Human Tracking And Applications In Surveillance

The aim of this project is implementing a human detection and tracking system,when amoving object is detected,algorithms will start working,many current methods for detectingand tracking human rely on color contrast or movement to segment the image,usingcolor,however,requires the target to be in constant motion relative to the background, oftenrequiring stationary cameras. Background subtraction is the most commonly used techniquefor object detection. Background subtraction techniques for object detection from videosequence use the concept of subtracting the background model or a reference model from thecurrent image.The methods considered in this project use various techniques for building thebackground model. There are different background subtraction algorithms like framedifferencing, approximiate median, Gaussian mixture this algorithms are tested for detectionmulti objects in different scene using recorder from stationary camera and compared theirperformance in speed and memory requirements for select the best one. after that trackingusing Kalman filter, without need to background subtraction for detect moving humanMotion detection and tracking can be used to locate moving objects and segment them fromthe background for this case we use optical flow technique. For simplicity to user tested ouralgorithms we present interface graphique GUI in MATLAB(R2010a) for detection andtracking process.