In this work, a computer vision algorithm for the detection and recognition of 2D kinematic diagrams, both from paper schemes and digital files, was developed. Furthermore, it runs even with hand-made diagrams, which can be correctly identified. The algorithm is mainly based on the use of the free computer vision library OpenCV, being able to identify each element of the kinematic diagram, its connection with the other elements and store its pixels, which will allow in future research the implementation of motion in the sketches themselves. Allowed elements are revolute, prismatic, fixed, cylindrical and rigid joints and rigid bars. The main applications of this work are focused on the teaching world, communication of ideas in a quickly and graphical way and for fast and preliminary designs of new mechanisms as people can draw the diagram in a Tablet or paper and simulate it in real time, avoiding the necessity to learn how to operate a specialized simulation software and the time it takes to prepare the virtual model and obtain its results.


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Automatic identification of kinematic diagrams with computer vision



    Erscheinungsdatum :

    2023-08-03



    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :

    DDC:    629





    Kinematic-Vision Residuals Analysis

    Nickels, Kevin / Bajracharya, Max / Trebi-Ollennu, Ashitey et al. | IEEE | 2007



    Automatic tractor guidance with computer vision

    Reid,J.F. / Saercy,S.W. / Univ.of Illinois,Agricultural Engineering Dep.,US et al. | Kraftfahrwesen | 1987


    Automatic tractor guidance with computer vision

    Reid, J.F. / Searcy, S.W. | Tema Archiv | 1987