Top Banner
CS 6825: Motion Part 2 CS 6825: Motion Part 2 – Optical Flow – Optical Flow
7

CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

Dec 21, 2015

Download

Documents

Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

CS 6825: Motion Part 2 – CS 6825: Motion Part 2 – Optical FlowOptical Flow

Page 2: CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

Recall: Optical FlowRecall: Optical Flow is an approximation of the 2D motion is an approximation of the 2D motion

field. field. Motion in the world usually occurs in 3D, Motion in the world usually occurs in 3D,

but, we have a 2D image sensor. So, we but, we have a 2D image sensor. So, we see the results as movement across the see the results as movement across the 2D image plane. 2D image plane.

Hence we are seeing the projection of the Hence we are seeing the projection of the 3D moving points onto the image plane. 3D moving points onto the image plane.

Page 3: CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

00

00

10

0

00

ˆby torelated

imagein induces , velocity has point Object

zr

rrrr

rv

rv

vv

f

dt

d

dt

d

P

ii

i

i

Motion Field and Optical Flow Motion Field and Optical Flow FieldField

Motion field: projection of 3D motion vectors on image planeMotion field: projection of 3D motion vectors on image plane

Optical flow field: apparent motion of brightness patternsOptical flow field: apparent motion of brightness patterns We equate motion field with optical flow fieldWe equate motion field with optical flow field

Page 4: CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

Brightness Constancy EquationBrightness Constancy Equation

Let P be a moving point in 3D:Let P be a moving point in 3D:• At time t, P has coords (X(t),Y(t),Z(t))At time t, P has coords (X(t),Y(t),Z(t))• Let p=(x(t),y(t)) be the coords. of its Let p=(x(t),y(t)) be the coords. of its

image at time t.image at time t.• Let E(x(t),y(t),t) be the brightness at p at Let E(x(t),y(t),t) be the brightness at p at

time t.time t. Brightness Constancy Assumption:Brightness Constancy Assumption:

• As P moves over time, E(x(t),y(t),t) As P moves over time, E(x(t),y(t),t) remains constant.remains constant.

Page 5: CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

Brightness Constancy Brightness Constancy EquationEquation

Taking derivative wrt time:Taking derivative wrt time:

Page 6: CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

Brightness Constancy EquationBrightness Constancy Equation

LetLet(Frame spatial gradient)(Frame spatial gradient)

(optical flow)(optical flow)

andand (derivative across frames)(derivative across frames)

Page 7: CS 6825: Motion Part 2 – Optical Flow Recall: Optical Flow is an approximation of the 2D motion field. is an approximation of the 2D motion field. Motion.

Brightness Constancy EquationBrightness Constancy Equation

Becomes:Becomes:

vvxx

vv

yy

rr E E

The Optical Flow The Optical Flow is CONSTRAINED is CONSTRAINED to be on a line !to be on a line !

-E-Ett/|/|rr E| E|Can calculate these: Different techniques to figure these out.

E is the spatial change in brightness in image i

Et is the difference in the brightness at (x,y) between image i and image i+1

We want to calculate v = [dx/dt dy/dt]