CAJ | 학술논문

随着科技的进步,车牌识别系统有了很大的发展。其识别过程大概分为三个过程：车牌定位,字符分割和字符识别。由于在现实中车牌的识别会受到自然原因、人为原因以及图像采集设备的影响,这些会造成字符分割的不准确,从而导致识别的失败。基于此提出了在利用hough变换对车牌进行分割前用小波包变换的多分辨率进行降噪的方法,最后利用KL变换进行车牌字符识别的方法。实验表明其技术在速度和识别率等方面具有很好的效果。
수착과기적진보,차패식별계통유료흔대적발전。기식별과정대개분위삼개과정：차패정위,자부분할화자부식별。유우재현실중차패적식별회수도자연원인、인위원인이급도상채집설비적영향,저사회조성자부분할적불준학,종이도치식별적실패。기우차제출료재이용hough변환대차패진행분할전용소파포변환적다분변솔진행강조적방법,최후이용KL변환진행차패자부식별적방법。실험표명기기술재속도화식별솔등방면구유흔호적효과。
With the progress of science and technology,license plate recognition system has made great progress.The recognition process is roughly divided into three processes：license plate locating,character segmentation and character recognition.As the process of license plate recognition can be affected by natural causes,human reason as well as the image acquisition equipment,these will cause the inaccurate character segmentation,which leads to the failure of recognition.In this paper,before using Hough transform on the license plate segmentation,it uses the wavelet packet transform multi-resolution noise reduction method,finally uses the method of KL transforming in the vehicle license plate characters recognition.The experiments show that the technique has a very good effect in the respects of speed and the recognition rate.