Hyperspectral Estimation Of Apple Tree Canopy Nitrogen Contents At Different Phenological Phases

Nitrogen is one of the important nutrients for plant growth, it played an important role in apple tree growth and the fruit quality. Previous research mainly focused on the field crops,little research study the apple tree and one stage of the growth phenology. Estimation nitrogen contents of the whole phenology stage of the apple tree could make us known the transportion and utilization of the nitrogen contents and determined the optimum estimation stage.Taking Qixia County of Shandong Province as the study area, Red Fuji was selected to be the research objects. The study was conducted in florescence stage(April), new shoots flourishing stage(May), spring shoots stop growing stage(June), autumn shoots stop growing stage(September) of 2013 and 2014. Canopy hyperspectral data was acquired with ASD Field Spec 4, meanwhile mensurating the nitrogen contents. Variation trends and the hyperspectral characteristics of the apple tree canopy at different phenological phasess were analyzed, water absorption bands were eliminated, data was transformed after standardization and smoothing. The main results were as follows:(1) The canopy nitrogen contents variation trends and the characteristics of the canopy hyperspectral variation trends were figured outThe mean canopy nitrogen contents have been fallen from florescence stage to autumn shoots stop growing stage. The reduction level of the canopy nitrogen contents is comparatively small from new shoots flourishing stage to spring shoots stop growing stage,the reflectance and the status of the nitrogen contents is relatively stable, so new shoots flourishing stage and spring shoots flourishing stage can be taken as the optimizing period for analyzing canopy nitrogen contents of the apple tree.Canopy reflectance varies at different phenological phases, but shows the similar changing trends. The reflecntance D-value almost the same in the florescence stage, new shoots flourishing stage and autumn shoots stop growing stage at 350~780 nm, the reflectance in the florescence stage was higher than those of other periods. The reflectance D-valuereached maximum at 680 nm(red valley). There formed a high near infrared reflectance platform at 1000~1300 nm. Reflectance of the florescence stage was significantly lower than those of other stages, the differences reached maximum and showed the trends that the reflectance of the new shoots flourishing stage was higher than reflectance of the spring shoots stop growing stage with the smallest reflectance appeared in the autumn shoots stop growing stage. There existed cross for four stages at 1419~2402 nm.(2) Sensitive bands and sensitive spectral indices of the apple tree canopy nitrogen contents at different phenological phases were determinedThe green light region(560~590 nm) and near infrared region(780~1100 nm) has been found to be sensitive to canopy nitrogen contents of the florescence stage. SDg, CIGREEN and NDVI could be taken as the sensitive spectral indices. Sensitive bands of the new shoots flourishing stage mainly concentrated on near infrared region and sensitive spectral indices were DVI, MSAVI and TVI. Red region(620~760 nm) and near infrared region were sensitive bands for estimating spring shoots stop growing stage, λr,MCARI,TCARI were selected as the seneitive spectral indices. Near infrared region was sensitive to canopy nitrogen contents of autumn shoots stop growing stage and DVI, MSAVI, TVI, MCARI were used as the sensitive spectral indices.(3)Optimized estimation models of apple canopy nitrogen contents at different phenological phases were establishedOptimized models were BP neural network models based on sensitive bands selected by SMLR from the accuracy of the models. Models of spring shoots stop growing stage were more accurate than models of new shoots flourishing stage, autumn shoots stop growing stage models was unstable with low prediction accuracy. Thus spring shoots stop growing stage chould be taken as the optimized stage for estimating canopy nitrogen contents.