Vehicle identification method based on Gabor feature extraction and sparse representation

Abstract

The invention discloses a vehicle identification method based on Gabor feature extraction and sparse representation. The method comprises the steps: firstly, classifying vehicles, selecting a certain number of vehicles of each class as sample vehicle images, extracting the Gabor features of the sample vehicle images through a 2D-Gabor filter, constructing a feature dictionary needed for sparse representation through the Gabor features, and decreasing the number of dimensions of the constructed feature dictionary according to the principal component analysis method; then, extracting the Gabor features of vehicle images to be detected through the same 2D-Gabor filter, carrying out dimension reduction on the Gabor features, and working out the sparse coefficient of the Gabor features of the vehicle images to be detected on the constructed feature dictionary after dimension reduction according to the orthogonal matching pursuit algorithm; finally, judging the classes of the vehicles to be detected according to the residual error reconstruction method. The vehicle identification method has high accuracy and can meet the requirement for vehicle identification.

Description

technical field [0001] The invention relates to a vehicle recognition method based on Gabor feature extraction and sparse representation, belonging to the technical field of vehicle recognition. Background technique [0002] In recent years, with the rapid development of the national economy and the continuous improvement of people's living standards, motor vehicles have increased rapidly, traffic jams and traffic accidents have occurred frequently, and traffic problems have become increasingly severe. In order to improve the orderliness and reliability of the operation of the transportation system and realize the intelligent monitoring and management of transportation services, the Intelligent Transportation System (ITS) is particularly important, and it has become one of the cutting-edge technologies in the current transportation field. [0003] At present, the main vehicle identification method is to bury piezoelectric sensing materials in advance on the road section to b...

AI Technical Summary

Problems solved by technology

[0003] At present, the main vehicle identification method is to bury piezoelectric sensing materials in advance on the road section to be detected. When a vehicle passes by, the piezoelectric material generates electricity proportional to the load-bearing pressure. According to the size of the electricity and the number of times it is generated, the vehicle The load capacity and the number of axles, etc., can be determined by the method of template matching, but this method embeds the sensor to soften the road surface, and the road surface is more likely to be damaged. Secondly, the response result of the sensor is easily affected by the surrounding environment and heavy traffic, and the piezoelectric sensor The service life is limited, generally two years, and the replacement of the sensor needs to destroy the traffic road and re-pave it, which will cost a high price. These have brought great challenges to the traditional vehicle identification management method

Images