کنفرانس مهندسی برق ایران

صفحه اصلی / سی و دومین کنفرانس بین المللی مهندسی برق

Classification of automotive radar targets using Gray Level Co-occurrence Matrix

نویسندگان :

Amin AghatabarRoodbary¹ MohammadHassan Bastani² Fereidoon Behnia³

1- دانشگاه صنعتی شریف 2- دانشگاه صنعتی شریف 3- دانشگاه صنعتی شریف

کلمات کلیدی :

Radar target classification،automotive radars،feature extraction،micro-Doppler signature،gray level co-occurrence matrix

چکیده :

Classification of targets has always been one of the most important topics in the field of radar. This issue is highlighted in the field of automotive radars, as the targets of interest are generally related to human life such as pedestrians, cars, cyclists, motorcyclists, traffic signs, etc. The reason for the need to classify targets is the difference in the decision-making process regarding each category of targets. The main challenge is to extract unique features from radar return signals to increase classification accuracy. Signatures such as Radar Cross Section(RCS), micro-Doppler, range-Doppler, etc. are used to extract features. One of the topics of interest is feature extraction from the micro-Doppler signature of targets, as these images have unique features that can be useful in classifying these targets. In this paper, the Gray Level Co-occurrence Matrix (GLCM), a powerful method for image texture analysis, is used for extracting the spatial features of gray images for multi-class classification problems. Moreover, the KNN(K-Nearest Neighbor) and ensemble learner with 200 weak learners are used to evaluate classification accuracy. Simulation results show that the proposed features extracted from GLCM of the micro-Doppler signatures of these six classes of targets have good classification performance and can discriminate the targets with about 90 percent of accuracy for Signal To Noise Ratio(SNR) 30dB.