In a previous study, H. Wang from our group developed a reference method for automatic and accurate measurement of cotton fiber length using image analysis. To improve on this, a new imaging system has been developed that will decrease the computation time for the length measurement algorithm and can also deliver higher-resolution images suitable for maturity measurements. The main idea is to develop a prototype system that can perform both length and maturity measurements of cotton fibers simultaneously, quickly and accurately so that a bivariate distribution of cotton can be obtained.

The improved system runs a faster version of the length algorithm that has already been validated on 20 cotton samples of 500 fibers each for a total of 10,000 fibers. For maturity algorithm, three different features are determined from a fiber image: 1) Change in fiber width, 2) Number of convolutions per unit length, and, 3) Light transmitted through fibers. Mature fibers generally have more convolutions per unit length and hence show more changes in fiber width. In addition, mature fibers transmit less light through it, and are more likely to show greater intensity variations along its length.