Fabric Defect Detection Algorithm Based on Image Saliency Region and Similarity Location

In order to solve the problem of defect detection and contour accurate segmentation periodic texture fabric images, a method based on saliency region similarity location is proposed. Firstly, image be detected was processed by color space conversion, Gaussian filtering, contrast enhancement, frequency-tuned (FT) salient algorithm used estimate map enhanced image. The divided into blocks same size with overlapping areas through sliding window, then statistical parameters each block were calculated. outliers in filtered out using inter-quartile range (IQR). Through positioning processing defects, abnormal elimination carried out, outline finally obtained. experimental results show that proposed this paper has better performance terms qualitative characterization Acc, Precision, Recall, F1 score.