Moving object detection and tracking based on fractional derivative and otsu thresholding

Abstract

Object detection and tracking in video surveillance is a challenging job due to the requirement of higher accuracy and efficiency in the presence of large volume of data The object detection attributes namely shape size colour velocity and the direction of the moving object can be incorporated in the object detection and tracking system The dynamic environmental conditions frequent changes in the motion of an object background intensity variations changes in colour and light illuminations appearance disparity occlusions etc are the main issues of object detection and tracking models A method is required to establish a threshold in a dynamic way with the consideration of pixel intensities of each frame Commonly used methods have been developed for deriving objects from the captured images and recorded videos works in an inhibited environment The major contribution presented in this thesis deals with markerless motion capturing spatiotemporal with color determination and fractional derivative based Otsu thresholding in typical object detection and tracking system The first part of the thesis deals with the Markerless Motion Capturing for Video Surveillance System MMCVS This model has been developed for optimal motion detection in a given video sequence. newline