Self-adaptive vision system

Stipančić, Tomislav and Jerbić, Bojan (2010) Self-adaptive vision system. = Self-adaptive vision system. In: IFIP Advances in Information and Communication Technology, 22-24.02.2010., Costa de Caparica, Portugal.

[img] Text
Self-adaptive Vision System.pdf - Published Version
Restricted to IP adress Jezik dokumenta:English

Download (315kB)
Official URL: https://link.springer.com/chapter/10.1007%2F978-3-...

Abstract

Light conditions represent an important part of every vision application. This paper describes one active behavioral scheme of one particular active vision system. This behavioral scheme enables an active system to adapt to current environmental conditions by constantly validating the amount of the reflected light using luminance meter and dynamically changed significant vision parameters. The purpose of the experiment was to determine the connections between light conditions and inner vision parameters. As a part of the experiment, Response Surface Methodology (RSM) was used to predict values of vision parameters with respect to luminance input values. RSM was used to approximate an unknown function for which only few values were computed. The main output validation system parameter is called Match Score. Match Score indicates how well the found object matches the learned model. All obtained data are stored in the local database. By timely applying new parameters predicted by the RSM, the vision application works in a stabile and robust manner.

Item Type: Conference or Workshop Item (Lecture)
Keywords (Croatian): doctoral conference on computing, electrical and industrial systems
Subjects: NATURAL SCIENCES > Mathematics
TECHNICAL SCIENCE > Computing
TECHNICAL SCIENCE > Mechanical Engineering
Divisions: 900 Department of Robotics and Production System Automation > 920 Chair of Manufacturing and Assembly System Planning
Indexed in Web of Science: No
Indexed in Current Contents: No
Date Deposited: 13 Apr 2015 08:03
Last Modified: 24 Sep 2018 09:03
URI: http://repozitorij.fsb.hr/id/eprint/3698

Actions (login required)

View Item View Item

Downloads

Downloads per month over past year