Cluster-based 3D reconstruction of aerial video

Scott M. Sawyer, Karl Ni, Nadya Bliss

Research output: Chapter in Book/Report/Conference proceedingConference contribution

3 Citations (Scopus)

Abstract

Large-scale 3D scene reconstruction using Structure from Motion (SfM) continues to be very computationally challenging despite much active research in the area. We propose an efficient, scalable processing chain designed for cluster computing and suitable for use on aerial video. The sparse bundle adjustment step, which is iterative and difficult to parallelize, is accomplished by partitioning the input image set, generating independent point clouds in parallel, and then fusing the clouds and combining duplicate points. We compare this processing chain to a leading parallel SfM implementation, which exploits fine-grained parallelism in various matrix operations and is not designed to scale beyond a multi-core workstation with GPU. We show our cluster-based approach offers significant improvement in scalability and runtime while producing comparable point cloud density and more accurate point location estimates.

Original languageEnglish (US)
Title of host publication2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012
DOIs
StatePublished - 2012
Externally publishedYes
Event2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012 - Waltham, MA, United States
Duration: Sep 10 2012Sep 12 2012

Other

Other2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012
CountryUnited States
CityWaltham, MA
Period9/10/129/12/12

Fingerprint

Antennas
Cluster computing
Processing
Scalability
Graphics processing unit

ASJC Scopus subject areas

  • Software

Cite this

Sawyer, S. M., Ni, K., & Bliss, N. (2012). Cluster-based 3D reconstruction of aerial video. In 2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012 [6408681] https://doi.org/10.1109/HPEC.2012.6408681

Cluster-based 3D reconstruction of aerial video. / Sawyer, Scott M.; Ni, Karl; Bliss, Nadya.

2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012. 2012. 6408681.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sawyer, SM, Ni, K & Bliss, N 2012, Cluster-based 3D reconstruction of aerial video. in 2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012., 6408681, 2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012, Waltham, MA, United States, 9/10/12. https://doi.org/10.1109/HPEC.2012.6408681
Sawyer SM, Ni K, Bliss N. Cluster-based 3D reconstruction of aerial video. In 2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012. 2012. 6408681 https://doi.org/10.1109/HPEC.2012.6408681
Sawyer, Scott M. ; Ni, Karl ; Bliss, Nadya. / Cluster-based 3D reconstruction of aerial video. 2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012. 2012.
@inproceedings{93ee33a8278e4e4eb83a7e51f77f5513,
title = "Cluster-based 3D reconstruction of aerial video",
abstract = "Large-scale 3D scene reconstruction using Structure from Motion (SfM) continues to be very computationally challenging despite much active research in the area. We propose an efficient, scalable processing chain designed for cluster computing and suitable for use on aerial video. The sparse bundle adjustment step, which is iterative and difficult to parallelize, is accomplished by partitioning the input image set, generating independent point clouds in parallel, and then fusing the clouds and combining duplicate points. We compare this processing chain to a leading parallel SfM implementation, which exploits fine-grained parallelism in various matrix operations and is not designed to scale beyond a multi-core workstation with GPU. We show our cluster-based approach offers significant improvement in scalability and runtime while producing comparable point cloud density and more accurate point location estimates.",
author = "Sawyer, {Scott M.} and Karl Ni and Nadya Bliss",
year = "2012",
doi = "10.1109/HPEC.2012.6408681",
language = "English (US)",
isbn = "9781467315760",
booktitle = "2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012",

}

TY - GEN

T1 - Cluster-based 3D reconstruction of aerial video

AU - Sawyer, Scott M.

AU - Ni, Karl

AU - Bliss, Nadya

PY - 2012

Y1 - 2012

N2 - Large-scale 3D scene reconstruction using Structure from Motion (SfM) continues to be very computationally challenging despite much active research in the area. We propose an efficient, scalable processing chain designed for cluster computing and suitable for use on aerial video. The sparse bundle adjustment step, which is iterative and difficult to parallelize, is accomplished by partitioning the input image set, generating independent point clouds in parallel, and then fusing the clouds and combining duplicate points. We compare this processing chain to a leading parallel SfM implementation, which exploits fine-grained parallelism in various matrix operations and is not designed to scale beyond a multi-core workstation with GPU. We show our cluster-based approach offers significant improvement in scalability and runtime while producing comparable point cloud density and more accurate point location estimates.

AB - Large-scale 3D scene reconstruction using Structure from Motion (SfM) continues to be very computationally challenging despite much active research in the area. We propose an efficient, scalable processing chain designed for cluster computing and suitable for use on aerial video. The sparse bundle adjustment step, which is iterative and difficult to parallelize, is accomplished by partitioning the input image set, generating independent point clouds in parallel, and then fusing the clouds and combining duplicate points. We compare this processing chain to a leading parallel SfM implementation, which exploits fine-grained parallelism in various matrix operations and is not designed to scale beyond a multi-core workstation with GPU. We show our cluster-based approach offers significant improvement in scalability and runtime while producing comparable point cloud density and more accurate point location estimates.

UR - http://www.scopus.com/inward/record.url?scp=84873539244&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873539244&partnerID=8YFLogxK

U2 - 10.1109/HPEC.2012.6408681

DO - 10.1109/HPEC.2012.6408681

M3 - Conference contribution

AN - SCOPUS:84873539244

SN - 9781467315760

BT - 2012 IEEE Conference on High Performance Extreme Computing, HPEC 2012

ER -