Fabric defect detection algorithm based on PHOG and SVM

Cuifang, Zhao ; Yu, Chen ; Jiacheng, Ma


In order to effectively improve the detection probabilityfor different types of fabrics and defects, a fabric defectdetection method based on pyramid histogram of edge orientationgradients (PHOG) and support vector machine (SVM) has beenproposed. The algorithm combines fabric texture statisticalmethod and machine learning method. It has two main parts,namely the feature extraction and classification. The detectionprocess mainly includes image segmentation, PHOG featureextraction, SVM model training and detection classification. Thesimulation results show that, based on the detection rate and thefalse alarm rate, the algorithm has a good detection andclassification effect, has a certain robustness, and can be appliedto the actual production department.


Defect detection;Fabric image;Pyramid histogramof edge orientation gradients;Support vectormachin

Full Text: PDF (downloaded 869 times)


  • There are currently no refbacks.
This abstract viewed 1586 times