We propose a deep convolutional object detector for automated driving applications that also estimates classification, pose and shape uncertainty of each detected object. The input consists of a multi-layer grid map which is well-suited for sensor fusion, free-space estimation and machine learning. Based on the estimated pose and shape uncertainty we approximate object hulls with bounded collision probability which we find helpful for subsequent trajectory planning tasks. We train our models based on the KITTI object detection data set. In a quantitative and qualitative evaluation some models show a similar performance and superior robustness compared to previously developed object detectors. However, our evaluation also points to undesired data set properties which should be addressed when training data-driven models or creating new data sets.


    Zugriff

    Zugriff prüfen

    Verfügbarkeit in meiner Bibliothek prüfen

    Bestellung bei Subito €


    Exportieren, teilen und zitieren



    Titel :

    Capturing Object Detection Uncertainty in Multi-Layer Grid Maps


    Beteiligte:


    Erscheinungsdatum :

    01.06.2019


    Format / Umfang :

    3883895 byte





    Medientyp :

    Aufsatz (Konferenz)


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    CAPTURING OBJECT DETECTION UNCERTAINTY IN MULTI-LAYER GRID MAPS

    Wirges, Sascha / Reith-Braun, Marcel / Lauer, Martin et al. | British Library Conference Proceedings | 2019



    Learned Enrichment of Top-View Grid Maps Improves Object Detection

    Wirges, Sascha / Yang, Ye / Richter, Sven et al. | IEEE | 2020


    Deep Generic Dynamic Object Detection Based on Dynamic Grid Maps

    Yan, Rujiao / Schubert, Linda / Kamm, Alexander et al. | IEEE | 2024


    Fully convolutional neural networks for dynamic object detection in grid maps

    Piewak, Florian / Rehfeld, Timo / Weber, Michael et al. | IEEE | 2017