Background change detection using wavelet transform

Detecting stationary objects in a scene has become a significant factor in image processing for the past few years due to its many applications in the field of safety and security. In most image surveillance systems, subtraction of images is the key to perform detection of objects in a scene. In this paper, the subtraction process will be utilizing the capabilities of Haar wavelet family. The wavelet transform will be employed in separating the foreground from the background as well as other operations and processes in order to come up with only the stationary objects in the scene. The study has three main processes namely: background modeling, subtraction and detection. The median function was used to model the background, Haar wavelet family for the subtraction process, and AND operation and Canny method for the edge detection process. The methods used resulted to satisfactory outputs giving the system a success rate of 95.11% with a confidence level of 100% in detecting 70% of the stationary objects added to or removed from the scene.