Monitoring of isparta oil rose (Rosa damascena mill.) gardens with unmanned aerial vehicles (UAVS) and development of yield prediction models

This thesis research was carried out in the 2019-2020 vegetation period in order to periodically determine the parameters effective on the yield in the areas where organic and conventional agriculture is grown in Isparta rose (Rosa Damascena Mill.) and to develop yield estimation models by combining them with spectral measurements. In the study, the phenological, morphological and biochemical changes of the Isparta rose were examined periodically, and the effects of biotic and abiotic stress factors during the vegetation period were determined in the trial and test area. By using hyperspectral data, the spectral signatures of the organic and conventionally cultivated Isparta rose were created. NDVI, NDRE, MCARI, TGI, VARI, GNDVI, GRVI, SAVI, LCI, plant vegetation indices were derived from multispectral images using Unmanned Aerial Vehicles (UAV) for the early yield evaluation of Isparta rose in organic ve conventional farming ve Soil moisture, pH, EC and soil temperature properties effective on yield were developed by using multivariate regression (MLR), multivariate adaptive regression curves (MARS), decision trees (CHAID, ExCHAID, CART, Random Forest) and artificial neural network algorithms (ANN) to develop prediction models. Early period (day 69) for organic rose of Isparta yield estimation; MARS 1 (0.907 R2), ExCHAID 1 (0.888 R2), CART 1 (0.931 R2) and RF 1 (0.909 R2) models were determined to be usable. ON the other hand, in conventional Isparta rose at Day 84 (3rd Period) it has been determined that MRL 3 (0.999 R2), MARS 3 (0.903 R2), CHAID (0.861 R2), ExCHAID 3 (0.861 R2), CART 3 (0.995 R2) and RF 3 (0.868 R2) models can be used for yield estimation. As a result of the study, models that can be used for different vegetation periods have been produced in the estimation of yield with UAV of Isparta rose grown under organic and conventional agriculture systems for the Agriculture 4.0 transition period in our country. In addition, it has been suggested that UAV and spectral data should be used for early yield estimation, especially in sustainable agricultural systems.