Title :
Identification of train wagon numbers
Author_Institution :
Depts. of Comput. Eng., Istanbul Commerce Univ., Istanbul, Turkey
Abstract :
In this work, we aimed to find and prepare to identify image parts, which include train wagon numbers. First, keypoints of wagon numbers are found. After that, the “candidate image”, about which we do not have any information, and the “reference image” are matched. The keypoints are found by the Scale Invariant Feauture Transform (SIFT) method of David G. Lowe. If enough keypoints are matched, than the area is controlled, where the information can be located. As a result of comparison, if there exists any matching areas, we check whether that area contains wagon number with 12 digits or not. After that, the part of selected area of image is prepared for identification utilizing image processing methods.
Keywords :
feature extraction; image matching; object recognition; SIFT method; candidate image; image matching; image parts identification; image processing methods; reference image; scale invariant feature transform; train wagon numbers identification; Arrays; Equations; Gray-scale; Mathematical model; Rail transportation; Standards; Transforms;
Conference_Titel :
Electrical and Electronic Engineering Conference (ElConRusNW), Proceedings of the 2014 IEEE NW Russia Young Researchers in
Conference_Location :
St. Petersburg
Print_ISBN :
978-1-4799-2593-3
DOI :
10.1109/ElConRusNW.2014.6839203