주성분 분석법을 이용한 지뢰인식 시스템

Abstract

GPR sensor를 이용한 지뢰탐지기에서 얻어지는 영상을 이용하여 보다 신속하고 정확하면서 종류와 수량까지 구분할 수 있는 시스템에 관해 연구하였다. 시스템 구현을 위해 기존의 GPR sensor 지뢰탐지기에 패턴인식 기술을 접목시켜 학습된 컴퓨터에 의해 전달된 영상을 분석하고 지뢰를 인식 및 분류하는 기술을 적용한다.|The NQR(Nuclear Quadrupole Resonance) and GPR(Ground Penetrating Radar) have been researched and developed to worldwide as the future technology of the detection landmines. The GPR among them have been interested in South Korea. The Mine detectors developed until now using GPR have the fault that takes long time to identify landmines. The reason is the fact that this GPR mothod can identify the landmines with only eye by persons in the monitor. This paper can overcome this problem using PCA(Pricipal Component Analysis). Features of landmines are extracted automatically using principal component analysis. Our system could classify the images automatically according to the proposed method in the experiment environment similar to real world. With principal component analysis method. This paper improved on the problem of longtime in landmines recognition. In the future, I will try the recognition with dynamic landmines images on real time.