Özet:

This study aims to develop a method using hyperspectral reflections for the determination of the sheet %N classes of the strategic importance of our country. For this purpose 3 different vegetation stages are represented by the Hoagland sulations (Hoagland and Arnon, 1938) with sufficient and more N content (Hoagland and Arnon, 1938) and 72 test plants in controlled greenhouse conditions, grown in perlit environment, spektroradiometer and 400-1000 nm spectral measurements and leaf samples for determination of %N were made. In the determination of the waves related to different periods and doses in the sheets of sugar pancreas, the stepwise multi-regression analysis has been applied and the analysis of the basic components of the determined 48 different waves reflective value has been selected for the 5 waves (474-517-652-721-961 nm) model that provides the highest contribution if the total varies. Coded using the specified waves, the KDA model identifies 72 plants with 92% accuracy to their actual classes (NNoksan; 92% NYeter; 88% and NFazla; 96%). The 36 test data used for the model validation was determined that the spectral reflections, which were differentiated and selected from the waves in the N% classes (NNoxan; 91, NYeter; 85% and NFazla; 92) with an accuracy of 89% and that the spectral reflections, which were selected from the waves, could be used in the determination of the need for sugar-panked nitrogen fertilization for different vegetation periods with the KDA model. The research resulted in spectral data and studies aimed at determining the plant nutrition status, hopeful findings for the use of discriminatory models were obtained and the KDA model was proposed to be developed using the experimental patterns to be installed for different plant types and nutrient elements. Keywords: nitrogen, hyperspectral reflection, square discriminant, spektroradiometer.

Anahtar Kelimeler: