Developing a PSO-Based Projection Algorithm for a Porosity Detection System Using X-Ray CT Images of Permeable Concrete

Yi Zeng Hsieh, Mu Chun Su, Jieh Haur Chen, Bevan Annuerine Badjie, Yu Min Su

Research output: Contribution to journalArticlepeer-review

7 Scopus citations


Permeable concrete is widely used as a road surfacing material due to its sturdiness and ability to be quickly repaired. The porosity of concrete has been used as a predictive indicator for the properties of the concrete. Traditional methods for measuring this porosity are feasible but can be time-consuming. In this paper, we propose a particle swarm optimization-based projection algorithm for visualization of the high-dimensional data as a 2-D scatter plot for detecting porosity in permeable concrete from X-ray computerized tomography images. We regard the proposed projection algorithm as an improved version of Sammon's nonlinear mapping. The projected scatter plot allows for a straightforward analysis of the inherent structure of clusters within scanned images. Several data sets, including artificial data sets and real-life imaging data, were tested to demonstrate the performance of the proposed projection algorithm. The model created in this paper can augment the traditional methods for examining porosity by providing visual images for decision makers to make correct decisions for future problems. With an accuracy of >99%, the visualized images provide a clearer understanding of the inner structure of pervious concrete and enhance the study of the correlation between the properties of the concrete.

Original languageEnglish
Article number8502039
Pages (from-to)64406-64415
Number of pages10
JournalIEEE Access
StatePublished - 2018


  • Cluster analysis
  • PSO algorithm
  • X-ray computerized tomography (CT)
  • automation
  • computational intelligence
  • projection algorithm


Dive into the research topics of 'Developing a PSO-Based Projection Algorithm for a Porosity Detection System Using X-Ray CT Images of Permeable Concrete'. Together they form a unique fingerprint.

Cite this