Sparse-PointNet : See Further in Autonomous Vehicles

Wang, Leichen; Goldlücke, Bastian

Sparse-PointNet : See Further in Autonomous Vehicles

Dateien

Zu diesem Dokument gibt es keine Dateien.

Datum

2021

Publikationstyp

Zeitschriftenartikel

Publikationsstatus

Published

Erschienen in

IEEE Robotics and Automation Letters. IEEE. 2021, 6(4), pp. 7049-7056. eISSN 2377-3766. Available under: doi: 10.1109/LRA.2021.3096253

Zusammenfassung

Since the density of LiDAR points reduces significantly with increasing distance, popular 3D detectors tend to learn spatial features from dense points and ignore very sparse points in the far range. As a result, their performance degrades dramatically beyond 50 meters. Motivated by the above problem, we introduce a novel approach to jointly detect objects from multimodal sensor data, with two main contributions. First, we leverage PointPainting [15] to develop a new key point sampling algorithm, which encodes the complex scene into a few representative points with approximately similar point density. Further, we fuse a dynamic continuous occupancy heatmap to refine the final proposal. In addition, we feed radar points into the network, which allows it to take into account additional cues. We evaluate our method on the widely used nuScenes dataset. Our method outperforms all state-of-the-art methods in the far range by a large margin and also achieves comparable performance in the near range.

Fachgebiet (DDC)

004 Informatik

Zitieren

ISO 690

WANG, Leichen, Bastian GOLDLÜCKE, 2021. Sparse-PointNet : See Further in Autonomous Vehicles. In: IEEE Robotics and Automation Letters. IEEE. 2021, 6(4), pp. 7049-7056. eISSN 2377-3766. Available under: doi: 10.1109/LRA.2021.3096253

BibTex

@article{Wang2021Spars-54548,
  year={2021},
  doi={10.1109/LRA.2021.3096253},
  title={Sparse-PointNet : See Further in Autonomous Vehicles},
  number={4},
  volume={6},
  journal={IEEE Robotics and Automation Letters},
  pages={7049--7056},
  author={Wang, Leichen and Goldlücke, Bastian}
}

RDF

<rdf:RDF
    xmlns:dcterms="http://purl.org/dc/terms/"
    xmlns:dc="http://purl.org/dc/elements/1.1/"
    xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
    xmlns:bibo="http://purl.org/ontology/bibo/"
    xmlns:dspace="http://digital-repositories.org/ontologies/dspace/0.1.0#"
    xmlns:foaf="http://xmlns.com/foaf/0.1/"
    xmlns:void="http://rdfs.org/ns/void#"
    xmlns:xsd="http://www.w3.org/2001/XMLSchema#" > 
  <rdf:Description rdf:about="https://kops.uni-konstanz.de/server/rdf/resource/123456789/54548">
    <dc:contributor>Wang, Leichen</dc:contributor>
    <dcterms:available rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-08-11T09:09:55Z</dcterms:available>
    <dc:creator>Goldlücke, Bastian</dc:creator>
    <bibo:uri rdf:resource="https://kops.uni-konstanz.de/handle/123456789/54548"/>
    <dc:creator>Wang, Leichen</dc:creator>
    <dcterms:title>Sparse-PointNet : See Further in Autonomous Vehicles</dcterms:title>
    <dc:contributor>Goldlücke, Bastian</dc:contributor>
    <dspace:isPartOfCollection rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dcterms:issued>2021</dcterms:issued>
    <dc:date rdf:datatype="http://www.w3.org/2001/XMLSchema#dateTime">2021-08-11T09:09:55Z</dc:date>
    <dcterms:isPartOf rdf:resource="https://kops.uni-konstanz.de/server/rdf/resource/123456789/36"/>
    <dc:language>eng</dc:language>
    <foaf:homepage rdf:resource="http://localhost:8080/"/>
    <dcterms:abstract xml:lang="eng">Since the density of LiDAR points reduces significantly with increasing distance, popular 3D detectors tend to learn spatial features from dense points and ignore very sparse points in the far range. As a result, their performance degrades dramatically beyond 50 meters. Motivated by the above problem, we introduce a novel approach to jointly detect objects from multimodal sensor data, with two main contributions. First, we leverage PointPainting [15] to develop a new key point sampling algorithm, which encodes the complex scene into a few representative points with approximately similar point density. Further, we fuse a dynamic continuous occupancy heatmap to refine the final proposal. In addition, we feed radar points into the network, which allows it to take into account additional cues. We evaluate our method on the widely used nuScenes dataset. Our method outperforms all state-of-the-art methods in the far range by a large margin and also achieves comparable performance in the near range.</dcterms:abstract>
    <void:sparqlEndpoint rdf:resource="http://localhost/fuseki/dspace/sparql"/>
  </rdf:Description>
</rdf:RDF>

Universitätsbibliographie

Ja

Begutachtet

Ja