摘要
鸡蛋体积与表面积传统手工检测方法速度慢、效率低,不能满足实际生产的需求。利用机器视觉技术,假设理想鸡蛋图像关于纵径对称,提出像素体积和Vp与像素表面积和Sp,给出在数字图像中的计算方法;最后建立了鸡蛋体积V、表面积S与Vp、Sp之间的关系模型。试验验证表明:鸡蛋体积预测模型相关系数为0.965,实测误差±1 cm3内准确率达92%;鸡蛋表面积预测模型相关系数为0.971,实测误差±1 cm2内准确率达88%。
The conventional manual methods for egg volume(V) and surface area'(S) detection are usually slow,inefficient and not able to meet the actual production requirements,so machine vision technology was researched to replace them.Assumed the ideal egg image was symmetric about longitudinal diameter,the definition of volume pixel(Vp),surface areas pixel(Sp) and the calculation were proposed.Then,the module between volume(surface area) and Vp(Sp) was established.Experimental results showed that the determination coefficient of volume module and surface area was 0.965 and 0.971,respectively;the test accuracy of egg volume module and surface area reached 92% with ±1cm3 error and 88% with ±1cm2 error respectively.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2010年第5期168-171,208,共5页
Transactions of the Chinese Society for Agricultural Machinery
基金
国家"863"高技术研究发展计划资助项目(2007AA10Z212
2007AA10Z203)
"十一五"国家科技支撑计划资助项目(2006BAD30B03)
关键词
鸡蛋
体积
表面积
机器视觉
Egg
Volume
Surface areas
Machine vision
作者简介
周平,博士后,主要从事农业自动化方向研究,E-mail:zhoup@nercita.org.cn
通讯作者:赵春江,研究员,博士生导师,主要从事数字农业技术相关研究,E-mail:zhaocj@nercita.org.cn