关键词: 高光谱, 氮素, 马铃薯, PLSR, 精细化农业

Abstract: Crop nitrogen status is the important evaluation index of crop growth status and soil nitrogen level,timely and non destructive monitoring of leaf nitrogen status is significant to optimize nitrogen management and improve grain yield and quality.In the present study,five nitrogen(N)fertilizer treatments were applied to build up levels of nitrogen variation in potato fields.Relationships between canopy spectral reflectance from the AISA Eagle airborne imaging spectrometer data after geometrical and radiometric correction and reconstruction of reflectance spectrum and nitrogen levels in potato leaves were studied.The leaves were sampled and analyzed for leaf N Leaf N content prediction models were developed using PLSR by analyzing the raw reflectance data(R)at 430-910nm,their first derivative D1(R),their second derivative D2(R),their absorbance equivalent\[log(1/R)\],its first derivative\[Dlog(1/R)\]and its second derivative\[D2log(1/R)\].The results indicated that the PLSR model with the first derivative gave the best performance with R2=0.82,RMSEC=0.38%.The best estimation model from PLSR was applied to all the potato pixels of the airborne images.The values of the leaf N distribution map ranged from 3.35% to 5.95%,which were quite consistent with those of laboratory measurements from 3.59% to 5.89%.The distribution agreed highly with the growth status distribution.The results provide an important theoretical and technical foundation for the future research and development of hyperspectral images based for distinguishing spatial variability in N status in potato fields.

Key words: Airborne hyperspectral image, Nitrogen, Potato, Partial least squares regression(PLSR), Precision agriculture