摘要
清选是现代农场作业的一个关键环节,从计算机视觉角度研究识别燕麦籽粒时出现的孔洞问题。选取产自山西省五寨县的晋燕8号燕麦籽粒为试验样品,首先采用Otsu将燕麦籽粒从背景分离,该方法分割籽粒时会出现影响识别的孔洞,采用形态学方法中的开运算进行孔洞填充,经试验,设置腐蚀操作的结构元素为9×9,可消除孔洞影响,为燕麦籽粒的进一步识别与统计提供支持。
Clearing is a key link in modern farm operations.This paper studies the hole problems that arise when identifying oat kernels from the perspective of computer vision.The Jinyan No.8 from Wuzhai County was selected as the sample.First,Otsu was used to separate the oat kernels from the background.When the method was used to divide the kernels,holes affecting the identification appeared,and the hole filling was performed by the opening operation in the morphological method.After testing,the structural element of the corrosion operation is set to 9×9,which can eliminate the influence of holes and provide support for further identification and statistics of oat kernels.
作者
燕红文
崔清亮
Yan Hongwen;Cui Qingliang(School of Information Science and Engineering,Shanxi Agricultural University,Taigu 030801,shanxi,China)
出处
《农业技术与装备》
2020年第4期10-11,共2页
Agricultural Technology & Equipment
基金
国家重点研发计划资助项目(2016YFD0701801)。
关键词
晋燕8号
清选
孔洞
分割
Jinyan No.8
clearing
hole
segmentation
