Motion Detection Surveillance System for Human Head

Abstract

Motion detection for human head and its coordinates is one of the important areas with multiple applications at the present time, such as control of devices for people with disabilities and other modern control devices estimation model to detect human head and follow up coordinates. This system uses frame subtracting to detect motion with development the background frame is to eliminate the problem of wrong alarm for motion when the background frame still without change. Gaussian smoothing filter also has been used to reduce noise caused by bad illumination or web camera low quality, so there is no need for kalman filter, then using Laplacian filter which is dependent on second derivation in order to detect edges of the frames(background and current) , at last detecting the human head using the sum of gradient magnitude around the perimeter of ellipses equation. The system has been tested on several samples indoor and out door , two of these samples for females it head hooded (with Islamic hijab) and the results were very good in detecting heads and tracking their coordinates. The system has been built using visual C# which is considered faster than interpreted MATLAB that have been used in other researches in the same area.