Yazar "Tunca, Emre" seçeneğine göre listele
Listeleniyor 1 - 12 / 12
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Accurate estimation of sorghum crop water content under different water stress levels using machine learning and hyperspectral data(Springer, 2023) Tunca, Emre; Koksal, Eyup Selim; Ozturk, Elif; Akay, Hasan; Taner, Sakine CetinThis study investigates the effects of different water stress levels on spectral information, leaf area index (LAI), and the performance of three machine learning (ML) algorithms in estimating crop water content (CWC) of sorghum. The results show that the spectral reflectance of sorghum varies with growth stage and irrigation treatment, but consistent patterns are observed for each treatment. The LAI of sorghum gradually increased throughout the growth stages, with the most significant variation observed during the flowering stage. In this study, three machine learning-based regression models, namely, extreme gradient boosting (XGBoost), random forest (RF), and support vector machine (SVM), were utilized to estimate sorghum CWC using hyperspectral measurements. Recursive feature elimination (RFE) method was used to select the optimal spectral reflectance wavelengths for the ML models, and principal component analysis (PCA) was used to reduce the dimensionality of the hyperspectral data. The results indicated that the RF model achieved the highest R-2 (0.90) and lowest of RMSE (56.05) value using selected wavelengths, while the XGBoost model demonstrated superior accuracy and reliability in estimating CWC using dimensionality-reduced hyperspectral data (r = 0.96, RMSE = 45.77). Also, the study highlights the importance of vegetation index (VI) in CWC estimate. Some VIs, such as NDVI and MSAVI, performed poorly, while others, such as CL_Rededge and EVI, performed better. The study provides valuable insights into the effects of water stress levels on spectral information, LAI, and the performance of ML algorithms in estimating the CWC of sorghum. The findings have significant implications for precision agriculture, as accurate and reliable estimates of CWC can help farmers optimize irrigation and fertilizer applications, leading to improved crop yields and resource efficiency.Öğe Accurate leaf area index estimation in sorghum using high-resolution UAV data and machine learning models(Pergamon-Elsevier Science Ltd, 2024) Tunca, Emre; Koksal, Eyuep Selim; Ozturk, Elif; Akay, Hasan; Taner, Sakine letinAccurate estimation of leaf area index (LAI) is essential for precision agriculture, yet traditional ground-based measurements are destructive, time-consuming, and limited in scale. This study aimed to address the need for rapid, non-destructive LAI monitoring over large areas by evaluating unmanned aerial vehicle (UAV) data and machine learning (ML) models. A field experiment with four irrigation treatments was conducted to obtain wide range of LAI values over two years. Multispectral and thermal UAV images were acquired throughout the growing season along with destructive LAI measurements. Five ML algorithms, including K-Nearest Neighbors (K-NN), Extra Trees Regressor (ETR), eXtreme Gradient Boosting (XGBoost), Random Forest (RF), and Support Vector Regression (SVR) were tested. Also, feature selection procedures were implemented to obtain useful information among the features used for the ML model. Results showed the K-NN model achieved the highest accuracy (R2 = 0.97, RMSE = 0.46, MAE = 0.197), followed by ETR. The analysis of feature selection revealed that the combination of Normalized Difference Vegetation Index (NDVI) and canopy height (NDVI x Hc) product had the highest importance, followed by Soil-Adjusted Vegetation Index (SAVI) and Green Normalized Difference Vegetation Index (GNDVI). Also, utilizing vegetation indices calculated from multiple spectral bands proved to be more effective than using individual bands alone. Overall, the study demonstrates that UAV data and ML techniques can estimate sorghum LAI precisely to support precision agriculture applications. Moreover, using low-cost UAVs equipped with multispectral sensors presents a cost-effective and reliable method for LAI estimation using ML models.