A Vehicle License Plate Recognition Algorithm Using Bilateral Filter and Wavelet Transformi

摘要：

This paper presents a new method to recognize the registered plate license of a vehicle using bilateral filter and wavelet transform. The method is based on an improved wavelet analysis method with point-point distance in eightdimension Euclidean space (EDES). There are three stages of the method. In the first stage, bilateral filter is used to remove image noises and enhance the plate information. In the second stage, some plate models are selected manually, and then, the wavelet transform is repeated to obtain eight channels of frequency information of the models. A recursive least square algorithm is used to estimate the approximate geometry plate model position (GPMP) in the EDES. In the last stage, each pixel of an input test image containing a plate is mapped into EDES using the same method. All pixels located within a threshold distance from the GPMP are considered the candidate plate pixels and the plate location is then found from the candidate plate pixels. The plate detection method is illustrated and verified by the experimental results obtained using real plate images and the method described in this paper.

关键词： Plate detection plate identification bilateral filter wavelet transform

作者: Yao Li Yan Hou Zhenhua Chai Hang Yang Minying Zhang Qi Li Samuel H. Chang

作者单位: School of Computer Software Changchun, Jilin Province, P.R.China School of Mathematics Jilin University, Changchun, Jilin Province, P.R.China

会议类型: 国际会议

会议名称: The 31st International Congress on Imaging Science(第31届国际影像科学大会 ICIS2010)

会议地点: 北京

会议语种:英文

页码: 118-121

在线出版日期: 2010-05-12（万方平台首次上网日期，不代表论文的发表时间）