Sistem pengesanan kerosakan pada jalinan tekstil /
Zalili binti Musa, Rohani Abu Bakar, Tuty Asmawaty Abdul Kadir
- Kuantan, Pahang : KUKTEM, 2006
- xvii, 211 p. : ill. ; 30 cm.
Research final report - vot RDU 05/1/17 (KUKTEM)
The purpose of this study was to determine the best methodology, techniques, and algorithms in detecting defects on textile webs through several tests on the prototypes developed in the laboratory. Existing methodologies are based on statistical methods that emphasize on pixel density and relationship of grey level among neighboring pixels. This method, however, is not very effective in detecting defects on textile webs. From a series of tests conducted in this study, a methodology that was enhanced with a pre-processor module (Lili Methodology) was able to detect defects on all the five types of textile webs at 100% level. Four main modules used in this methodology were image acquisition, pre-processor, extraction, and texture analysis (see Figure 1). For each module, several enhanced techniques and algorithms that were based on existing techniques and algorithms were used to produce optimum results.