Interpretation of synthetic aperture radar images of concrete by combined uses of image parameters

Abstract

Synthetic aperture radar (SAR) imaging has become an emerging technique in the remote subsurface sensing of construction materials (dielectrics) such as concrete, FRP-concrete, timber, and masonry. Increased center frequency, bandwidth, and synthetic aperture can improve the resolution of SAR images with better contrast and signal-to-noise ratio for condition assessment. However, since most reported SAR applications are for surface sensing, little is known about how to interpret SAR images for subsurface sensing of construction materials. In this paper, extraction, interpretation, and application of parameters from SAR images are presented for condition assessment, using two sides of a CFRP-concrete specimen as an example. A 10-GHz center frequency laboratory SAR imaging system was used for data collection. Seven incident angles were considered (0-deg, 15-deg, 20-deg, 30-deg, 35-deg, 45-deg, and 60-deg.) Multi-dimensional SAR image parameters were defined and applied to the fourteen SAR images of the CFRP-concrete specimen. We have found that performance of SAR image parameters depends on the type of condition assessment problems at hand. Combined use of SAR image parameters of different dimensionalities is encouraged, but a systematic understanding on the physical meaning of each SAR image parameter is necessary. Maximum SAR amplitude is easily affected by background noise. Integrated SAR amplitude in theory performs better than the maximum SAR amplitude. Range curves of SAR images are similar to the A-scan curves in GPR, but they contain more coupling effects from geometry and material's. Cross-range curves of SAR images can also used for surface profiling of concrete specimens. Critical contours of SAR images show the spatial distribution of the backscattering response of concrete. K-R-I curves of SAR images extract features of the images and facilitate quantitative comparison between two SAR images collected at different orientations and image resolutions.