Investigation of the determination of nitrogen content with hyperspectral detection techniques in tomato plant

In this thesis, the determination of the nitrogen content of the tomato plant by hyperspectral methods in controlled greenhouse conditions was investigated. The study was carried out in the greenhouse conditions, two different fertilizers amonyum nitrat and ammonium sulfate were applied to the experimental plants and 5 different doses were applied. A total of 120 flower pots were sampled in four different periods (15 August, 25 August, 5 September and 15 September) for 3 replications. Spectral reflections were measured by using ASD FieldSpec HandHeld spectroradiometer-plant probe apparatus and nitrogen (N), phosphours (P) and potassium (K) analyzes were performed on sampled tomato plant leaves. Spectral reflectance measurements and 1st derivative values and N nutrient levels were evaluated by stepwise multiple regression analysis method and the highest R2 estimation models were formed. Wavelengths which can be used in the prediction of the N element which is the subject of research are determined. Reflection graphs of ammonium nitrate and ammonium sulfate applications, derivative graphs and estimation model equations are compared. Correlation analysis was performed to determine the relationship between the values of N values of the tomato plants fertilized with ammonium nitrate and ammonium sulfate and the values of the vegetation indices. Graphs showing the relation between vegetation indices and N nutrient contents were obtained. According to the findings, significant statistical differences were determined between 1 % and 445, 705, 720, 735, 750, blue (450-500 nm) and red (610-700 nm) wavelengths of spectrum ammonium sulfate applications. There were no significant statistical differences between ammonium nitrate applications and the wavelengths of the spectrum. Ammonium sulfate application, except for the CI index, mND705, SR705, RI indices were statistically significant (P <0.001) was found. Plant indices provide information about plant health. Therefore ammonium sulfate treated plants have been found to be healthier than ammonium nitrate treated plants. In the application of ammonium nitrate, it is concluded that plant indices should be investigated in larger data sets. As a result, it has been found that 75 % of the wavelengths in the visible region of the spectral reflection graph of the application of ammonium nitrate in the determination of N content by hyperspectral detection methods of tomato plant, 25 % of the wavelengths in the near infrared region, 54.2 % of the wavelengths in the visible region of the spectral reflection graph in the ammonium sulfate application, 45.8 % of the wavelenghts in the near infrared region can be determined. It is estimated that espically in the 4th period, ammonium nitrate and ammoniım sulphate applications increase the reflectance percentages in the visible region due to the decreasing does, thus the N content of tomato plant can be determined in this period.