Title :
Simple solution for public transport route number recognition based on visual information
Author :
Mikolaj Leszczuk;Lukasz Skoczylas;Andrzej Dziech
Author_Institution :
AGH Univ., Krakó
Abstract :
Freedom of movement is a major challenge for people with visual impairments. Movement in urban spaces is a particular problem. Hence, the aim of this study was to develop a method for locating and recognizing a public transport route number on the basis of visual information collected using smart-phone cameras. The proposed image processing algorithm, based on photos of oncoming buses or trams, generates text information of the route number of the oncoming vehicle using Computer Vision (CV) and Optical Character Recognition (OCR) techniques. The accuracy of the public transport route number identification based on a single image is approximately 60%. However, by collecting the results of a series consecutive shots, a recognition performance of over 85% is obtained.
Keywords :
"Optical character recognition software","Optical filters","Image recognition","Optical imaging","Filtering algorithms","Support vector machines","Image resolution"
Conference_Titel :
Signal Processing: Algorithms, Architectures, Arrangements, and Applications (SPA), 2013
Electronic_ISBN :
2326-0319