A modified motion detection algorithm in complex background

Indoor and outdoor real-time monitoring of scenes has become a hot topic for decades. This paper analyzes the median filter and its shortcomings and proposes a modified detection method for moving objects with complex background. Using the method based on decision-making data gives the weights of a consecutive sequence of frames to calculate the weighted average of the corresponding values of pixel in them and the background is modeled for real-time updating. Besides, the thrice threshold processing used to detect and extract objects is presented, which is by using a combination of standard deviation in weighted algorithm and the adaptive threshold algorithm. The experiment results indicate that adopting this method can separate the object from the background and remove the background mistaken for objects and meet the system's real-time requirements of monitoring.