Scalable Track Detection in SAR CCD Images
Abstract
Existing methods to detect vehicle tracks in coherent change detection images, a product of combining two synthetic aperture radar images ta ken at different times of the same scene, rely on simple, fast models to label track pixels. These models, however, are often too simple to capture natural track features such as continuity and parallelism. We present a simple convolutional network architecture consisting of a series of 3-by-3 convolutions to detect tracks. The network is trained end-to-end to learn natural track features entirely from data. The network is computationally efficient and improves the F-score on a standard dataset to 0.988, up fr om 0.907 obtained by the current state-of-the-art method.
- Authors:
-
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
- Publication Date:
- Research Org.:
- Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
- Sponsoring Org.:
- USDOE National Nuclear Security Administration (NNSA), Office of Defense Nuclear Nonproliferation
- OSTI Identifier:
- 1347496
- Report Number(s):
- SAND2017-2423
651687
- DOE Contract Number:
- AC04-94AL85000
- Resource Type:
- Technical Report
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 47 OTHER INSTRUMENTATION
Citation Formats
Chow, James G, and Quach, Tu-Thach. Scalable Track Detection in SAR CCD Images. United States: N. p., 2017.
Web. doi:10.2172/1347496.
Chow, James G, & Quach, Tu-Thach. Scalable Track Detection in SAR CCD Images. United States. https://doi.org/10.2172/1347496
Chow, James G, and Quach, Tu-Thach. 2017.
"Scalable Track Detection in SAR CCD Images". United States. https://doi.org/10.2172/1347496. https://www.osti.gov/servlets/purl/1347496.
@article{osti_1347496,
title = {Scalable Track Detection in SAR CCD Images},
author = {Chow, James G and Quach, Tu-Thach},
abstractNote = {Existing methods to detect vehicle tracks in coherent change detection images, a product of combining two synthetic aperture radar images ta ken at different times of the same scene, rely on simple, fast models to label track pixels. These models, however, are often too simple to capture natural track features such as continuity and parallelism. We present a simple convolutional network architecture consisting of a series of 3-by-3 convolutions to detect tracks. The network is trained end-to-end to learn natural track features entirely from data. The network is computationally efficient and improves the F-score on a standard dataset to 0.988, up fr om 0.907 obtained by the current state-of-the-art method.},
doi = {10.2172/1347496},
url = {https://www.osti.gov/biblio/1347496},
journal = {},
number = ,
volume = ,
place = {United States},
year = {Wed Mar 01 00:00:00 EST 2017},
month = {Wed Mar 01 00:00:00 EST 2017}
}
Save to My Library
You must Sign In or Create an Account in order to save documents to your library.