车牌识别外文翻译文献综述.docx

立刻移除水印 立刻移除水印 车牌识别外文翻译文献综述 p0f element License Plate Recognition Based On Prior Knowledge Abstract - In this paper、a new algDrilhm based on improved BP (back pnopagation) neural network for Chinese vehicte license pEate recognition (LPR^ is described. Jhc proposed approach provides a solulicni fcr the vehicle license plale^ (VLP} which were tie graded severity. WhaE il nMnarkably differs from the tiadilional methods is Ehc appheadon of prior knowledge of license phle to the procedure of localioti, siigmentaiiot] and neeoignitioTi. Co hr cullacdlion is u^ed to Locate Lhe license plate? in Lhc image. Diinensions of each cliaratlcr are corihlant. which is uacd. to segment the charaeler of VLPs. The Layuui of the Chinese VLP is an important feature, which is used IO construct a dEassiGer fbr nxDgnizing. The Cxpcrimenldl resulb shtnv that The improved aigerithm is eflective underlhe condilioE that the license plaEts were degraded severely. Index Termi - License plate recognition, prior knowkiige. vehicle license plates, neural network. i. INTRODUCTION Vehicle Liccflse-Plate (VLP) reccgniliou is a xcry interesting but diOlcuh problcin. It is important in a number of applications such as wcighl-and-speed-limitT red [raffic infrirtgementT ruad surveys and park sceurily [1J. VLP recognition system cunsists of the phle lucattcn, the characters s(^Tncnla[ionT and the characters recognilicjn. These tasks become marc sephi^tieaced when deating with plale images cakcn in varicus inclined angles ot under various lighting, weather condilioti and cleanliness of lhe plaits Because this problem usttaHy used in real-time syslcmst ii nequirca nol only aeeuFsey but also fhsl processing. Most existing VLP reccgnilioH methadi [2]. [3], [4]. [5] reduce the complexity and increase lhe recognition rate by using some speci fic icalurcs of lotai VLPs and establishing some constrains on the posiLionT distance from the camera lo vehicles, and the inclined angks. In 3dditionh neural nelwork was u