LEE- Geomatik Mühendisliği-Doktora
Bu koleksiyon için kalıcı URI
Gözat
Yazar "Erten, Esra" ile LEE- Geomatik Mühendisliği-Doktora'a göz atma
Sayfa başına sonuç
Sıralama Seçenekleri
-
ÖgeAiding agricultural practices with the exploration of earth observation data via machine learning(Graduate School, 2023-08-04) Çelik, Mehmet Furkan ; Erten, Esra ; 501162604 ; Geomatics EngineeringThe rapid growth of the global population, coupled with the decline in available agricultural fields, the effects of climate change, and soil degradation, pose significant threats to food security. As the population continues to rise, the demand for food and agricultural products increases, putting pressure on optimizing limited resources and their use. The human-made global climate crisis, primarily driven by fossil fuel emissions, worsens the issue, causing extreme weather events and displacing communities. Minimizing environmental damage and maximizing agricultural efficiency is crucial for ensuring a sustainable supply of essential food resources and the well-being of humanity. Obtaining accurate information about agriculture is vital for decision-makers, but traditional in-situ measurements are insufficient to represent the fields and are time-consuming. Remote sensing satellite images provide a solution by offering comprehensive and reliable data, overcoming the limitations of traditional methods, and enabling effective monitoring of agricultural fields on a regional or larger-scale level. Remote sensing satellite imaging technologies, including synthetic aperture radar (SAR) and multi-spectral imaging (MSI) satellites, provide valuable information for Earth Observation (EO) studies. SAR satellites are able to operate in any weather condition, day or night, and penetrate cloud cover, making them highly effective for monitoring Earth's surface. Despite their reliance on clear skies and solar energy, MSI satellites play a crucial role in agricultural monitoring due to their value with a wide range of spectral bands. Both satellite systems have a significant role in observing agricultural fields; SAR satellites are sensitive to detecting morphological changes in crops and MSI satellites have the capability to monitor chemical changes in vegetation. The satellite images offer insights into crop health, growth stages, and potential yield prediction through parameters derived from MSI and SAR images. Utilizing machine learning (ML) algorithms to analyze remote sensing data for agricultural research has opened up a wide range of possibilities for conducting comprehensive studies based on the ability of these algorithms to grasp nonlinear relationships associated with electromagnetic radiation and vegetation. Agricultural planning authorities and researchers can obtain critical insights into many aspects of agriculture and make informed decisions by utilizing the power of these advanced computing approaches. For this purpose, in order to address the critical challenges in monitoring agricultural fields and understanding the interrelation between environmental factors and agricultural activities, three-stage research that implements state-of-art ML and deep learning (DL) methods on remote sensing images has been conducted within the scope of this thesis. These challenges include various aspects of agricultural analysis and can be effectively tackled using the power of ML and DL algorithms that explain the models' behavior in an easy format to understand. In the first study, regression analysis was used to examine the estimation of biophysical parameters using only SAR remote sensing satellite data. Among the regression methods, polynomial chaos expansion (PCE) is one of the reliable and interesting ones due to its tight relationship with uncertainty quantification. One of the advantages of PCE is that global sensitivity analysis (GSA) with Sobol's method can be analytically computed from polynomial coefficients if the input space is statistically independent. However, most of the phenomena include dependent features, either statistically or physically. Therefore, an independent and uncorrelated input space must be created before the regression analysis. In this paper, we performed PCE-based regression analysis for the estimation of biophysical parameters of crops. The study was conducted in the experimental fields of field pea, barley, canola, and oat of the AgriSAR2009 campaign. The input parameters of the regression model were formed by creating polarimetric features derived from RADARSAT-2 imagery. The estimated biophysical parameters were based on the discrete in-situ measurements of leaf area index (LAI) and normalized difference vegetation index (NDVI), scattered semi-randomly in each crop field. We implemented neighborhood component analysis (NCA) to create an independent and uncorrelated input space by eliminating correlations. Once the model was created, we investigated the importance of features that drive the PCE-based regression models applying GSA with Sobol's method. Besides the individual effects of each feature, their interactions were found to be significant. In the second study, time series analysis was conducted to obtain short-term soil moisture in field scale, integrating satellite imaging, climate, and auxiliary data. The recent advancements in different types of satellite imagery coupled with deep learning-based