Transcript ppt

CSE 473/573
Computer Vision and Image
Processing (CVIP)
Ifeoma Nwogu
[email protected]
Lecture 4 – Image formation(part I)
Schedule
• Last class
– linear algebra overview
• Today
– Image formation and camera properties
• Readings for today: Forsyth and Ponce 1.1,
1.4, Szeliski 2.1 and 2.3.1 (optional).
Physical parameters of image
formation
• Optical
– Sensor’s lens type
– focal length, field of view, aperture
• Geometric
– Type of projection
– Camera pose
• Photometric
– Type, direction, intensity of light reaching sensor
– Surfaces’ reflectance properties
What is an image?
• Till now: a function – a 2D pattern of intensity
values
• Today: a 2D projection of 3D points
What is a camera?
• Some device that allows the projection of light
from 3D points to some “medium” that will
record the light pattern.
1st known photograph
View from the Window at le Gras,
Joseph Nicéphore Niépce 1826
Reproduction, 1952
Heliograph- a pewter plate coated with bitumen of Judea (an asphalt
derivative of petroleum); after at least a day-long exposure of eight hours, the
plate was removed and the latent image of the view from the window was
rendered visible by washing it with a mixture of oil of lavender and white
petroleum which dissolved away the parts of the bitumen which had not
been hardened by light. – Harry Ransom Center UT Austin
Image formation
• Let’s design a camera:
– Put a film in front of an
object
– Will we get a reasonable
image?
– Why? Why not?
Turning a room into a camera
obscura
Hotel room, contrast enhanced
View from hotel window
Accidental pinholes produce images that are unnoticed or
misinterpreted as shadows
A. Torralba and W. Freeman, Accidental Pinhole and Pinspeck Cameras, CVPR 2012
Image formation
• Let’s design a camera:
– Put a film in front of an object
– Add a barrier with an opening to block off most
of the rays (reduce blurring)
– Opening is called aperture
Ist known camera
• Known to Aristotle (384-322 B.C.)
• According to DaVinci “When images of illuminated objects ...
penetrate through a small hole into a very dark room ... you
will see [on the opposite wall] these objects in their proper
form and color, reduced in size, in a reversed position, owing
to the intersection of the rays".
• Depth of the room is the “focal length”
• How does the aperture size affect the image?
Shrinking the aperture
Pinhole too big many directions are
averaged, blurring the
image
Pinhole too smalldiffraction effects blur
the image
Generally, pinhole
cameras are dark, because
a very small set of rays
from a particular point
hits the screen.
Slide by Steve Seitz
Shrinking the aperture
Pinhole too big many directions are
averaged, blurring the
image
Pinhole too smalldiffraction effects blur
the image
Generally, pinhole
cameras are dark, because
a very small set of rays
from a particular point
hits the screen.
Adding a lens - concept of focus
• A lens focuses light onto the film
– There is a specific distance at which objects are “in focus”
– other points project to a “circle of confusion” in the image
• Changing the shape or relative locations of the lens
elements changes this distance
The thin lens
The thin lens
Sign is +ve when incident lens surface is convex, and –ve when concave
Depth of field
http://www.cambridgeincolour.com/tutorials/depth-of-field.htm
Depth of field is the range of distance within the subject that is acceptably sharp.
Slide by A. Efros
How can we control the depth of field?
• Changing the aperture size affects depth of field
– A smaller aperture increases the range in which the
object is approximately in focus
– But small aperture reduces amount of light – need to
increase exposure
Slide by A. Efros
Field of View (FOV)
• FOV is the extent of the observable world that
is seen at any given moment.
• For cameras, it is a solid angle through which a
detector is sensitive to light
– the area of the inspection
captured on the camera’s imager.
Zooming and Moving are not the
same…
Large FOV, small f
Camera close to car
Small FOV, large f
Camera far from the car
Real lens systems
Lens flaws: chromatic aberration
A lens can have different refractive indices for
different wavelengths: causes color fringing
Near Lens Center
Near Lens Outer Edge
Lens flaws: Spherical aberration
• Spherical lenses don’t focus light perfectly
• Rays farther from the optical axis focus closer
Lens flaws: Spherical aberration
Left: image showing low level of
spherical aberration and right:
image showing high level of
spherical aberration
http://www.mto-ophtalmo.ch/intraocular-lenses/neutralasphericity/
Radial distortion
– Caused by imperfect lenses
– Deviations are most noticeable near the edge of the lens
No distortion
Pin cushion
Barrel
Lens flaws: Vignetting
Digital camera
• A digital camera replaces film with a sensor array
– Each cell in the array is light-sensitive diode that converts photons to electrons
– Two common types
• Charge Coupled Device (CCD)
• Complementary metal oxide semiconductor (CMOS)
– http://electronics.howstuffworks.com/digital-camera.htm
Slide by Steve Seitz
CCD vs. CMOS
•
CCD: transports the charge across the chip and reads it at one corner of the array. An
analog-to-digital converter (ADC) then turns each pixel's value into a digital value by
measuring the amount of charge at each photosite and converting that measurement to
binary form
•
CMOS: uses several transistors at each pixel to amplify and move the charge using more
traditional wires. The CMOS signal is digital, so it needs no ADC.
http://electronics.howstuffworks.com/digital-camera.htm
http://www.dalsa.com/shared/content/pdfs/CCD_vs_CMOS_Litwiller_2005.pdf
Geometric projections
Types of 3D projections
• 3D projection is any method of mapping threedimensional points to a two-dimensional plane.
– Perspective projections
• objects in the distance appear smaller than those close by
• Parallel lines converge at an image point in infinity, on the
horizon
– Weak perspective projections
• perspective effects, not over the scale of individual objects
– Orthographic projections
• objects in the distance appear same size as those close by
• parallel lengths at all points are of the same scale
regardless of distance from the camera
Distant objects are smaller
Effects of perspective projection:
• Apparent size of object depends on their distance e.g. B’ and C’ have the same
height but in reality A and C are half the size of B
• Distance d from pinhole O to the plane of C is half the distance from O to plane
of A and B.
Parallel lines meet
It is common to draw the image plane (or film)
in front of the focal point. Moving the film
plane merely scales the image.
Projection of 2 parallel lines lying in the same plane:
• The projections of 2 parallel lines in the same plane F appear to converge on h
• h is a horizontal line formed by the intersection of image plane P and a plane
parallel to F passing through the aperture O.
• The line L in plane F and parallel to image plane P has no image
Vanishing points
• Each set of parallel lines (=direction) meets at a
different point
– The vanishing point for this direction
• Sets of parallel lines on the same plane lead to
collinear vanishing points.
– The line is called the horizon for that plane
• Good ways to spot faked images
– scale and perspective don’t work
– vanishing points behave badly
– supermarket tabloids are a great source.
Example of a scene vanishing point
Perspective projection
• Consider a coordinate system (O, i, j, k) attached to the camera whose origin O
coincides with the camera aperture.
• O is located at a distance d along the vector k.
• The line passing through the aperture and perpendicular to P is the optical axis
• The point c where this line intersects with the plane P is the image center. This is
often the origin of the image plane coordinate frame.
Perspective projection equations
• In image space, z = d
• Since P, O, and p are collinear, Op = lOP for
some l,
• x = lX, y = lY, d = lZ OR l =
𝑥
𝑋
• Therefore,
x= d
𝑋
𝑍
and y = d
𝑌
𝑍
=
𝑦
𝑌
=
𝑑
𝑍
Weak perspective
• An even coarser approximation of image formation
– Consider front-parallel plane Po defined by Z = Zo
– For any point P in Po
𝑑
– x = -mX, y = -mY, where m = 𝑍𝑜
– m is the positive magnification associated with plane Po
Weak perspective
• Issue
– perspective effects, but not over the scale of
individual objects
– collect points into a group at about the same
depth, then divide each point by the depth of its
group
– Advantage: easy
– Disadvantage: wrong
Orthographic projection
•
•
•
•
•
No reversal of image features
m = -1 (unnatural negative magnification)
All light rays are parallel to the k-axis and orthogonal to P
x = X, y = Y
Useful for creating to-scale drawings for construction and
engineering (showing details)
Modeling projection
y
d
z
x
Projection equation:
x y
( x, y , z )  ( d , d )
z
z
Source: J. Ponce, S. Seitz
Homogeneous coordinates
x
y
( x, y , z )  ( f , f )
z
z
• Is this a linear transformation?
• no—division by z is nonlinear
Trick: add one more coordinate:
homogeneous image
coordinates
homogeneous scene
coordinates
Converting from homogeneous coordinates
Slide by Steve Seitz
Perspective Projection Matrix
• Projection is a matrix multiplication using homogeneous
coordinates
f
0

