Please use this identifier to cite or link to this item:
|Title:||Algorithms and architectures for low-cost license plate recognition system||Authors:||Ku, Wei Chiet||Keywords:||DRNTU::Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision||Issue Date:||2012||Source:||Ku, W. C. (2012). Algorithms and architectures for low-cost license plate recognition system. Master’s thesis, Nanyang Technological University, Singapore.||Abstract:||License plate recognition system has its application in many areas such as recording parking statistics, identifying car thefts and monitoring traffic flow. Though several effective solutions have been proposed, the computation complexity has been one of the main limitations. Commercial mobile license plate recognition systems continue to suffer from the shortcomings of high cost and bulky implementation. This thesis presents several novel algorithms in an attempt to realize a low-cost license plate recognition system capable of real-time performance. In this research, the license plate recognition process is divided into three stages: license plate localization, license plate identification and license plate character recognition. The license plate localization was realized by first finding the vertical edges in the grayscale vehicle image using Sobel edge detector. A series of morphological operation was then applied on the vertical edges to extract rectangular regions with dense concentration of vertical edges. An efficient technique was then introduced and tested for the removal undesirable license plate regions. This technique employs Hough Transform based line detector to detect vertical lines and horizontal lines in the license plate candidates. A unique relationship between the number of horizontal and vertical lines were established to isolate the license plate more accurately. This made it possible for the elimination of other candidates with similar edge characteristics.||URI:||https://hdl.handle.net/10356/50543||DOI:||10.32657/10356/50543||Fulltext Permission:||open||Fulltext Availability:||With Fulltext|
|Appears in Collections:||SCSE Theses|
Page view(s) 20498
Updated on Aug 2, 2021
Updated on Aug 2, 2021
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.