全球旧事资料 分类
龙源期刊网httpwwwqika
comc
高光谱技术在常规水稻种子活力检测中的应

作者:吴小芬赵光武祁亨年来源:《安徽农业科学》2017年第29期
摘要目的实现常规水稻种子活力的快速自动化检测。方法采用高光谱成像技术(波长范围874~1734
m),提取甬籼69和中早392种常规水稻种子未老化、老化48h和老化72h的光谱反射率,在提取样本光谱时采用小波变换(WT)剔除像素点光谱噪声部分,并基于全波段光谱建立了支持向量机(SVM)判别分析模型。结果未老化种子与老化种子可以准确识别,而老化48h种子与老化72h种子之间无法准确识别,与基于种子活力参数的测量结果相符,且不同水稻品种对老化的反应存在差异。结论高光谱成像技术结合化学计量学方法用于种子活力的快速自动化无损检测是可行的。
关键词常规水稻;种子活力;高光谱成像技术;支持向量机;特征波长
中图分类号S3文献标识码A文章编号05176611(2017)29001203
I
spectRiceSeedVigorofCo
ve
tio
alRicebyHyperspectralImagi
gwithChemometricMethods
WUXiaofe
1,ZHAOGua
gwu2,QIHe
g
ia
1
(1SchoolofI
formatio
E
gi
eeri
g,HuzhouU
iversity,Huzhou,Zhejia
g313000;2KeyLaboratoryofAgriculturalProductsQualityImproveme
tTech
ologyi
Zhejia
gProvi
ce,FacultyofAgriculturala
dFoodScie
ce,Zhejia
gAFU
iversity,Li
’a
,Zhejia
g311300)
AbstractObjectiveTheaimwastorealizefasta
dautomaticdectio
ofriceseedvigorMethodHyperspectralimagi
gcoveri
gthespectralra
geof8741734
mwasappliedtodetectco
ve
tio
alriceseedvigorTworiceseedcultivars(
amedYo
gxia
69a
dZho
gzao39)wereusedfora
alysisSpectrali
formatio
ofu
treatedriceseeds,riceseedsacceleratedagi
gfor48hoursa
d72hourswasextractedfromhyperspectralimagesWavelettra
sformwasappliedtoelimi
ateobviousra
dom
oisesofpixelwisespectrabeforespectraextractio
Supportvectormachi
e(SVM)wasappliedtobuilddiscrimi
a
tmodelsusi
gfullspectraTheSVMmodelsofthetworicecultivarsshowedacceptableresultsResultTheu
treatedseedsa
dtreatedseedscouldbeaccuratelydiscrimi
ated,whileseedsacceleratedagi
gfor48hoursa
d72hourscould
otbediscrimi
atedTheresultsmatchedwiththevigortestbytraditio
almethods,a
dtheacceleratedagi
gshoweddiffere
ceso
differe
tseedcultivarsCo
clusio
Thehyperspectralimagi
gcombi
edwithchemometricmethodscouldbeusedforfast,automatica
d
o
i
vasivedetectio
ofriceseedvigor
f龙源期刊网httpwwwqika
comc
KeywordsCo
ve
tio
alrice;Riceseedvigor;Hyperspectralimagi
gtech
ology;SVM;Characteristicwaveba
ds
基金项目浙江省自然科学基金项目(Y14C130046);国家公益性行业(r
好听全球资料 返回顶部