An Image Analysis for Inspecting Nonwoven Defect

Abstract

Quality and uniformity in nonwoven fabrics is very important, especially when utilized for highly technical purposes. Therefore, the standards for uniformity of nonwovens are growing stricter in applications like filters and battery separators. Traditionally, uniformity evaluation for nonwovens has been the analysis of coefficient of variance (COV), which is often found to be insufficient for inspecting and identifying small defects in the fabric. This paper presents a novel technique based on image segmenting and watershed analysis to inspect defects in the nonwoven. The exact defect areas can be quantified and described through precise calculations. A number of defective nonwoven samples have been tested, showing the method's successful detection and quantification of defects. In addition, the size of investigated image does not appear to greatly affect the result of analysis. Based upon the inspiring findings above, the method should be applicable for on-line monitoring during nonwoven manufacture and promises to reduce manual inspection and costs.