中文摘要：

研究基于Adaboost cascade算法的车牌检测技术,针对收费站车辆照片,选用局部矩形特征,利用cascade算法得到了98%的识别率,误识对象一般都集中在真实目标附近,从而验证了方法的可行性.

英文摘要：

This paper is mainly concerned with license plate detection that based on Adaboost cascade algorithm. Using part rectangle feature, we experimented on vehicle images from toll station and got the following results： 98 % recognition rate, the objects identified wrong generally muster around the target object. The results indicate that this method is feasible.