Moving object detection using registration for a moving camera platform

Abstract

In this research work, an accurate and fast moving object detector that can detect all the moving objects from Unmanned Aerial Platform (UAV) is proposed. Because of the distance of the UAV to the objects and the movement of the platform, object detection is a challenging task. In order to achieve best results with low error, at first the camera motion has to be estimated so, by using the Rosten and Drummond technique the corners is detected and then by using the corners the camera motion is compensated. After motion compensation, by subtracting the registered frame from the reference frame all the moving objects are detected and extracted.