SUBMISSIONS

SUBMISSION DETAIL

Bekir YILDIRIM, Gkhan YAAR, Berrin AHN
 


Keywords:



STATISTICAL FEATURE EXTRACTION BASED METHOD FOR PILLING GRADE EVALUATION
 
Pilling is one of the important quality problems for the consumer which is to be constantly monitored and controlled in textile industry. Traditionally, pilling is evaluated by comparing fabrics after the pilling process with reference to a set of standard images according to EN ISO 12945-2:2000. This subjective evaluation sometimes results in inconsistency not only in between different laboratories but also even in the same laboratory between different evaluators. Objective evaluation approaches using image processing algorithms are actively researched. The main aim of this research is to develop an objective evaluation system using image processing methods and statistical features to obtain fabric image surface features. First step of the proposed method is image preprocessing where non-uniform illumination effects and noise are eliminated. The second step is to extract texture features of the fabric images. Texture features of the non-processed and processed fabric surface images will reveal differences which will lead to classification. For this purpose various statistical features are extracted. In the last step, obtained features are further used to compare with the standard image features. Supervised classification methods are used for pilling grade assessment. The presented research work explains the statistical features extracted from images to pilling grade evaluation.

Anahtar Kelimeler: Pilling Grade, Objective Evaluation, Statistical Features, mage Processing