Akademik Veri Yönetim Sistemi
ICHEAS 10TH INTERNATIONAL CONFERENCE ON HEALTH, ENGINEERING AND APPLIED SCIENCES, Dubai, Birleşik Arap Emirlikleri, 1 - 03 Kasım 2025, ss.213-228, (Tam Metin Bildiri)
The objective of this two-year study was to determine and predict nitrogen (N) deficiency stress in (Zea mays L.) maize with high accuracy, addressing the dual challenge of maximizing yield and minimizing environmental pollution from excess fertilizer. The research utilized an integrated approach combining multispectral remote sensing and Artificial Neural Networks (ANN) within a Precision Agriculture framework. A controlled field experiment was established in Samsun-Bafra over two years (2024-2025) using a randomized complete block design with five distinct N fertilizer doses (0, 2.5, 5, 10, and 15 kg/da). Weekly imaging was performed using a UAV-mounted multispectral camera to capture plant indices (NDVI, NDRE, GNDVI) across all growth stages. Results showed that all plant indices increased systematically with N dosage, validating the strong correlation between fertilizer level and plant health, with the 10 $kg/da$ dose identified as close to the economic optimum. The ANN model successfully processed these complex, non-linear relationships to estimate the Chlorophyll Concentration Index and Nitrogen Index. Crucially, the ANN-derived stress maps revealed widespread nitrogen stress (indicated by low index values and blue tones), highlighting the model's capacity to detect subtle, spatial variations in N status that are often missed by visual assessment or simple index mapping. This integrated methodology provides a robust, non-estructive, and high-precision tool, offering farmers a scientific basis for site-specific fertilization decisions, thereby optimizing resource efficiency and enhancing environmental sustainability.