A Background Reconstruction Algorithm Based on Pixel Intensity Classification

Construction of a Robust Background Model for Moving Object Detection in Video Sequence

“…Optical flow is a complex algorithm, it required a huge computation and it is poor for real time, also it is very sensitive to noise, besides, a special hardware is needed for this method, thus, it is practically very poor [4,6,7 ] . 2.2.…”

“…complete region of object cannot be completely extracted, it can extract the boundary of the object only). In addition, it is extra sensitive to threshold value, and it need a supportive algorithm that can be used to detect stationary objects [4,6,7]. 2.3 Background Subtraction: this method subtract the current frame from reference frame ( background image ) which contains non moving object.…”

“…The simplest method to background modeling is Time Average Background Image (TABI) in which a background image is obtained by averaging a number of consecutive frames, this method is efficient with simple situations with no moving objects, but it is not robust to complex scene which contain a number of moving objects, it combine foreground objects with background image [3,4 ]. Then background reconstruction methods has been extended by modeling the pixel intensity as a Mixture of Gaussian ( MoG) , this method can cope with slow changes in illumination, and long term changes in the scene, but on the other hand , it is computationally expensive , and it's parameters required carful estimation [3,2,5].…”

Implementation of moving object detection using angle change method This paper conclude the work in terms of input and output parameter that has been considered while detecting the moving object in video surveillance. It also provides with a look up of our work area. This paper presents a discussion of various techniques like Background subtraction, Block matching, Frame differencing, Optical Flow used in motion detection and also include the comparative analysis of two algorithms first is Colour method(Sp

“…In proposed algorithm, the pre-training of no-moving object in background and the models of background and target aren't needed, and only one parameter is adjusted. Simulation results to real video surveillance sequences show that background can be reconstructed correctly, so target can be extracted perfectly and tracked successfully [2]. Masayuki Yokoyama, TomasoPoggio (2005) [3] purposed a fast and robust approach to the detection and tracking of moving objects.…”

Robust Background Modeling with Kernel Density Estimation

“…A robust background modeling is used to represent each pixel of the background image over time by a mixture of Gaussians [1]. This approach was first proposed by Stauffer and Grimson [2,3] and has become a standard background updating procedure for comparison. Instead of modeling the feature vectors of each pixel by a mixture of several Gaussians [4], Elgammal proposed evaluating the probability of a background pixel using a nonparametric kernel density estimation (KDE) based on very recent historical samples in the image sequence [5].…”