Robust license-plate recognition method for passing vehicles underoutside environment

摘要

Novel methods to recognize license plates robustly are presented.nA sensing system with a wide dynamic range has been developed to acquirenfine images of vehicles under varied illumination conditions. Thendeveloped sensing system can expand the dynamic range of the image byncombining a pair of images taken under different exposure conditions. Innorder to avert blurring of images against fast passing vehicles, a prismnbeam splitter installed a multilayered filter, and two charge-coupledndevices are utilized to capture those images simultaneously.nFurthermore, to extend the flexibility of camera placement, anrecognition algorithm that can be applied to inclined plates has beenndeveloped. The performance of recognizing registration numbers onnlicense plates has been investigated on real images of about 1000nvehicles captured under various illumination conditions. Recognitionnrates of over 99% (conventional plates) and over 97% (highly inclinednplates) showed that the developed system is quite effective fornlicense-plate recognition