Premium
Edge‐Enhanced Matching for Gradient‐Based Computer Vision Displacement Measurement
Author(s) -
Luo Longxi,
Feng Maria Q.
Publication year - 2018
Publication title -
computer‐aided civil and infrastructure engineering
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.773
H-Index - 82
eISSN - 1467-8667
pISSN - 1093-9687
DOI - 10.1111/mice.12415
Subject(s) - computer vision , artificial intelligence , robustness (evolution) , computer science , contrast (vision) , enhanced data rates for gsm evolution , orientation (vector space) , segmentation , tracking (education) , displacement (psychology) , edge detection , matching (statistics) , process (computing) , pattern recognition (psychology) , image processing , mathematics , image (mathematics) , psychology , pedagogy , biochemistry , chemistry , statistics , geometry , psychotherapist , gene , operating system
Computer vision‐based displacement measurement for structural monitoring has grown popular. However, tracking natural low‐contrast targets in low‐illumination conditions is inevitable for vision sensors in the field measurement, which poses challenges for intensity‐based vision‐sensing techniques. A new edge‐enhanced‐matching (EEM) technique improved from the previous orientation‐code‐matching (OCM) technique is proposed to enable robust tracking of low‐contrast features. Besides extracting gradient orientations from images as OCM, the proposed EEM technique also utilizes gradient magnitudes to identify and enhance subtle edge features to form EEM images. A ranked‐segmentation filtering technique is also developed to post‐process EEM images to make it easier to identify edge features. The robustness and accuracy of EEM in tracking low‐contrast features are validated in comparison with OCM in the field tests conducted on a railroad bridge and the long‐span Manhattan Bridge. Frequency domain analyses are also performed to further validate the displacement accuracy.