The analysis of recognition and maintaining of moving object methods is conducted. A general requirement for these algorithms, the use of real-time is specified. Methods frame-to-frame difference method, method of separation background, recognition method using templates, morphological processing method, correlation method and algorithm of tracking the center of mass are considered. Compared and analyzed these methods in terms of ease of implementation and performance. Found out a necessity to develop modified methods of classification of moving objects in the video for the implementation of real-time. А modification of method of classification and maintaining moving target based on the histogram estimation method and identify the moving points method based on frame-to-frame difference method is offered. Based on this information technology developed software Video Object Tracking. The software recognizes the moving points in the video frame sand maintaining moving targets


    Access

    Download


    Export, share and cite



    Title :

    Mathematical providing of recognition and maintaining of moving objects in real time for video




    Publication date :

    2013




    Type of media :

    Article (Journal)


    Type of material :

    Electronic Resource


    Language :

    Unknown





    Tracking of fast moving objects in real time

    Kruglov, A. V. / Kruglov, V. N. | British Library Online Contents | 2016


    Real-Time Segmentation of Moving Objects in a Video Sequence by a Contrario Detection

    Dibos, F. / Pelletier, S. / Koepfler, G. | British Library Conference Proceedings | 2005


    Real-time Video Stabilization for Moving Platforms

    Vermeulen, E. | British Library Conference Proceedings | 2007