eprintid: 166
rev_number: 4
eprint_status: archive
userid: 6
dir: disk0/00/00/01/66
datestamp: 2008-10-07
lastmod: 2015-05-29 19:48:32
status_changed: 2009-04-08 16:55:11
type: report
metadata_visibility: show
item_issues_count: 0
creators_name: Cheng, Kell
contributors_name: Bona, Andrej
contributors_name: Bose, Chris
contributors_name: Engler, Wolgan
contributors_name: Gong, Minglun
contributors_name: Guyot, Cyril
contributors_name: Ivanescu, Christian
contributors_name: King, John
contributors_name: Kenway, Dan
contributors_name: Krislock, Nathan
contributors_name: Laflamme, Claude
contributors_name: Malill, Abid M.
contributors_name: Pillai, Suresh
contributors_name: Savu, Anamaria
contributors_name: Ting, Fridolin
contributors_name: Tomoda, Satoshi
title: The Tennis Ball Problem
ispublished: pub
subjects: other
studygroups: ipsw4
companyname: VisionSmart
full_text_status: public
abstract: Stereoscopic vision is a well-established phenomenon: biological evolution showed its utility in ancient times. In this workshop, we have examined some subtleties and limitations in applying this old concept to an entirely new application: with modern technology, we attempt to track the position of an early segment of a flying object, and then extrapolate its later trajectory.
problem_statement: Imagine that you wish to track the location of a tennis ball in space using just two cameras. Each camera produces a two-dimensional image at discrete time intervals, and these images are in fact composed of discrete pixels. The problems are:

1. How do we account for the inevitable distortions in the cameras?
2. How do we use our imperfect discrete data to best estimate the actual path?
3. Can we take into account the spin of the ball?
date: 2000
date_type: published
pages: 17
citation:   Cheng, Kell  (2000) The Tennis Ball Problem.  [Study Group Report]     
document_url: http://miis.maths.ox.ac.uk/miis/166/1/tennis_ball.pdf