Evaluating Hyperspectral Vegetation Indices for Estimating Nitrogen Concentration of Winter Wheat in Different Growth Stages

Authors : Metin AYDOĞDU, Hakan YILDIZ, Ediz ÜNAL, Seda KÜLEN

Pages : 325-334

Doi:10.30897/ijegeo.673038

View : 15 | Download : 6

Publication Date : 2020-12-06

Article Type : Research Paper

Abstract :Plant health and plant density can be monitored through indices values calculated from reflectance charateristics of plants. This study aims to investigate the effects of different rates of nitrogen fertilizer on spectral reflectance characteristics of dryland winter wheat and to determine and select the most reliable vegetation indices for Nitrogen insert ignore into journalissuearticles values(N); status assesment insert ignore into journalissuearticles values(N contents of leaves); of winter wheat canopy in early, late and whole development stages. The field experiment was carried out in randomized block design with three replications and 0, 40, 80, 120, 160 kg/ha N doses applied between 2012-2013 years in İkizce Farm experiment field. In each plot, spectroradiometer readings were taken during growing period from planting to harvest various plant vegetation indices insert ignore into journalissuearticles values(NDVI, RDVI, SAVI, MTVI, MCARI-1, MCARI-2, TCARI, TVI, SIPI, NPCI, Red Edge insert ignore into journalissuearticles values(750-700);, Red Edge insert ignore into journalissuearticles values(740-720);); were calculated from measured spectroradiometric values. A total of 90 different indices were calculated to obtain the relationship between nitrogen accumulation and hyperspectral indices insert ignore into journalissuearticles values(structural, chlorophyll pigment, red edge indices);. The simple linear determination coefficients insert ignore into journalissuearticles values(R²); between those indices and leaf nitrogen contents in different development stages of early, late, and whole season were calculated. During the early period of tillering to bolting, insert ignore into journalissuearticles values(Feekes 4-7); NDI-1, NDI-2, SR-7, Red Edge indices insert ignore into journalissuearticles values(708-850 nm.); have the highest determination coefficients insert ignore into journalissuearticles values(R²); of 0.643, 0.641, 0.620 with RMSE values of 5.996, 6.039, 6.129 and relative percentage error values insert ignore into journalissuearticles values(%RE); of 23.07, 23.23, 23.58 % respectively. During the period of heading to ripening insert ignore into journalissuearticles values(Feekes 8-10);, PhRI, NDV-3, NDVI-4 visible insert ignore into journalissuearticles values(Green Zone);, red, red edge and near infrared insert ignore into journalissuearticles values(NIR); indices insert ignore into journalissuearticles values(531-800 nm.); showed the highest determination coefficients of 0.734, 0.708, 0.699 with RMS values of 3.089, 3.205, 3.149 and relative percentage error values of 40.37, 41.89, 41.16 % respectively. Considering all growing period in 2013 of tillering to ripening, insert ignore into journalissuearticles values( Feekes 4-10);; SR-14, SRPI, TVI visible area insert ignore into journalissuearticles values(blue + green); and Red Edge indices insert ignore into journalissuearticles values(415-750 nm.); have determination coefficients of 0.742, 0.699, 0.646 and RMS values 1.203, 0.902, 0.697 and relative percentage error values of 7.15, 5.36, 4.14 %, respectively. High determination coefficient insert ignore into journalissuearticles values(R²); between plant nitrogen uptake and reflectance charecteristics were attained as HVI insert ignore into journalissuearticles values(r= 0.806 p<0.0.1 **);, OSAVI insert ignore into journalissuearticles values(r= 0.794, p<0.0.1**);, NDVI insert ignore into journalissuearticles values(r= 0.794, p<0.0.1**); and HNDVI insert ignore into journalissuearticles values(r= 0.793, p<0.0.1**);.
Keywords : Hyperspectral, spectroradiometer, vegetation indices, remote sensing, wheat, leaf area indice LAI,