Vehicle model identification method based on T-nearest neighbor spectral clustering under heterogeneous calculation model

The invention discloses a vehicle model identification method based on T-nearest neighbor spectral clustering under a heterogeneous calculation model, and relates to the field of vehicle information detection. The method comprises the following steps: step 1, extracting a characteristic value of a vehicle type frequency change waveform acquired by an induction coil; step 2, clustering different values of each attribute by adopting a T-nearest neighbor spectral clustering algorithm; and step 3, taking a clustering result of each attribute in the previous step as a different value of each attribute of the classifier, obtaining a vehicle model identification rule according to a Bayesian classifier algorithm, and identifying the vehicle model according to the rule in work. According to the invention, vehicle model identification of different vehicles can be effectively realized in real time. 本发明公开了一种异构计算模型下基于T近邻谱聚类的车型识别方法，涉及车辆信息检测领域。该方法包括如下步骤：步骤1，提取由感应线圈采集到的车型频率变化波形的特征值；步骤2，对每个属性的不同取值采用T近邻谱聚类算法进行聚类；步骤3，将上一步每个属性的聚