摘要
为建立智能烘烤系统,自制图像采集系统,实时采集了烘烤过程中的烟叶图像,并对图像进行自适应维纳滤波和基于k均值聚类的彩色分割处理,图像处理结果显示能较好地将烟叶从背景和阴影干扰中分离出来。对分割后烟叶图像在线提取了9个颜色特征,分析了烘烤过程中烟叶颜色的变化,并初步探究了颜色与烘烤工艺曲线的相关联系。分析表明在烘烤过程中烟叶图像的R分量和a*分量逐渐增加,G分量和H分量逐渐降低,且在烘烤后期的变化相对缓慢;B分量、b*分量变化不大;图像经过滤波处理后,I分量和L*分量变化也不大。利用规律变化的H分量和a*分量建立烤烟颜色与烘烤曲线的关系,为密集烤房烘烤操作的智能化提供了有意义的数值依据。
In order to build intelligent curing system,an image acquisition system is made for real- time acquisition of tobacco leaves image during curing. The image is preprocessed by adaptive wiener filtering and k- means clustering algorithms. The results demonstrate that image of tobacco leaves can be separated from the background appropriately. Nine color features are extracted in real time and how they changed during curing is analyzed. The relationship of color and bulk curing schedule is also discussed. Analysis show that R and a* components gradually increase during curing,while G and H components decrease with B and b* components varying slightly. For the reason of image filtering,I and L* components make little change during curing. The relationship of tobacco color and bulk curing schedule is established making use of H and a* components,which provides meaningful value basis for intelligent operation of tobacco bulk curing.
出处
《激光杂志》
北大核心
2015年第6期64-67,71,共5页
Laser Journal
基金
国家自然科学基金(61071191)
重庆市教育科学"十二五"规划项目(2013-ZJ-081)
关键词
图像处理
密集烤房
烘烤曲线
图像特征
Image processing
Bulk curing barn
Bulk curing schedule
Image features
