בוריס צ'רבאצקי, הרצאה סמינריונית למגיסטר
יום רביעי, 13.6.2012, 10:30
We address the problem of robot navigation using natural visual features in a
planar environment. The algorithm gets as an input a target, which is represented
by a captured image from the targets pose, and a source image, captured from the
starting position and should make an accurate move towards the target which can
be decomposed as a rotation and translation on the plane.
This algorithm can be used as a step in a higher level homing algorithm for
visual navigation. In addition when all the visual features are located on a
plane, orthogonal to the floor, there is a special structure for the Essential
Matrix and a Homography, and a novel algorithm was proposed for finding both of
them from only 2 point correspondence which gives a higher presicion and solves
numerical stability issues.
A theoretical connection between those entities was given, and a fast algorithm
for finding one from another was proposed. This technique also gives us the
ability to reconstruct the bounding walls of the environment