摘要
使用光学显微镜观察棒状颗粒时,采集的样品图像中不可避免的会出现颗粒叠加交叉的情况,为准确测量棒状颗粒的形态参数,必须实现棒状交叉颗粒的分离。首先提取颗粒角点,再通过图像细化找到骨架交叉点,计算出颗粒角点到骨架交叉点的距离;依据设定的判断准则,筛选出棒状颗粒的轮廓交叉点,然后根据提出的棒状交叉颗粒分离算法,实现棒状交叉颗粒的分离。实验数据表明所提算法对棒状交叉颗粒分离的准确性大于95%。
Collected sample images unavoidably include overlapping particles,the rod-like particles are observed by using microscope. Therefore,a separation algorithm is proposed in order to get morphological parametersof the rod-like particles. Corners of particle are firstly detected,then crossing points of skeleton of particle are extracted by thinning the image. Secondly,intersections of particle outline are selected by comparing the distance between corners and crossing points of skeleton. Finally,the overlap particles can be separated by using the proposed algorithm. Experimental data show that the accuracy of the separation algorithm for overlapping rod-like particles is more than 95%.
出处
《科学技术与工程》
北大核心
2016年第18期225-229,237,共6页
Science Technology and Engineering
基金
国家科技支撑计划(2012BAK08B04-03)课题资助
关键词
图像处理
棒状颗粒
交叉颗粒
分离算法
image processing
rod-like particle
overlapping particle
separation algorithm
