A vehicle classification method and system

Abstract

The invention discloses a vehicle classification method and system, wherein the vehicle classification method includes: determining a preset first training set as a target training set, determining projection matrices of all moving targets in the target training set, and establishing a feature space according to the projection matrix ;Compare the coordinates of all moving objects and preset moving objects in the feature space; according to the comparison results, classify all moving objects into the first moving object and the second type of moving object, and save the first type of moving object to the second Two training sets, the second moving object is saved to the third training set; the second training set and the third training set are respectively determined as target training sets, and the first type of moving object and the second type of moving object are further classified. According to the type of moving target, the training set is divided into six sub-training sets in detail, and the type of moving target to be classified is judged more specifically according to the moving target information in the six sub-training sets, and better classification accuracy is obtained.

Description

technical field [0001] The present invention relates to the technical field of pattern recognition, and more specifically, relates to a vehicle classification method and system. Background technique [0002] As an important part of the intelligent transportation system, the vehicle classification system is mainly combined with the detection, tracking, technology and classification of road vehicles. Traditional vehicle classification methods are mainly divided into two categories, one is based on radar and pressure sensors, and the other is image-based vehicle classification methods. Among them, the image-based vehicle classification method has been more and more applied with the rapid development of computer graphics. [0003] Existing image-based vehicle classification systems mainly use methods based on linear discriminant analysis (LDA Linear Discriminant Analysis) and principal component analysis (PCA Principal Component Analysis) to classify the acquired video sequence...

AI Technical Summary

Problems solved by technology

[0003] Existing image-based vehicle classification systems mainly use methods based on linear discriminant analysis (LDA Linear Discriminant Analysis) and principal component analysis (PCA Principal Component Analysis) to classify the acquired video sequences, although the existing image-based Although vehicle classification avoids the expensive cost of installing and maintaining sensors, the method of vehicle classification based on LDA and PCA only performs second-order statistical information on the training sample set in the process of classifying vehicles, ignoring high-order statistical information. , so the classification accuracy of the vehicle classification method based on LDA and PCA is low

Images