frameworks have paved the way for large-scale SM estimation. This research combined high spatial resolution Sentinel-1 (S1) backscatter data and high temporal resolution Soil Moisture Active Passive (SMAP) SM data to create short-term SM predictions that can accommodate agricultural activities. We created a deep learning model to forecast the daily SM values using time series of climate and radar satellite data, soil type, and topographic data. The model was trained with static and dynamic features that influence SM retrieval. While the topography and soil texture data were taken as stationary, SMAP SM data and S1 backscatter coefficients, including their ratios and climate data were fed to the model as dynamic features. As a target data to train the model, we used \textit{in-situ} measurements acquired from the International Soil Moisture Network (ISMN). We employed a deep learning framework based on Long Short-Term Memory (LSTM) architecture with two hidden layers with 32 unit sizes and a fully connected layer. The model's performance was also evaluated concerning above-ground biomass, land cover classes, soil texture variations, and climate classes. The model prediction ability was lower in areas with high normalized difference vegetation index (NDVI) values. Moreover, the model can predict better in dry climate areas, such as arid and semi-arid climates, where precipitation is relatively low. The daily prediction of SM values based on microwave remote sensing data and geophysical features was successfully achieved using an LSTM framework to assist various studies such as hydrology and agriculture. In the third study, the importance of the input features was investigated during the cotton phenological cycle in order to predict yield using an explainable artificial intelligence. The potential cotton yield can be predicted by integrating the climatic factors, soil parameters, and biophysical parameters observed by high temporal and spatial resolution remote sensing satellites. This study used a multisource dataset to create an explainable and accurate predictive model for cotton yield prediction over the continental US (CONUS). A recently proposed glass-box method called Explainable Boosting Machine (EBM), which provides transparency, reliability, and ease of interpretation, was implemented. Accuracy performance was compared with well-known ML methods for predicting cotton yields. The EBM showed higher accuracy against other glass-box methods and competitive results with black-box models. With the help of the EBM, the importance of individual features and their pairwise interactions was revealed without applying any post-hoc methods. The study findings showed that the precipitation (P), enhanced vegetation index (EVI), and leaf area index (LAI) are the three most important dynamic features. The dynamic features are the driver of the created model with 78% of the overall feature importance, followed by pairwise interactions of the features with 16% contribution. Lastly, static features contribute 6% to the overall feature importance. The study highlights the importance of using multisource data and interactions of the input features and providing an interpretable model to understand the inner dynamics of cotton yield predictions.
-
ÖgeEstimating forest parameters using point cloud data(Graduate School, 2022-08-05) Arslan, Adil Enis ; Erten, Esra ; 501112601 ; Geomatics EngineeringThe spatial distributions and statistical properties of stand attributes must be understood in order to characterize the dynamic forest ecosystem. In this context dendrometry is an invaluable tool in forestry when quantitative characterisation of forests or individual trees are required. Diameter at Breast Height (DBH) and Tree Height (TH) are two significant parameters in dendrometry and heavily correlated with Leaf Area Index. Leaf Area Index (LAI) is described as a dimensionless parameter that has a significant impact in forestry applications and characterising the canopy's structural vegetation in general. With conventional methods, LAI can be calculated with destructive sample collection or with a relatively new non-destructive method called hemispherical photography. Conventional measurements of DBH and TH, although not destructive, are also very time and manpower consuming. With the engagement of modern surveying instruments in forestry, obtaining forest stand parameters for large areas in short time has recently become more prominent and possible with the use of LiDAR technology. Although promising, LiDAR data evaluation techniques for forest stand parameters calculation are still subject to development. This thesis work aims to make a comparative evaluation of existing novel techniques with newly proposed methods for estimating forest stand parameters, namely DBH, TH and LAI. For this purpose Point Cloud Data (PCD) from different sources such as Airborne LiDAR Systems (ALS), Terrestrial Laser Scan (TLS), and Unmanned Aerial Vehicle (UAV) have been evaluated. These data sources have been chosen since they are greatly preferred for forestry operations, and their results can be quantitatively compared against the conventional method results. In-situ data was collected to assess LAI, DBH and TH estimations from PCD through varying sample locations including deciduous, coniferous, mixed forest type. Sampling zone spans from northern parts