Novel Gaussian mixture model background subtraction method for detecting moving objects

Abstract

Moving object detection is the focus of research and application in the field of computer vision. Background subtraction method is one of the most commonly used methods for moving object detection, in which moving objects in image sequences are detected by comparison of the background model with the current frame. In the process of moving object detection, there are many challenges, such as the interference of clutter background, the influence of illumination, noise and shadow. In this paper, a novel Gaussian mixture model background subtraction method based on wavelet blocks is proposed for the challenge of object detection. This method can not only reduce the influence of illumination, noise and shadow, but also adapt to the dynamic change of natural scene. The contribution lies in the following aspects: (1) A Gaussian background modeling method with less running time is proposed in the background modeling stage. The background is reconstructed based on Gaussian mixture model (GMM) of the mean images of image blocks, aiming to simplify the calculations so as to improve the speed of the corresponding operations. (2) In the foreground detection stage, a wavelet-based de-noising method with the semi-soft threshold function is applied to de-noise the object images of the foreground. Experimental results show that the computational complexity is reduced, while the adaptability and performance are improved by using the proposed method. It was more efficient and robust than traditional approaches.