Keywords: automated rebar counting, polar transformation, radial-transition voting, image-to-image translation, projection profile, reinforced-concrete inspection.

Accurately ensuring the quantity of reinforcing bars before concrete pouring is critical for quality assurance in reinforced concrete structures. However, on-site rebar inspections are still carried out manually and restricted to a small number of sample areas, making them time-consuming and potentially unreliable. Consequently, such methods fall short in delivering reliable, and comprehensive quality assurance. To overcome these limitations, this study proposes a novel method for automatically counting rebars from grayscale-converted field images. The binary image in Cartesian coordinates is transformed into a polar coordinate system, such that vertical rebars appear as concentric circles. Instead of counting peaks from the vertical projection profile of rebar pixels in the binary image—as in conventional approaches—this method estimates the number of rebars by counting the number of circles. To enable accurate estimation of the number of concentric circles, we propose a radial transition voting algorithm. This algorithm performs a 360-degree scan from the image center and estimates the number of rebars by counting the intensity transitions along the radial direction in the binary image. This approach enables robust and fully automated rebar counting without the need for threshold tuning or parameter adjustment.

download the full-text of this paper (PDF, 9 pages, 633 Kb)

go to the previous paper
go to the next paper
return to the table of contents
return to the volume description