zoom Given focal length (fm=100mm), what Z100 to make the height of the person remain the same as fm=50mm? h p H p f100 = 100 mm h f 50 50 Y Z 50 h f 100 100 Y Z 100 Z 100? s.t. h h 100 50 Z 50 =157.41m

Dolly Zoom Given focal length (fm=100mm), what Z100 to make the height of the person remain the same as fm=50mm? h p H p f100 = 100 mm h f 50 50 Y Z 50 h f 100 100 Y Z 100 Z 100? s.t. h h 100 50 Z 50 =157.41m Z f f Z 100 100 50 50 100 Z 100 157.41 314.8m 50

Dolly Zoom (Jaws 1975) Dolly Zoom (Vertigo Effect)