|
Title |
Characteristics of Carbonated Concrete Image and Automatic Detection Technique
|
|
DOI |
https://doi.org/10.4334/JKCI.2017.29.6.589 |
|
Keywords |
탄산화 ; 콘크리트 ; 검출 carbonation ; concrete ; detection |
|
Abstract |
The purpose of this study is to propose an image processing technique for automatically detecting the carbonation region of concrete surface images. The technique consists of five steps: extraction of green intensity image, identification of a partially carbonated image using a silhouette value, detection of carbonated region using threshold and morphology techniques, identification of detection results using a convex hull, and identification of the carbonated regions. Validation tests were performed to evaluate the detection performance of the developed technique using five different partially carbonated concrete images. Test results showed that the developed technique provided accurate detection of the carbonated region except in the image in which aggregates were placed at the boundary between the carbonated region and noncarbonated region.
|