Developing a concept for evaluating defect level in visual quality inspection through digital image processing and fuzzy logic theory

Abstract

Several items in defect check list of the quality inspection on architectural works only use subjective visual inspection, especially those involving aesthetic faults. The inspectors are not able to quantify the value of defect. Subjective evaluations depend on individual experience because they lacks of the quality evaluation standard. The different perception on quality evaluation leads to unreliable results and causes conflicts between people involved in quality inspection. Therefore, the objective of this research is to present a concept of developing defect evaluation system for supporting the visual quality inspection in aesthetic issue of architectural work. The system is divided into two main components: (1) the defect detection and quantification system, (2) the defect level evaluation system. First, the digital image processing technique is applied to analyze the defect feature and quantify the defect value. Second, the standard of defect level evaluation using within organization is developed by using Fuzzy logic, AHP and Knowledge Management concepts. The inspection of tiling work was chosen as a case study for developing a prototype system. The result of the implemented system on an actual construction site was given the most accuracy of position detection at 79% when compared with actual defect positions. The result from defect level evaluation system is more reliable than it from inspectors. The potential benefits of the system are to increase the reliability of visual quality inspection by reducing the subjective human judgment of aesthetic faults, and to reduce conflict about acceptable judgment of defect level.