摘要
对影响计算机视觉检测玉米籽粒品质的尖端和胚部的识别两个问题进行了研究。利用玉米籽粒尖端的形态特征和胚部图像的亮度特征,分别提出了相应的识别算法。算法对4个品种xhg、xn12、wn14、wc玉米籽粒尖端、表面特征的综合识别率分别为92.50%和89.58%,为特征参数计算奠定了基础。
Two problems were researched, the identification of tip cap and germ surface of corn kernel, which are important to inspect its quality by computer vision. By analyzing morphological features of tip cap and intensity feature of germ surface, two algorithms were established. The algorithms provide overall successful identification of 92.50% for tip cap and 89.58% for germ surface feature among four corn varieties of xhg, xn12, wn14 and wc. This constitutes a great step towards computing feature parameters of corn kernel.
出处
《农业工程学报》
EI
CAS
CSCD
北大核心
2004年第3期117-119,共3页
Transactions of the Chinese Society of Agricultural Engineering
关键词
玉米籽粒
尖端
胚部
识别
计算机视觉
corn kernel
tip cap
germ surface
identification
computer vision
