Detection of Moving Object: A Modular Wavelet Approach

In video surveillance, identification is a very significant element for target tracking, activity recognition, traffic monitoring, military etc. The identification process classifies the pixels into either foreground or background and a common approach used to achieve such a classification is background removal. A Novel method is proposed for the moving object detection based on Modular Wavelet approach, where two consecutive image from image sequences are divided into four parts and then, the Wavelet Energy (WE) is applied to each sub image. The sub image in turn has two energy values of WE, namely, the percentage of energy corresponding to the approximation and the detail. Comparing the energy values corresponding to the detail, the moving object is recognized. Since the discrete wavelet transform has a pleasant property that it can divide an image into four different frequency bands without loss of the spatial information and most of the fake motions in the background can be decomposed into the high frequency wavelet sub-band. Proposed method is compared with existing methods and proposed algorithm gives an enhanced performance.