![[ Visit Client Website ]](images/banner.jpg)
Fiber cross sectional analysis provides fundamental measurements (such as cell wall thickness, area and perimeter) which are directly related to cotton maturity, and hence is regarded as the reference for other indirect methods [1, 2]. Accurate and unbiased cross-section data are crucial for establishing the reference database.
USDA Southern Regional Research Center and International Textile Center established a laboratory protocol of cross-sectioning fibers and acquiring images in a project sponsored by Cotton, Inc. Over 100 species of cotton were collected and their cross-section images were analyzed by using the Fiber Image Analysis Software (FIAS) developed in University of Texas at Austin in 2004 [1, 2, 3]. However, a recent independent study revealed that there is an overestimation of maturity by around 8-9% using the laboratory protocol and the FIAS software because 10-40% of immature fibers could not be detected correctly in the image analysis [4]. Immature fibers have thinner wall, and are more easily to be scratched/shredded by the cutting blade, increasing the difficulty of edge detection.
In this study, we further revised the FIAS to increase the detection rate of immature fibers and accuracy of lumen identifications. We added effective routines to automatically repair damaged fiber edges so that immature fibers could have equal opportunity to be included in the analysis as mature fibers. The image was segmented using an adaptive thresholding algorithm, and small gaps between edge ends were filled before the “flooding” with the background. We re-analyzed all the images of 12 cotton samples used in a previous study, and fiound the new FIAS can retain more than 80% of missed immature fibers. In a future study, we will collaborate with Dr. Eric Hequet to process cross-section images in the cotton maturity reference database created in 2006.