基于机器视觉的软体纤维丝集束智能计数系统  

Intelligent counting system of soft fibrils collection based on machine vision

作　　者：邢笑雪[1,2] 刘富[1] 马冬梅[3] 翟微微[4] 王芳荣[1] 

机构地区：[1]吉林大学通信工程学院,长春130022 [2]长春大学电子信息工程学院,长春130022 [3]中国科学院长春光学精密机械与物理研究所,130022 [4]吉林大学物理学院,长春130022

出　　处：《吉林大学学报（工学版）》2012年第5期1267-1272,共6页Journal of Jilin University:Engineering and Technology Edition

基　　金：吉林省科技厅重点项目(20080317)

摘　　要：设计了一种基于机器视觉的可以对软体纤维丝集束实现自动计数的检测系统。系统采用自制半球形LED光源照明,对软体纤维丝集束切片,经显微光学系统放大,再由电荷耦合器件(Charge coupled device,CCD)采集放大图像至计算机,通过图像处理系统计算出集束中软体纤维丝数量,并与标准值比较,自动给出是否合格的判别结果。并提出了一种图像处理方法,该方法首先采用基于区域熵值最大的原则将不同光强照射下获取的源图像进行融合,再对融合后的图像利用自组织特征映射(Self organization feature map,SOFM)神经网络求取分割阈值,然后使用求取的分割阈值作为测度指导源图像实现二值化融合,最后采用基于统计量的边界分离和计数方法实现纤维丝集束的计数。实验证明,该系统检测误差不大于1%,重复测量标准差不超过0.07,实现了对软体纤维丝集束的智能计数。Based on machine vision,an intelligent counting system on soft fibrils collection was designed.The hemispherical LED source was used in the system,and the fiber slice was manufactured.The fiber slice was imaged by a microscope optical system,recorded by a CCD camera,and transferred by the image grabber into a computer.Then the quantities of the soft fibrils collection were calculated by the image processing system,and whether the numbers were qualified or not could be judged.A novel image processing method was proposed.The original images obtained in different light intensity could be fused based on maximum region entropy.The optimal threshold of the fusion image could be got based on SOFM neural networks.Based on the above threshold,the binary images of the originals could be refused.After that the quantities of the soft fibrils collection could be counted through boundary separation and counting algorithms based on statistical value.The experimental results show that the detection error of the system is less than 1%,the maximum standard deviation is no more than 0.07 and the automatic and intelligent counting function on the soft fibrils collection can be accomplished.

关 键 词：计算机应用 区域熵值最大 自组织特征映射 图像融合 软体纤维丝集束 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术] TP317.4[自动化与计算机技术—计算机科学与技术]