Development Of Image Filtering System For Car Sticker Detection

Elysa Sakinah , Md. Amin (2013) Development Of Image Filtering System For Car Sticker Detection. Project Report. UTeM, Melaka. (Submitted)

	PDF (24 Pages) CDR_09034-24_pages.pdf - Submitted Version Download (546kB)
	PDF (Full Text) CDR_09034.pdf - Submitted Version Restricted to Registered users only Download (5MB)

Abstract

The purpose of this report is to study a filtering algorithm that could be used to assist the current UTeM’s sticker verification system. The current verification system is not able to detect stickers in images taken under variable environmental behavior and of unfixed position on the windscreen of a car. The filtering algorithm for this project was done by using MATLAB 2013a. The methods that were used in this filtering technique include Color Correlation Technique, Accelerated Sum Squared Difference using Cross Correlation Technique and the Point Match Features using SURF Features Technique. These techniques have been tested using image samples having different range of brightness and different degrees of rotation. The result of the experiment showed one of the technique produced a good performance compared to others.

Actions (login required)

View Item