Abstract: Object detection is widely applied in automatic control fields. AdaBoost is a learning algorithm to construct accurate classers. In this paper , AdaBoost algorithm is improved and used to resolve real object detection problem. The algorithm given by this paper keeps the original accuracy in less detection time, thus it is more fit to be used in the real-time process of input system. In the end, the experimental results prove the efficiency of this method.

摘要： 目标检测在自动化控制领域具有广泛的应用。AdaBoost是一个构建准确分类器的学习算法。文章改进了AdaBoost算法，并将其运用于目标检测中。改进的算法不但保持了原来的准确性，而且速度也大大提高，更适合采集装置的实时处理要求。实验结果证明该方法有效。

关键词: AdaBoost算法, 目标检测, 误识率, 检测率

CLC Number: 

• TP274.4