Öğe Calibrating UAV thermal sensors using machine learning methods for improved accuracy in agricultural applications(Elsevier, 2023) Tunca, Emre; Koksal, Eyup Selim; Taner, Sakine CetinAccurate temperature measurements are essential for detecting crop stress, managing irrigation, and monitoring vegetation health. However, various factors can affect thermal sensors that can introduce measurement errors. To address this, machine learning (ML) algorithms were used to calibrate unmanned air vehicle (UAV) thermal sensor measurements. In this study, commercially available two different types of UAV thermal sensors, including Micasense Altum and Flir Duo Pro-R (FDP-R), have been tested and evaluated its performance by comparing the calibrated ground thermal measurements. For this purpose, five different ML algorithms, namely Random Forest, Support Vector Machine, K-NN and XGBoost, were used to calibrate UAV thermal sensors. Results showed that, after thermal calibration with XGBoost, the RMSE decreased by 2.84 degrees C (from 4.23 degrees C to 1.39 degrees C) for Micasense Altum and by 2.51 degrees C (from 3.84 degrees C to 1.33 degrees C) for FDP-R, while R2 increased from 0.89 to 0.96 for Micasense Altum and from 0.87 to 0.94 for FDP-R. In addition, we conducted correlation analyses between the calibrated temperature measurements and various sorghum phenotype parameters, such as leaf area index, crop height, and soil moisture. The results indicate that both sensors have performed well in terms of correlation coefficients. Micasense Altum has shown slightly better performance for crop height and soil moisture (r = -0.78 and r = -0.59, respectively), while FDP-R has performed better for leaf area index (r = -0.70). This study demonstrates the potential of using calibrated UAV thermal sensors for precision agriculture tasks and highlights the importance of validating the calibration with ground measurements.Öğe Evaluating machine learning and deep learning for upscaling instantaneous latent heat flux to daily scale(Springer, 2026) Tunca, EmreThe aim of this study is the evaluation of machine learning (ML) and deep learning (DL) models as an alternative to traditional methods for upscaling hourly latent heat flux (LEi) to daily latent heat flux (LEd). For these purposes, Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Multilayer Perceptron (MLP), and Long Short-Term Memory-Convolutional Neural Network (LSTM-CNN) models were developed using the FLUXNET2015 dataset and compared the performance of these models with the evaporative fraction (EF) approach. Also, the most suitable combinations of time and features were determined to upscale LEi to LEd. Results indicated that the model performance was stable during midday hours, with the highest correlation at 11:00 (R-2=0.75, RMSE = 24.25 W m(-)(2), and MAE = 17.58 W m(- 2)). To determine the best features, Pearson Correlation analysis was conducted, and results revealed that hourly LEi, net radiation (Rn-i), reference evapotranspiration (ETo(i)), incoming solar radiation (SWin), air temperature (Ta-i) and daily net radiation (Rn-d) had the greatest influence on LEd. Among the generated models, MLP performed best (R-2=0.85, RMSE = 15.96 W m(-)(2) and MAE = 11.43 W m(-)(2)) to estimate LEd. However, in terms of computational efficiency, XGBoost outperformed other models with significantly lower training times, whereas DL models required intensive computational effort. This study showed that ML and DL models are robust and more accurate for upscaling LE compared to traditional methods. Specifically, MLP is prominent as a suitable model for accuracy.