DOE Patents title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Superpixel edges for boundary detection

Abstract

Various embodiments presented herein relate to identifying one or more edges in a synthetic aperture radar (SAR) image comprising a plurality of superpixels. Superpixels sharing an edge (or boundary) can be identified and one or more properties of the shared superpixels can be compared to determine whether the superpixels form the same or two different features. Where the superpixels form the same feature the edge is identified as an internal edge. Where the superpixels form two different features, the edge is identified as an external edge. Based upon classification of the superpixels, the external edge can be further determined to form part of a roof, wall, etc. The superpixels can be formed from a speckle-reduced SAR image product formed from a registered stack of SAR images, which is further segmented into a plurality of superpixels. The edge identification process is applied to the SAR image comprising the superpixels and edges.

Inventors:
;
Issue Date:
Research Org.:
Sandia National Lab. (SNL-NM), Albuquerque, NM (United States)
Sponsoring Org.:
USDOE
OSTI Identifier:
1261606
Patent Number(s):
9389311
Application Number:
14/927,066
Assignee:
Sandia Corporation (Albuquerque, NM)
Patent Classifications (CPCs):
G - PHYSICS G01 - MEASURING G01S - RADIO DIRECTION-FINDING
G - PHYSICS G06 - COMPUTING G06T - IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
DOE Contract Number:  
AC04-94AL85000
Resource Type:
Patent
Resource Relation:
Patent File Date: 2015 Oct 29
Country of Publication:
United States
Language:
English
Subject:
47 OTHER INSTRUMENTATION; 99 GENERAL AND MISCELLANEOUS

Citation Formats

Moya, Mary M., and Koch, Mark W. Superpixel edges for boundary detection. United States: N. p., 2016. Web.
Moya, Mary M., & Koch, Mark W. Superpixel edges for boundary detection. United States.
Moya, Mary M., and Koch, Mark W. Tue . "Superpixel edges for boundary detection". United States. https://www.osti.gov/servlets/purl/1261606.
@article{osti_1261606,
title = {Superpixel edges for boundary detection},
author = {Moya, Mary M. and Koch, Mark W.},
abstractNote = {Various embodiments presented herein relate to identifying one or more edges in a synthetic aperture radar (SAR) image comprising a plurality of superpixels. Superpixels sharing an edge (or boundary) can be identified and one or more properties of the shared superpixels can be compared to determine whether the superpixels form the same or two different features. Where the superpixels form the same feature the edge is identified as an internal edge. Where the superpixels form two different features, the edge is identified as an external edge. Based upon classification of the superpixels, the external edge can be further determined to form part of a roof, wall, etc. The superpixels can be formed from a speckle-reduced SAR image product formed from a registered stack of SAR images, which is further segmented into a plurality of superpixels. The edge identification process is applied to the SAR image comprising the superpixels and edges.},
doi = {},
journal = {},
number = ,
volume = ,
place = {United States},
year = {2016},
month = {7}
}

Works referenced in this record:

Oil spill detection based on a superpixel segmentation method for SAR image
conference, July 2014


Traversability classification for UGV navigation: a comparison of patch and superpixel representations
conference, October 2007


Building detection in SAR imagery
conference, May 2015