一种融合区域化多特征的车辆检测方法  

A Vehicle Recognition Method Based on Regional Multi-Feature Fusion

作　　者：曹磊[1] 李璟[1] 洪留荣[1] 

机构地区：[1]淮北师范大学计算机科学与技术学院,安徽淮北235000

出　　处：《计算机工程与科学》2012年第6期123-126,共4页Computer Engineering & Science

基　　金：安徽省教育厅自然科学基金资助项目(KJ2012Z354)

摘　　要：为了提高车辆识别的效率,本文提出了一种基于车辆多种局部特征的融合识别算法。算法采用优化的PCA方法对车辆的多种局部特征进行抽取、分析及融合来实现车辆的识别检测。实验结果表明,该算法可有效降低车辆阴影及周围环境对识别效率的影响,提高车辆在部分遮挡情况下的检测效率,具有较好的精确性和稳定性。In order to improve the efficiency of vehicle recognition, this paper put recognition algorithm based on a variety of local features of the vehicles. The algorith mized Principal Component Analysis （PCA） to extract, analyze, and fuse the local fe cles so as to realize the recognition of vehicles. The results of the experiments indicate can effectively reduce the influence prove the recognition efficiency und rithm has pretty good accuracy and s forward a fusion m adopts the opti- atures of the vehi- that the algorithm of the shadows of vehicles and the surrounding environment, and im er the circumstance of partial hloeking of a vehicle, and that the algo stability.

关 键 词：车辆检测 特征区域 PCA 灰度均衡 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]