Image Detection with Parallel Structure Background Addition Using Robust Techniques

Abstract

Motion detection is a very important task for detecting the moving objects in a video sequence, based on background subtraction approaches in computer vision based applications. A background model should be generated at the first time before subtracting it from each image of the sequence. The idea of background subtraction is to subtract or difference the current image from a reference background model. This approach is effective and accurate for moving objects detection and the use of spatial color information was robust to environmental illumination changes. The experimental results are also compared with the results of the background estimation algorithm with Spatial Detection (SD) and Motion Detection with Pyramid Structure of background model (MDPS). Finally, the image is enhanced by Super Resolution Algorithm.