基于BP神经网络的车型识别研究  

Research on Vehicle Type Recognition Based on BP Neural Network

作　　者：吴志攀[1] 

机构地区：[1]惠州学院计算机科学系,惠州516007

出　　处：《现代计算机》2013年第3期38-41,共4页Modern Computer

摘　　要：车型识别是智能交通系统中的一个重要组成部分,近年来已成为国内外研究热点之一。提出一种基于特征提取的车型识别方法,该方法对车辆图像进行预处理,通过图像边缘检测、图像纵横填充、图像修正方法进行车型特征值提取,得到车型分类特征字空间,利用BP神经网络进行车型分类识别。实验结果表明,该方法高效可行,并对低质量和背景复杂图像有着良好的处理效果。The vehicle type recognition is a very important part of intelligent transport system and becomes one of the domestic and foreign research hotspots in recent years. Presents a way of the vehicle type recognition based on feature extraction. The way includes a pre-processing to the vehicle＇s in,age; and then a edge detecting, a horizontal and vertieal filling, a image correcting to get the vehicle＇s feature and the feature＇s word vector space; and finally a vehiele type recognizing based on BP-ANN. Through the MatLab simulation results, the algorithm based on image difference is good for the background image segmentation, and also can commendably reach the goal based on BP ANN, and have a good effect in low-quality or complex background images.

关 键 词：车型识别 特征提取 图像处理 边缘检测 BP神经网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]