Speeded-up robust features based moving object detection on shaky video

Abstract

In this paper we propose a novel moving object detection method on shaky video employing speeded-up robust features (SURF). SURF features are extracted and tracked in each frame to estimate projective transformation parameters. We adopt RANdom SAmple Consensus (RANSAC) to improve the estimation accuracy. We use an efficient background subtraction method to detect moving objects in the scene. The background template is registered and updated in each frame to ensure the stability of the system. Experimental results performed on various video sequences demonstrate that our method can detect moving objects accurately and quickly. The proposed algorithm has the potential to achieve real-time performance. © Springer-Verlag 2011.