Generating 3d Point Clouds from a Single SAR Image Using 3D Reconstruction Network

Obtaining the three-dimensional data of the target is very useful for the interpretation and application of the SAR target. This paper proposes a deep learning framework to recover the three-dimensional structure of the target from a single SAR image, which is expressed in the form of 3D point cloud. Due to the small data set of SAR images, the network is combined by two parts. First, the twodimensional image in the optical perspective is predicted from the SAR target image, and then the 3D points of the target is reconstructed based on the pre-trained 3D reconstruction network model from the optical images. The experiment is based on the MSTAR datasets. The results confirm the effectiveness of the three-dimensional reconstruction method.

[1]  Leonidas J. Guibas,et al.  ObjectNet3D: A Large Scale Database for 3D Object Recognition , 2016, ECCV.

[2]  Alec Jacobson,et al.  Thingi10K: A Dataset of 10, 000 3D-Printing Models , 2016, ArXiv.

[3]  H Leung,et al.  Detection, Location, and Imaging of Fast Moving Targets Using Multifrequency Antenna Array SAR , 2001 .

[4]  Kostas Papathanassiou,et al.  First demonstration of airborne SAR tomography using multibaseline L-band data , 1999, IEEE 1999 International Geoscience and Remote Sensing Symposium. IGARSS'99 (Cat. No.99CH36293).

[5]  Silvio Savarese,et al.  3D-R2N2: A Unified Approach for Single and Multi-view 3D Object Reconstruction , 2016, ECCV.

[6]  Jiajun Wu,et al.  Learning a Probabilistic Latent Space of Object Shapes via 3D Generative-Adversarial Modeling , 2016, NIPS.

[7]  Xiang-Gen Xia,et al.  Detection, location, and imaging of fast moving targets using multifrequency antenna array SAR , 2004, IEEE Transactions on Aerospace and Electronic Systems.

[8]  Vladlen Koltun,et al.  A Large Dataset of Object Scans , 2016, ArXiv.

[9]  Jianxiong Xiao,et al.  3D ShapeNets: A deep representation for volumetric shapes , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[10]  Alexei A. Efros,et al.  Image-to-Image Translation with Conditional Adversarial Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[11]  Hao Su,et al.  A Point Set Generation Network for 3D Object Reconstruction from a Single Image , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[12]  John A. Saghri,et al.  Three-dimensional target modeling with synthetic aperture radar , 2010, Optical Engineering + Applications.

[13]  Leonidas J. Guibas,et al.  ShapeNet: An Information-Rich 3D Model Repository , 2015, ArXiv.