 0
0
f
0
 x
0 0    f x 
y

0 0     f y 
z
1 0    z 
1 
x
y
(f , f )
z
z
divide by the third
coordinate
In practice: lots of coordinate transformations…
2D
point
(3x1)
=
Camera to
pixel coord.
trans. matrix
(3x3)
Perspective
projection matrix
(3x4)
World to
camera coord.
trans. matrix
(4x4)
3D
point
(4x1)
Orthographic projection (sort of…)
M.C. Escher's
waterfall
http://glasnost.itcarlow.ie/~powerk/GeneralGraphicsNotes/projection/orthographicprojection.html
Orthographic Projection
• Special case of perspective projection
– Distance from center of projection to image plane
is infinite
– Also called “parallel projection”
– What’s the projection matrix?
Slide by Steve Seitz
Physical parameters of image
formation
• Optical
– Sensor’s lens type
– focal length, field of view, aperture
• Geometric
– Type of projection
– Camera pose
• Photometric
– Type, direction, intensity of light reaching sensor
– Surfaces’ reflectance properties
Slide Credits
• David Forsyth – UIUC, slides accompanying Forsyth and
Ponce – Computer Vision book, 2/e
• Rob Fergus – NYU
• AaronBobick – GA Tech
• Svetlana Lazebnik - UIUC
Next class
• More on image formation (photometric)
• Readings for next lecture:
– Forsyth and Ponce 2.1, 2.2.4; Szeliski 2.2 (optional)
• Readings for today:
– Forsyth and Ponce 1.1, 1.4; Szeliski 2.1 and 2.3.1,
(optional)
Questions