论文部分内容阅读
提出融合局部二值模式(LBP)和Hu矩特征的车型识别算法。LBP特征能够很好地对车辆局部纹理进行描述,Hu矩属于全局特征,反映了车辆的形状轮廓信息,将这两种互补特征结合能更好地表达车型特征。设计了融合特征的提取方法,并结合支持向量机分类器构建了车型识别系统。实验结果表明,融合算法比单一的特征算法性能更优,提高了车型识别率。
A vehicle type recognition algorithm based on Local Binary Pattern (LBP) and Hu moments is proposed. The LBP feature can describe the vehicle’s local texture well. The Hu moment belongs to the global feature and reflects the shape and contour information of the vehicle. Combining these two complementary features can better express the vehicle features. The fusion feature extraction method is designed and a vehicle recognition system is constructed based on SVM classifier. The experimental results show that the fusion algorithm has better performance than the single feature algorithm and improves the vehicle recognition rate.