This paper proposes a recognition network for small-scale objects in cluttered point clouds. The network consists of two components: improved semantic segmentation for large-scale 3D point clouds and an adaptive instantiation algorithm. In semantic segmentation, based on the backbone, we introduce the grid sampling module and the normal-angle feature to improve the efficiency and accuracy of segmentation respectively. Then the network outputs point-wise semantic labels. After that, we propose an adaptive instantiation algorithm to group points that are closely packed together and obtain the objects. In this way, our network completes the recognition of the small-scale objects. We conducted experiments on real aero-engine datasets and the results reveal that the proposed network can recognize a small-sized component in the cluttered point cloud scene of aero-engine.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.