摘要
针对光学显微镜采集的纤维图像存在很多纤维交叉现象的问题,提出一种基于局部二进制模式(local binary pattern,LBP)特征的交叉纤维分离算法。通过分割灰度图像提取得到目标纤维区域,经形态学处理,计算交叉点处各个纤维灰度像素的圆形LBP特征之间汉明距离,依据同一根纤维的汉明距离差异较小的原理完成纤维的配对,进而实现交叉纤维的有效分离。为验证算法的准确性,选用不同批次的羊毛纤维,在分离交叉纤维后,根据距离变换公式计算纤维的直径及其变异系数,并与标准测试值进行对比。试验结果表明,该算法具有较高的准确性。
A cross-fiber separation algorithm based on local binary pattern(LBP)was proposed for the presence of many fiber crossover phenomena in the fiber image acquired by optical microscope.The target fiber region was extracted by splitting the grayscale image,and then morphologically processed to calculate the Hamming distance among circular LBP features of each fiber gray scale pixel at the intersection.The fiber matching was done according to the principle that the Hamming distance difference of the same fiber is small,thus achieving effective separation of the cross fiber.To verify the accuracy of the algorithm,different batches of wool fibers were selected.The fiber diameter and its variation coefficient were calculated by distance transformation formula and compared with the standard test values.The experimental results show that the algorithm proposed in this paper has a high degree of accuracy.
作者
洪津津
王荣武
HONG Jinjin;WANG Rongwu(College of Textiles,Donghua University,Shanghai 201620,China)
出处
《东华大学学报（自然科学版）》
CAS
北大核心
2020年第3期395-400,共6页
Journal of Donghua University(Natural Science)
关键词
局部二进制模式
交叉纤维
特征提取
汉明距离
local binary pattern
cross fiber
feature extraction
Hamming distance
