Principal Component Analysis Based on Shock Wave from Supersonic Projectiles and K-means Algorithm
Gengchen Shi · Tance yu kongzhi xuebao · 2004
A classification technique based on shock wave from supersonic projectiles is proposed in this paper. Signal time-domain features is extracted. By experimental analysis for 5. 56 mm, 7. 62 mm and 12. 7 mm projectiles dimensions about signal feature variables is reduced with principal component analysis (PCA). K-means class assignments are used to class projectiles' classification. In comparision with the effect of classification clustered by original features and clustered after PCA, it is confirmed that PCA is effective and identifying supersonic projectiles based on shock wave signal is feasible.