Öğe Evaluating the impact of different UAV thermal sensors on evapotranspiration estimation(Elsevier, 2024) Tunca, Emre; Koksal, Eyup SelimAccurate evapotranspiration (ET) estimation is vital for precise irrigation management. Remote sensing provides a unique method for obtaining spatial and temporal ET information. With technological advancements, several unmanned aerial vehicle (UAV) thermal sensors have been developed. However, the impact of thermal sensors on ET estimation is unclear. This study evaluated the impact of different UAV thermal sensors, including Micasense Altum and Flir Duo Pro-R (FDP-R), on ET estimation using the Two Source Energy Balance (TSEB) model. A field experiment was conducted during the 2021 sorghum growing period, with irrigation treatments consisting of four different regimes: full irrigation (S1), 70 % of S1 (S2), 40 % of S1 (S3), and rainfed (S4). The results revealed no statistically significant differences between the estimated ET values using Micasense Altum and FDP-R thermal sensors. The TSEB model's performance was entirely satisfactory for full irrigation, with RMSE values of 5.63 mm for Micasense Altum and 7.17 mm for FDP-R, in 10 days. However, the accuracy deteriorated with increasing water stress, reaching 29.02 mm for Micasense Altum and 25.12 mm for FDP-R, in 10 days in rainfed plots. The study results highlight the capability of both Micasense Altum and FDP-R thermal sensors to provide comparable ET estimates, particularly under full irrigation conditions. However, the decline in accuracy with increased water stress underlines a potential limitation of the TSEB model when applied to varying irrigation regimes. These insights emphasize the importance of adjustment of TSEB input parameters such as alpha PT coefficient, resistance terms etc. and sensor technologies, particularly in water-stressed environments, to ensure accurate ET estimation. This study demonstrated the potential of high-resolution UAV thermal sensors for precision irrigation management tasks. Further studies with different thermal sensors are needed to understand this technology's benefits fully. The impact of different climate conditions on ET estimation should also be explored for accurate results.Öğe Evaluating the performance of the TSEB model for sorghum evapotranspiration estimation using time series UAV imagery(Springer, 2023) Tunca, EmreEvapotranspiration (ET) is a vital process involving the transfer of water from the Earth's surface to the atmosphere through soil evaporation and plant transpiration. Accurate estimation of ET is important for a variety of applications, including irrigation management and water resource planning. The two-source energy balance (TSEB) model is a commonly used method for estimating ET using remotely sensed data. This study used the TSEB model and high-resolution unmanned aerial vehicle (UAV) imagery to estimate sorghum ET under four different irrigation regimes over two growing seasons in 2020 and 2021. The study also validated net radiation (Rn) flux through hand-held radiometer measurements and compared the estimated ET with a soil water balance model. The study outcomes revealed that that the TSEB model capably estimated Rn values, aligning well with ground-based Rn measurements for all irrigation treatments (RMSE = 32.9-39.8 W m(-2) and MAE = 28.1-35.2 W m(-2)). However, the TSEB model demonstrated robust performance in estimating ET for fully irrigated conditions (S1), while its performance diminished with increasing water stress (S2, S3, and S4). The R-2, RMSE, and MAE values range from 0.64 to 0.06, 10.94 to 17.04 mm, and 7.09 to 11.43 mm, respectively, across the four irrigation treatments over a 10-day span. These findings not only suggest the potential of UAVs for ET mapping at high-resolution over large areas under various water stress conditions, but also highlight the need for further research on ET estimation under water stress conditions.Öğe Farklı Sulama Seviyeleri Altında Yetiştirilen Dolmalık Biberin Yaprak Alanının Tahmininde Makine Öğrenmesi Yaklaşımlarının Karşılaştırılması(2025) Tunca, Emre; Koksal, Eyüp SelımYaprak Alanı (LA), bitki gelişiminin izlenmesi ve bitki su tüketimi tahmini için oldukça önemli bir parametredir. Bu çalışmada, dolmalık biberin LA değerlerini farklı makine öğrenmesi (ML) modelleri ile tahmin etmek ve farklı sulama düzeylerinin model performansına etkisini değerlendirmek amaçlanmıştır. ML algoritmaları olarak Rastgele Orman (RF), eXtreme Gradient Boosting (XGBoost), En Yakın Komşular (KNN) ve Çok Katmanlı Algılayıcı (MLP) kullanılmıştır. Geniş bir veri seti elde edebilmek için, dört farklı sulama düzeyinde yetiştirilen dolmalık biberin yaprak alanı değerlerinden yararlanılmış ve bu kapsamda 2 yıllık bir tarla denemesi yürütülmüştür. Birinci yıl (2017) elde edilen yaprak alanı değerleri (6757 örnek) model eğitiminde, ikinci yıl (2018) elde edilen değerler ise (6128 örnek) test verisi olarak kullanılmıştır. Elde edilen sonuçlara göre, kullanılan ML algoritmaları arasında en yüksek performansı MLP (R²=0.90, RMSE=0.20 cm² ve MAE=0.15 cm²) sergilerken; en düşük performans XGBoost (R²=0.90, RMSE=0.20 cm² ve MAE=0.15 cm²) tarafından gösterilmiştir. Ayrıca, farklı sulama düzeylerinden elde edilen yaprak alanı değerleri ayrı ayrı değerlendirildiğinde benzer performansların elde edildiği görülmüştür. Sonuçlar, ML yöntemlerinin yaprak alanını geleneksel yöntemlere alternatif olarak hızlı ve doğru şekilde tahmin edebildiğini ortaya koymaktadır.