快速Gabor滤波器在车型识别中的应用

Vehicle classification based on fast Gabor filters

车型识别在视频交通流量检测系统中是最重要、最关键的一环。为了提高车型识别速度和识别率,在分析了偶对称Gabor滤波器函数特征的基础上简化了求二维卷积的运算,用递归运算方法来实现。并提出一种新的适应分割采样策略,提取Gabor特征。从实验结果来看,此方法实现简单,能够有效降低Gabor滤波器的计算量,从而满足实时系统的需要;从有效性来看,该方法也能增强识别率和鲁棒性。

It is an important and key part to recognize moving vehicle in intelligence transportation system.In order to get a faster algorithm and good discrimination ability of vehicle classification,the properties of the real part of Gabor filter was analyzed,so the calculation of two-dimensional convolution calculation was simplified and implemented recursively.Then segmented sampling was used to get Gabor features for classification on the basis of the edge features in vehicles.Experiment results confirm that this proposed method is efficient and reliable,and its computation cost can also satisfies the real-time vehicle recognition system.