This essay covers image recognition based object tracking system improvement for UAV. Established system, using computer vision, detects and tracks the object and controls the flying robot. Similar systems were analyzed in the scientific articles, which are using the methods and algorithms for recognition and object tracking. Considered already used markers and their resistance to difficult environmental conditions. In this paper we verified the applicability of methods and algorithms to the flying robot system. In this paper we verified the applicability of methods and algorithms to the flying robot system. Also in this paper we consider image recognition based object tracking system, were studied a variety of methods for the detection and tracking, components necessary to build a system were selected, for designing engineering and software parts of the system, the results of the system and the algorithms used were estimated. The experiment found that the improved adaptive threshold is best to improve the process of recognizing, identifying and tracking an object. Using an improved method of adaptive threshold, object recognition marker accuracy of flying robot system increased from 66.0% to 87.3%. The accuracy of identification of the object marker increased from 76.1% to 88.3%.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Skraidančio roboto vaizdo atpažinimu grindžiamos objekto sekimo sistemos tobulinimas ; Image recognition based object tracking system improvement for UAV


    Beteiligte:

    Erscheinungsdatum :

    2016-06-07


    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Lithuanian , Englisch



    Klassifikation :

    DDC:    629