Öğe Integration of UAV images and ensemble learning for root zone soil moisture estimation in sorghum(Springer, 2025) Tunca, Emre; Koksal, Eyup Selim; Taner, Sakine CetinAccurate estimation of root-zone soil moisture (SM) is critical for agricultural water management and sustainable crop production. This study develops and evaluates a methodology to estimate SM in sorghum root zones by high-resolution unmanned aerial vehicle (UAV) multispectral and thermal imagery with machine learning (ML). A two-season field experiment (2020-2021) with four irrigation regimes provided UAV data and concurrent ground-based SM measurements. This study conducted a comparative analysis of four ML algorithms: Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Light Gradient Boosting Machine (LGB), and K-Nearest Neighbors (KNN). The models were evaluated both as standalone predictors and as components of ensemble structures. Among single models, RF achieved the highest test performance (R-2 = 0.84, RMSE = 11.22 mm/90 cm, MAE = 9.32 mm/90 cm). An ensemble combining XGBoost, LGB, and KNN yielded a slight improvement (R-2 = 0.85, RMSE = 11.124 mm/90 cm, MAE = 8.775 mm/90 cm), indicating that ensemble learning can modestly enhance model performance. The proposed workflow offers a practical approach for field-scale SM monitoring, demonstrating potential applications in irrigation scheduling and agricultural water management.Öğe Optimizing Reference Evapotranspiration Estimation in Data-Scarce Regions Using ERA5 Reanalysis and Machine Learning(Mdpi, 2026) Tunca, Emre; Novak, Vaclav; Sarec, Petr; Koksal, Eyup SelimThis study aims to optimize the estimation of reference evapotranspiration (ETo) in data-scarce regions by integrating ERA5-Land reanalysis data with machine learning (ML) models. Daily meteorological data from 33 stations across Turkey's diverse climate zones (1981-2010) were utilized to train and validate three ML models: Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Extreme Learning Machine (ELM). The methodology involved rigorous quality control of ground-based observations, spatial correlation of ERA5-Land grids to station locations, and performance evaluation under various data-limited scenarios. Results indicate that while ERA5-Land provides highly accurate solar radiation (Rs) and temperature (T) data, variables like wind speed (U2) and relative humidity (RH) exhibit systematic biases. Among the used models, XGBoost demonstrated superior performance (R2 = 0.95, RMSE = 0.43 mm day-1, and MAE = 0.30 mm day-1) and computational efficiency. This study provides a robust, regionally calibrated framework that corrects reanalysis biases using ML, offering a reliable alternative for ETo estimation in areas where local measurements are insufficient for sustainable water management.Öğe PlanetScope ve Landsat-8 Uydu Görüntülerinde YOLOv9 Algoritması ile Dairesel Hareketli Sulama Sistemlerinin Tespit Edilmesi(Çanakkale Onsekiz Mart University, 2024) Tunca, EmreDünya nüfusundaki hızlı artış, sürdürülebilir tarımsal üretimin önemini ve suyun etkin kullanımını kritik hale getirmektedir. Suyun verimli kullanılması ise basınçlı sulama sistemlerinin kullanımını gerektirmektedir. Bu sulama sistemleri arasından dairesel hareketli sulama sistemi (DHSS) etkinliği ile öne çıkmaktadır. Söz konusu sistemlerin sayısı ve kullanımı gibi bilgiler su kaynakları yönetimi konusunda oldukça önemlidir. Bu çalışmada DHSS’nin farklı konumsal çözünürlüğe sahip PlanetScope ve Landsat 8 uydu görüntülerinde Sadece Bir Kez Bakarsınız_v9 (YOLOv9) algoritması kullanılarak tespiti amaçlanmıştır. Bu amaçla yakın