Water bodies are among most sensitive ecological environments. In order to ensure good water quality and establish framework for their protection European Parliament was adopted the Water Framework Directive (WFD) (Directive 2000/60/EC). The biological, hydro morphological and physic chemical quality parameters which are relevant for assessment of ecological status of water body are defined in Annex V of WFD. Traditionally, quality of surface water bodies are monitored by in situ measurements resulting in low spatial and temporal resolution of historical data. Remote sensing has great potential for monitoring and identification of water bodies over large scale regions in a more effective and efficient manner. In order to provide reliable monitoring of water quality, surface reflection derived by multispectral sensors need to be integrated with in situ measurements. Relationship between remote sensing and in situ data is usually modeled by using empirical, machine learning or deep learning algorithms. In this study, a 4-year (2013-2016) result of in situ monitoring of surface water bodies in Serbia are used for calibration and validation of algorithm for water quality monitoring based on Landsat 8 satellite image. The Turbidity, Suspending Sediments, Total Phosphorus and Total Nitrogen (physic chemical parameters) in region of Vojvodina, Republic of Serbia are monitored. The Neuron Networks and Supported Vector Machine are used to analyzing correlation between in situ measurements and Landsat 8 atmospherically corrected satellite images. Feature more, capabilities of Landsat 8 are compared with Sentinel 2 images (2-years, 2015-2016). In situ data are provided by Agency for environment protection of Serbia.
Floods are one of the most serious and common natural disasters that cause fatalities and considerable economic losses worldwide every year. In order to reduce and manage risk that floods pose to human health, the environmental and economy flood risk map, as a crucial element of flood risk management, need to be generated. Most often flood risk is estimated based on Digital Elevation Model and projected water levels therefor DEM’s resolution and accuracy highly influence on the reliability of flood risk map especially in lowlands area, where the offset of few decimeters in the elevation data have a significant impact. Airborne light detection and ranging (LiDAR) remote sensing has been a widely used method that provides high-resolution topographical datasets. However LiDAR data are expensive and hard to acquire, usually limited by availability of technology and legal constrains. The main aim of this paper is to present usability of DEM, crated based on UAV RGB images, for flood risk assessment in Vojvodina Porvince, Republic of Serbia therefor flood risk assessment by using the UAV DEM was compared with a flood risk assessment based on LiDAR DEM. Additionally, UAV point cloud was compared with high resolution LiDAR point cloud.
The Water Framework Directive of the European Union aims to protect water bodies from feature degradation. Monitoring is essential for assessment and comprehensive overview of water status. Annex V of WFD define tree type of water quality parameters which need to be monitored (biological and two supported one – hydro morphological and physic chemical) in order to assess ecological status of water bodies. Remote sensing data can be used for monitoring and identification of water bodies over large scale regions in a more effective and efficient manner. However, this technique must to be integrated with traditions in situ sampling method and field surveying in order to provide precise results. Various empirical, semi-analytics and machine learning algorithms exist to derive relationship between multi spectral image surface reflectance and water quality indicators derived from in situ measurement. In this study we evaluate the capabilities of Landsat 8 satellite image for assessment of abundance of phytoplankton’s (biological parameters) and Turbidity, Dissolved oxygen, Total Phosphorus and Total Nitrogen (physic chemical parameters) in region of Vojvodina, Republic of Serbia. The Neuron Networks are used to analyzing correlation between in situ measurements and 7 Landsat 8 atmospherically corrected satellite images acquired in 2013. In situ data are obtained from Agency for environment protection of Serbia. Our results shows that satellite-based monitoring, in combination with in situ data, provide an improved basis for more effective monitoring of large number of water bodies over large geographical area. Relationship between derived and WFD quality parameters is established in order to provide usage of remote sensing data for ecological status classification according to WFD.
The aim of our study was to verify the impact that pansharpening (PS) methods produce on vegetation indices. We used images with both moderate (Landsat 7, Landsat 8) and high (World View2, Ikonos) spatial resolution on which we performed three methods of PS (Brovey transform, Gram-Schmidt and Principal component). The study is based on the differences of vegetation indices (VI) values before and after the pansharpening method is applied. The difference is quantified as an root mean square error. Vegetation indices used in this study were: NDVI, MSAVI2, EVI2, GNDVI, OSAVI and SAVI. Statistical analysis is carried out by calculating coefficients of correlation, root mean square errors and bias calculations for every vegetation index before and after pansharpening procedure is done. The results imply that the BT gave the most diverse results between original VI values and the PS VI values, while the GS and PC methods preserved the values of pixel bands, and that the effect of any PS method is most evident when using Ikonos bands.
The aim of our study was to verify the accuracy and the usability of Moderate resolution imaging spectroradiometer
(MODIS) 13Q1 product for corn yield estimation on a local level for 2014 year. Product 13Q1 consists of Normalized
Difference Vegetation Index (NDVI) and Enhanced Vegetation Index (EVI) 16-day composites with 250 m spatial
resolution. The estimation is based on ground truth data (sowing structures for 8 years) which was provided by local
agricultural organization in Vojvodina, Serbia. The indices were used in linear regression, where the average yield for
corn was the dependent variable, NDVI and EVI were independent variables. Average corn yield was estimated
approximately 15 days before the beginning of the harvest and compared with official results. Depending on the used
linear method, relative errors ranged from 0.6 % to 7.4 %. Overall, coefficients of determination (R<sup>2</sup>) ranged from 0.66
to 0.75 and were significant at 0.05. The smallest difference between official results for corn yield and our estimate when
using NDVI was 0.59 t/ha, when using EVI the smallest difference was 0.07 t/ha. Paper showed that NDVI and EVI
from MODIS follow linear relationship with average corn yield and can be used in estimation of crop yields in Serbia
and also that EVI produces better prediction results than NDVI. The crop yield estimation can be used for similar
cultivated plants in Serbia and for longer period dataset.