Moving objects detection from time-varied background: an application of camera 3D motion analysis

Abstract

Motion detection from a moving observer is a very important technique for 3D dynamical image analysis. Because of continuous background variations, detecting real moving objects has become very difficult and employs an optical flow method to measure the flow vector difference between the background and moving objects. Unfortunately, there is a huge calculation cost for obtaining accurate optical flow vectors. A new motion detecting method based on camera 3D motion analysis is proposed. With the special motion features of an on-board camera and FOE, our method can detect real moving objects simply by matching two adjacent frames. The camera 3D motion can theoretically be determined in our method with only three matching pairs, which makes it fast and more efficient for real-time applications.