動画予測を用いた異常診断手法の開発


英字タイトル:
Development of Anomalous Detection using Movie Prediction
著者:
榊原 洋志 Yoji SAKAKIBARA 出町 和之 Kazuyuki DEMACHI 河合 理城 Masaki KAWAI リッチャタクリ Ritu CHHATLULI 上赤 一馬 Kazuma KAMIAKA
発刊日:
公開日:
カテゴリ: 第9回
キーワードタグ:

概要

In this research, the new method to predict the near-future of the movie images captured by video camera based on the combination of t he Principle Component Analysis (PCA) and the Singular Spectral Analysis (SSA). In the normal condition of machines, the real-time ca ptured movie is supposed to correspond to the predicted one. If the error between the both becomes significantly large, it may suggest some anomalous motion of the machines. So the movie prediction method has a possibility of the sensitive anomalous detection system.


全文掲載のPDFファイルダウンロード