5.1.1. Elimination of Structure Unknowns
Since we assume rigidity we know that the motion can be represented by a rotation
and a translation
P = RP + T
where P and P are the position vectors of a world point before and afte
Since we got a nice taste of what regularization is, lets see a generic problem of the
type that is usually handled with regularization. The simplest such problem is smooth
interpolation in one dimension. This i
Linear Image Operations
1. Simple Smoothing Filter
Consider the following problem. We are given images that are corrupted by random
noise and we are asked to clean them. The content of the images can be described as pictures with rather smooth v
1. Image Sequences
It is said that a picture is worth a thousand words. Whoever said this was overly
optimistic about image compression technology, but assuming it is correct we can only
imagine the worth of a sequence of images.
Introduction to Statistical Estimation
1. The Simplest Kind of Statistics
Least Squares is the simplest and most intuitive kind of statistics and often the most
useful. The most straightforward application is as follows. We have a set of quantit
6. Hand Eye Coordination
Whenever we have a robot arm and a camera we need to the spatial relation betwen
the camera and the arm, so that the camera can give meaningful feedback to the arm.
Consider for examplesomeone leaving a s
1. Video Input Devices
A Computer Vision system senses its environment through a video camera and a
digitizer, and like its biological counterpart, the camera converts an optical image to a format that can be processed by a computi
5. Absolute Orientation
From the name alone one can guess that the algorithm is from the early twentieth
century, if not older and it must be a cousin of the relative orientation problem. Hard to
say how old it is because it has
Lab Assignment I.
Due Date: Oct. 5, 9:30am
1. Cheap Webcam Simulation
A. What is the matrix of the intrinsic parameters (calibration matrix) of a simple webcam with 640
480 resolution and 60 degrees horizontal viewing angle? Show the
Lab Assignment II.
Due Date: Oct. 19, 9:30am
1. Play with corners
Take two images from the same scene scene with the camera in two slightly different positions.
Make sure that nothing in the image moves more than a few pixels. Then det