of Istanbul Urban forest area to a research forest under the supervision of Istanbul University-Cerrahpasa, in Istanbul, Turkey. In-situ measurements were accepted as ground truth, and the results obtained from PCD evaluation were compared against them in terms of their overall error statistics, as well as their performances due to the computational cost and challenges in data acquisitions. The results obtained from the study show that segmentation and removal of wood materials from TLS based PCD by using neural network algorithms and connected component analysis methods, albeit, complex and computer resource demanding, have a promising future on the calculation of effective LAI values of large areas in a very short time span. Similarly, the forestry PCD obtained by TLS has the best performance among other PCD at both DBH and TH estimation
-
ÖgeImproving the performance of remote sensing-based water budget components across mid- and small- scale basins(Graduate School, 2022-07-19) Kayan, Gökhan ; Erten, Esra ; Türker, Umut ; 501152601 ; Geomatics EngineeringIn the last few decades, many global basins have been threatened by rapid urban growth and global warming, resulting in changes in their climate regime. Climate change has increased the incidence of extreme weather events, uncertain water availability, water scarcity, and water pollution. Remote sensing (RS) has emerged as a powerful technique that provides estimations with high spatiotemporal resolution and broad spatial coverage. In recent years, the efficacy of RS products for water budget (WB) analysis has been widely tested and implemented in global and regional basins. Although RS products provide high temporal and spatial resolution images with a near-global coverage, uncertainty is still a significant problem. The main goal of this study is to utilize two different approaches to minimize the uncertainty of the products and to improve RS-based WB estimations in mid- and small- scale basins. The first approach aims to improve the efficacy of water WB estimations from various hydrological data products in the Sakarya basin by; (1) Evaluating the uncertainties of hydrological data products, (2) Merging four precipitation (P) and six evapotranspiration (ET) products using the error variances, and (3) Employing the Constrained Kalman Filter (CKF) method to distribute residual errors (r) among WB components based on their relative uncertainties. The results showed that applying bias correction before the merging process improved estimations of P products with decreasing root mean square error (RMSE), except PERSIANN. VIC and bias-corrected CMORPH products outperformed other ET and bias-corrected P products, respectively, in terms of mean merging weights. The terrestrial water storage change (ΔS) is the primary reason for non-closure errors. This is mainly caused by the two facts. First, the Sakarya basin is a relatively small basin that GRACE can not simply resolve. Second, while P, ET, and Q mostly describe the surface water dynamics, ΔS includes both the surface water and ground water. It is well known that surface water and ground water have completely different dynamic behaviors. The change in surface water is much faster than the change in groundwater. The CKF results were insensitive to variations in uncertainties of runoff (Q). P derived from the CKF was the best output, with the highest correlation coefficient (CC) and the smallest root mean square deviation (RMSD). In the second approach, the annual r in the WB equation arising from the uncertainties of the RS products was minimized by applying fuzzy correction coefficients to each WB component. For analysis, three different fuzzy linear regression (FLR) models with fourteen different sub-models were used in the two basins having different spatial characteristics, namely Sakarya and Cyprus basins. The performance of sub-models is better in the Sakarya basin than that in the Cyprus basin, which has a higher leakage error due to across ocean/land boundary. Moreover, the Cyprus basin is too small for some low-resolution RS-based products to resolve. The Zeng and Hojati sub-models outperformed Tanaka sub-models in the Sakarya basin, whereas Zeng Case-I, Zeng Case-II, and Hojati (degree of fitting index (h) =0.9) sub-models showed the best performance in the Cyprus basin. The best fuzzy sub-models reduced the error up to 68% and 52% in terms of mean absolute error compared to non-fuzzy model in the Sakarya and Cyprus basins, respectively. Further evaluations showed that the best sub-model P well captured the temporal patterns of gauge observations in both basins. Moreover, they have the best consistency with gauge observations in terms of RMSE, Kling-Gupta efficiency (KGE), and percent bias (PBIAS) in the both basins. The results proved that the second approach will provide valuable insights into WB analysis in ungauged basins by incorporating the fuzzy logic approach into hydrological RS products. In general, the FLR and CKF derived P, ET, and Q showed similar seasonal variation with peak and bottom values appeared in nearly the same years. In terms of CC, RMSE, and bias, fuzzy outputs show closest agreement with CKF outputs for Q, with slightly less agreement for P and ET, and much less agreement for ΔS. It can be concluded that the majority of the errors in the second approach are caused by fuzzy ΔS.