基于稀疏取样和梯度分布特征的车标识别  

Vehicle Logo Recognition Based on Sparse Sampling and Gradient Distribution Features

作　　者：周斌斌 高尚兵 潘志庚 王亮亮[1] 王洪阳 

机构地区：[1]淮阴工学院计算机与软件工程学院,淮安223001 [2]杭州师范大学数字媒体与人机交互研究中心,杭州311121

出　　处：《系统仿真学报》2017年第9期2035-2042,共8页Journal of System Simulation

基　　金：国家自然科学基金(61402192;61332017);国家重点研发计划(2015BAK04B05);江苏省六大人才高峰资助项目(XYDXXJS-011);江苏省333工程资助项目(BRA2016454);江苏省青蓝工程资助

摘　　要：传统方法将车标定位与识别分开进行,定位的误差将会给后续的识别带来影响,并且车标图像具有低分辨率、低质量的特点。提出了一种新颖的车标定位和识别有机整合的方法。通过稀疏取样对样本图像进行取点采样,将点集分为邻近点集和非邻近点集,分别对其提取梯度特征和明暗特征,构建特征库,对车标粗定位区域进行多尺度扫描。实验结果表明,该方法在车标检测识别效率方面具有更大的优势,而且对于不同类型的车标图像都具有鲁棒性。The vehicle logo location and recognition are separated in the traditional method, the location errors will affect the subsequent recognition, at the same time the vehicle logo images are with low resolution and poor quality. Thus, a novel method was proposed which integrated the vehicle logo location and recognition organically. The sample images were sampled by sparse sampling, and then the point set was divided into adjacent point set and non adjacent point set, and the gradient feature and light and dark feature were extracted respectively, constructing the feature library. The logo coarse location area was multi-scale scanned. The experimental results show that the proposed method is superior to other advanced algorithms on the vehicle detection and recognition efficiency, and robust to the different types of logo images.

关 键 词：图像处理 灰度分布 梯度分布 多尺度检测 车标定位 车标识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]