Intelligent and objective measurement of wool fibre diameter

澳大利亚每年生产的50多万吨羊毛都由人类羊毛分级员进行目视评估。 这种纤维分级过程是主观的，严重依赖于分级者的经验。 在这个项目中，我们将通过提取人类羊毛分类器使用的特征并结合图像处理和人工智能来客观地测量羊毛纤维直径。 通过基于分形的纹理分析计算的分形维数将与纤维直径相关。这种方法将提供一个洞察到在农场和/或在棚羊毛纤维直径的客观测量。

More than a half million tones of wool produced in Australia per year are visually evaluated by human woolclassers. This fibre-classing process is subjective and heavily dependent on the experience of the classers. In this project, we will objectively measure wool fibre diameter by extracting features used by human woolclassers and by combining image processing and artificial intelligence. The fractal dimension calculated by fractal based texture analysis will be correlated to fibre diameter. This approach will provide an insight into an on farm and/or in shed objective measurement of wool fibre diameter.