基本動作認識による屋外駐車場の不審行動検知システム

摘要

Studies for video surveillance applications for preventing various crimes such as stealing and violence have become a hot topic. This paper proposes a new video surveillance system that can detect suspicious behaviors such as a car break-in and vandalization in an open space parking, and that is based on image processing. The proposed system has the following features: it 1)deals time series data flow, 2)recognizes "human elemental actions" using statistic features, and 3)detects suspicious behavior using Subspace method and AdaBoost. We conducted the experiments to test the performance of the proposed system using open space parking scenes. As a result, we obtained about 10.0% for false positive rate, and about 4.6% for false negative rate.%様々な社会不安の要素がもたらす治安悪化への懸念が叫rnばれるようになった今日，治安対策に注目されている技術rnのーつとして監視カメラを挙げることができる。近年ではrn映像監視を自動化するためのシステム開発が盛んであrnり，画像認識の分野からも応用研究が進められている。