基于双端须丛影像法长度直方图计算原棉长度  

Calculation of Raw Cotton Length with Length Histogram Based on Double-side Cluster Image Method

作　　者：金敬业[1] 李莹[2] 杨欢[1] 王府梅[1] 

机构地区：[1]东华大学,上海201620 [2]陕西省纺织科学研究所,陕西西安710038

出　　处：《棉纺织技术》2017年第4期18-21,共4页Cotton Textile Technology

基　　金：国家自然科学基金(NSFC 51673036);中央高校基本科研业务费专项基金(CUSF-DH-D-2014009)

摘　　要：探讨基于双端须丛影像法的原棉长度直方图的原棉长度计算方法。双端须丛影像法是一种快速、精确、低成本的原棉长度检测方法;本文提出将其相对根数曲线(类似HVI曲线)转化为长度质量分布直方图的算法,并通过计算指标,考查了新算法的稳定性与正确性。试验发现:进行5次重复检测时各个长度组质量分数的极差平均不超过1%;t检验也证明,本算法直方图得到的平均长度和变异系数与AFIS值无显著差异。认为:利用本文算法,基于双端须丛影像法将能提供更详细的长度分布信息。Calculation method of raw cotton length with length histogram based on double-side cluster image method was discussed. The double-side cluster image method was a rapid,accurate and low cost test method for raw cotton length. It was put forward its relative counts curve （similar with HVI curve） was transferred to length quality distribution histogram. Through calculation indexes, the stability and correctness of new algorithm were examined. The experiments show the range of mass fraction for each length group are all less than 1 G when five times of duplicate detection are proceeded. T-test shows there is no difference between the average length and variable coefficient obtained from the calculation histogram and AFIS. It is considered length distribution information with more details can be provided based on double-side cluster image method when the algorithm mentioned is used.

关 键 词：双端须丛 图像处理 长度分布 棉纤维 长度直方图 

分 类 号：TS111.9[轻工技术与工程—纺织材料与纺织品设计]