Use of Hyperpectral Data For Chlorophyll Estimation Based on Leaf Area Indices (LAI) in Wheat

Spectral Indexes are frequently used in the estimation of green parts of plants, usually developed to reduce the spectral effects of external factors such as atmosphere and soil. The aim of this study was to evaluate the ability of different spectral indices to estimate chlorophyll in wheat according to phenological developmental stages and to calculate their optimal band combinations. In this study, chlorophyll-pigment related indices primarily used in chlorophyll estimation, as well as structural and red edge indices were used. Spectral reflectance values obtained for different phenological periods were correlated with SPAD (Minolta-502) values and Partial Least Square (PLS) model was used to calculate the prominent hyperspectral indices and their optimal band combinations. Hyperspectral multispectral data were used to calculate the amount of chlorophyll in wheat according to phenological periods, and the correlation (R²), standard error (RMSE), relative error (% RE) values obtained by evaluating the old and newly developed spectral indices together to calculate the amount of chlorophyll according to phenological periods. Many vegetation indices are strongly influenced by the unfavourable reflection properties of the soil at low LAI. As the Leaf Area Index (LAI) increases, the predictive power of the spectral index increases as the saturation effect decreases. In this study, the responses and sensitivities of different spectral indices for chlorophyll estimation against LAI change in phenological periods were investigated. As a result, the indices that were least and most affected by saturation changes were revealed. Thus, the power of the indices to predict the chlorophyll content of the canopy was demonstrated. Leaf Area Index values obtained from the experimental area in Haymanada between 2013-2014 varied between 1.08-2.81 in the early period (13 May 2014). In chlorophyll estimation, NDVI (705,750) was the least affected by the saturation change due to the increasing LAI value in the early period and showed a high correlation (LAI= 2.63, R²= 0.554). This was followed by Red Edge (740-720), (LAI= 2.63,1.722), NDVI (550,780), (LAI= 2.63,0.733), SRPI (430,680) (LAI= 2.63,0.661), LCCI (705,750) (LAI= 2.63,0.554), and NPCI (430,680) (LAI= 2.63, 0.203). These indices, which showed high correlation in the early period, were in the range of R2=0.836-0.761. In Haymana in the late period between 2013-2014 (26 May,04-12-24 June 2014) LAI values vary between 0.63-3.38 and correlation values are between R2 =0.892-0.862. MSR (705,750) was the least affected by the saturation change due to the increasing LAI value in the late period and showed high correlation (LAI=1.904, 0.906). This was followed by NDVI670 (LAI=1.904,0.703), NDVI550 (LAI=1.904,0.651) and LCCI (LAI=1.904,0.448).