tarihli PlanetScope ve Landsat 8 uydu görüntüleri YOLOv9 algoritması ile eğitilmiş ve modellerin başarısı kesinlik, duyarlılık ve F1 skoru ile değerlendirilmiştir. Ayrıca modellerin tespit ettiği DHSS sayıları ile manuel sayılan DHSS sayıları ve modellerin eğitim süreleri de karşılaştırılmıştır. Elde edilen sonuçlara göre PlanetScope ve Landsat 8 uydu görüntülerinde kesinlik, duyarlılık ve F1 skoru değerleri sırasıyla 0,970, 0,928, 0,945 ve 0,966, 0,851, 0,897 olarak hesaplanmıştır. PlanetScope ve Landsat 8 görüntülerine dayalı olarak oluşturulan modeller, DHSS sayısı tahminlerinde benzerlik göstermiştir; bu oranlar sırasıyla %96,1 ve %93,2 olarak belirlenmiştir. Ancak, modelin eğitim süreleri arasında önemli bir farklılık gözlemlenmiştir. PlanetScope görüntülerinin model eğitim süresi 1,810 saat olarak kaydedilirken, Landsat 8 görüntülerinin model eğitim süresi 1,414 saat olarak tespit edilmiştir. Araştırmadan elde edilen sonuçlar, YOLOv9 algoritmasının PlanetScope ve Landsat 8 uydu görüntülerinde DHSS’yi benzer başarı oranları ile tespit edilebildiğini ve bu yöntemin su kaynaklarını yönetiminde potansiyel bir araç olarak kullanılabileceğini ortaya koymuştur.Öğe UAV IMAGE-BASED ESTIMATION OF SURFACE ENERGY BALANCE COMPONENTS IN SORGHUM UNDER DIFFERENT IRRIGATION REGIMES(2025) Tunca, EmrePrecise knowledge of plot-scale surface energy partitioning is fundamental for agricultural water management, yet conventional ground or satellite techniques rarely resolve the heterogeneity for the plot scale areas. This study couples multispectral and thermal imagery acquired weekly–bi-weekly by a DJI Matrice-300 RTK/MicaSense-Altum platform with the physically based Two- Source Energy Balance (TSEB) model to quantify net radiation (Rn), soil heat flux (G), sensible heat flux (H) and latent heat flux (LE) in a randomized sorghum experiment comprising four irrigation levels (100, 70, 40 % ETc and rain-fed) in the Bafra Plain, Türkiye (May– October 2021). The model produced consistent seasonal trends: midday Rn peaked at 797 W m⁻², while G/Rn declined from ≈0.20 at emergence to <0.10 under full canopy. Strong irrigation-induced contrasts were detected; fully irrigated plots reached maximum LE of 692 W m⁻² and H of 10 W m⁻², whereas rain-fed plots dropped to LE of 215 W m⁻² and exceeded H of 450 W m⁻² during peak stress. Flux magnitudes and partitioning agreed with published eddy covariance and lysimeter studies, indicating that UAV-driven TSEB reliably bridges the scale gap between point sensors and satellites. The approach offers significant potential for real-time irrigation scheduling and water resource optimization, with applications extending to diverse agricultural systems and climate conditions.Öğe Yüksek Çözünürlüklü Termal Görüntülerin Üretimi ve Değerlendirilmesi: Landsat 8 ve PlanetScope Uydu Verileri Örneği(Ondokuz Mayıs Üniversitesi, 2024) Tunca, EmreBu çalışma, Landsat 8 ve PlanetScope uydu görüntüleri kullanılarak yüksek konumsal çözünürlüğe sahip yapay termal görüntülerin üretilmesi ve bu görüntülerin doğruluğunun değerlendirilmesini amaçlamaktadır. Sultansuyu Tarım İşletmesi arazileri örneği üzerinde yürütülen araştırmada, Normalize Edilmiş Vejetasyon İndeksi (NDVI) ve yüzey sıcaklığı (Ts) haritaları kullanılarak geliştirilen regresyon modeli PlanetScope görüntülerine uygulanmış ve yüksek çözünürlüklü Ts haritası oluşturulmuştur. Araştırmadan elde edilen sonuçlara göre PlanetScope görüntülerinin yüksek çözünürlüklü Ts haritalarında parsel sınırları daha net bir şekilde belirlenirken, Landsat 8 Ts görüntülerinde söz konusu detay ortaya konulamamıştır. İstatistiksel analizler, her iki uydu verisinin de benzer NDVI değerleri ürettiğini ve tutarlı sonuçlar sağladığını doğrulamıştır (R2=0.90). Ancak, PlanetScope verileri, Landsat 8'e göre, genel olarak, daha yüksek NDVI değerleri ve daha geniş bir varyans sergilemiştir. PlanetScope ile üretilen yapay Ts haritaları, homojen bölgelerde Landsat 8 ile benzer sonuçlar üretmesine rağmen, özellikle sulama yapılan düşük örtü yüzdesine sahip alanlarda hatalı Ts tahminleri yapılmaktadır. Gerçekleştirilen bu çalışma, uydu verilerinin tarımsal izleme ve çevresel analizlerde etkin bir şekilde kullanılabilmesi için metodolojik bir temel sağlamaktadır.
