摘要
选用扫描仪获取水稻叶片的数字图像,通过比较第1和第3完全展开叶(L1和L2)颜色参量的空间分布,研究基于机器视觉技术的水稻氮素诊断的最佳叶位和位点选择。结果表明基于机器视觉的水稻氮素营养诊断是有理论依据的,能反映出叶片的营养状况;选择B、b、b/(r+g)、b/r、b/g作为最优颜色特征参量;比较颜色特征参量对应的变异系数CV,得到低氮处理的CV明显高于正常氮素水平,同时CV随着叶位的增加而减小;不同位点的CV其叶尖和叶基的变化幅度较为接近,不同位点间差异不显著。初步研究选择第3完全展开叶作为水稻无损氮素诊断的最佳叶位。
Prior research indicated that leaves at different positions responds differentially to the spectral characteristics under different nitrogen rates.A method based on the computer vision technology was proposed,by comparing the spatial differences of color parameters which was captured from the scanned images of upper fully expanded leaves.The result illustrated that the diagnosis of rice based on the scanned image under different N rates is able to partly reflect the hyperspectral properties.And the B、b、b/(r + g)、b/r、b/g were selected as the optimum color parameters.The coefficient of variation(CV) of the color parameters is higher at low N condition than normal.Furthermore,CV decreases with the increased leaf position.Meanwhile,the difference of CV at different part of the leaf is not obviously.The preliminary research concluded that the third fully expanded leaf can be applied as the ideal indicator to quantify the different status of nitrogen.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2010年第4期179-183,共5页
Transactions of the Chinese Society for Agricultural Machinery
基金
国家自然科学基金资助项目(30571112)
国家"863"高技术研究发展计划资助项目(2006AA10Z204)
浙江省科技计划项目(2007C2308
2008C33008)
关键词
水稻
氮素
机器视觉
叶位
位点
变异系数
Rice
Nitrogen
Computer vision
Leaf position
Different part of the leaf
Coefficient of variation