Detecting Central Region in Weld Beads of DWDI Radiographic Images Using PSO

Detecting Central Region in Weld Beads of DWDI Radiographic Images Using PSO

Fernando M. Suyama (Federal University of Technology, Paraná, Brazil), Andriy G. Krefer (Federal University of Technology, Paraná, Brazil), Alex R. Faria (Federal University of Technology, Paraná, Brazil) and Tania M. Centeno (Federal University of Technology, Paraná, Brazil)
Copyright: © 2015 |Pages: 15
DOI: 10.4018/ijncr.2015010103


This paper presents a methodology to detect the central region of weld beads on petroleum pipelines in double wall double image (DWDI) radiographic images. The method is based on three steps: pre-processing (to isolate selected regions), optimization (to define the ellipse that best fits in selected region), and decision (to choose the best region). Results show that the Particle Swarm Optimization (PSO) algorithm converges satisfactorily to the selection of the region that is most similar to the central region of the weld on the optimization and decision steps (to balance the weights of the classifier). The scientific contribution of this research is the improvement of the method applied in the search of candidate regions through ellipses' attributes analysis.
Article Preview


In petrochemical facilities, networks of fluid-conductive pipes are constructed by attaching pipes and other components by means of welded joints (Telles, 2001). These facilities are designed to withstand harsh conditions of temperature and pressure, although failures still occur. If not dealt with in time, they can result in leaks, unplanned shutdowns, production losses, fires or even accidents with serious environmental damages (Felisberto, 2007). However, periodic inspection programs through non-destructive testing help to prevent such failures. Non-destructive testing (NDT) includes methods and procedures employed to examine materials and products without interfering with its functionality (Kroetz, 2012).

Computer vision and artificial intelligence techniques have been widely explored in research involving the analysis and interpretation of radiographs of welds through digital radiographic inspection, (e.g. Liao & Ni, 1996; Liao & Tang, 1997; Liao & Li, 1998; Wang & Liao, 2002; Mery & Berty, 2003; Carrasco & Mery, 2004; Padua, 2004; Silva, Calôba, Siqueira, & Rebello, 2004; Zhang, Xu, & Ge, 2004; Felisberto, Lopes, Centeno, & Arruda, 2006; Silva & Mery, 2007a; 2007b; Vilar, Zapata, & Ruiz, 2009; Mahmoudi & Regragui, 2009a; 2009b; Thiruganam, Anouncia, & Kantipudi, 2010; Li, Wang, Xu, & Tan, 2010; Valavanis and Kosmopoulos, 2010; Kroetz, Centeno, Delgado, Felisberto, Lucas, Dorini, Fylyk & Vieira, 2012; Kroetz, 2012; Suyama, Krefer, Faria, & Centeno, 2013). Thus, computer-based radiographic inspection techniques may be suitable. A decision support system may help specialists by decreasing the human error and increasing productivity on inspections.

The process of weld detection and extraction is the first step in a NDT system for a fully automatic detection and identification of flaws (e.g. Liao & Ni, 1996; Liao & Li, 1998). This step consists in isolating the weld as a region of interest (ROI), in which defects can be identified and segmented using a diversity of image processing techniques. This research presents a method to search the central region of weld beads in double wall double image (DWDI) automatic detection on welded joints through digital radiographic inspection. This method seeks to reduce the search space, working as a support technique for weld detection in DWDI radiographic images.

The main goal of this research is to identify the central region of weld beads in DWDI radiographic images as regions of interest (ROI) using the Particle Swarm Optimization (PSO) algorithm. The PSO algorithm is applied in such a way that the search does not depend on the initial detection of the pipe in the image. It is an extension of the work presented by Suyama (Suyama et al., 2013) working as a support technique for weld detection and extraction in DWDI radiographic images.

Complete Article List

Search this Journal:
Open Access Articles: Forthcoming
Volume 8: 4 Issues (2019): 1 Released, 3 Forthcoming
Volume 7: 4 Issues (2018)
Volume 6: 2 Issues (2017)
Volume 5: 4 Issues (2015)
Volume 4: 4 Issues (2014)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing