A key challenge for self-driving vehicle researchers is to curate massive instrumented vehicle data sets. A common task in their development workflow is to extract video segments that meet particular criteria, such as a particular road scenario or vehicle maneuver. We present a novel approach for detecting vehicle maneuvers from monocular dash-cam video building upon a deep learning visual odometry model (DeepV2D) to estimate frame-accurate ego-vehicle movement. We classify movement sequences against reference maneuvers using dynamic time warping and simple heuristics. We show that using deep learning visual odometry to estimate location is superior to consumer-grade high-resolution GPS for this application. We describe and implement a greedy approach to classify maneuvers and evaluate our approach on non-trivial road maneuvers, finding an overall AUROC value of 0.84.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Classifying Ego-Vehicle Road Maneuvers from Dashcam Video


    Beteiligte:


    Erscheinungsdatum :

    01.10.2019


    Format / Umfang :

    2605115 byte




    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Analysis of Dashcam Video for Determination of Vehicle Speed

    Marquez, Alvaro / Leifer, Jack | SAE Technical Papers | 2020


    Analysis of Dashcam Video for Determination of Vehicle Speed

    Leifer, Jack / Marquez, Alvaro | British Library Conference Proceedings | 2020


    Analysis of Dashcam Video for Determination of Vehicle Speed

    Leifer, Jack / Marquez, Alvaro | British Library Conference Proceedings | 2020


    Two-stream neural architecture for unsafe maneuvers classification from dashcam videos and GPS/IMU sensors

    Simoncini, Matteo / de Andrade, Douglas Coimbra / Salti, Samuele et al. | IEEE | 2020


    Classifying Passing Maneuvers

    Jenkins, J. M. / Rilett, L. R. | Transportation Research Record | 2005