Strategic decisions about the construction of engineering structures along the river which are essential for the management of sediment entering the reservoir will be facilitated by understanding the behavior and characteristics of the sedimentation of rivers leading to large dam reservoirs. Multi-temporal ...
Read More
Strategic decisions about the construction of engineering structures along the river which are essential for the management of sediment entering the reservoir will be facilitated by understanding the behavior and characteristics of the sedimentation of rivers leading to large dam reservoirs. Multi-temporal and spectral remote sensing technology has been applicable for detecting of the rivers morphological changes. However, the specific nature of the narrow and shallow rivers is responsible for increasing the complexity of morphology with available data. This study was undertaken in order to detect narrow and shallow rivers by assessing the ability of six famous water indices, including: Normalised Difference Water Index, Modified Normalised Difference Water Index, Automated Water Extraction Index no shadow, Automated Water Extraction Index shadow, Enhanced Water Index and Water Index 2015 which were derived from two Landsat ETM+ and OLI sensors. The optimal threshold for each of these indices was determined using ROC curves and validation process was carried out using Google Earth images captured in August 2013. The accuracy of results was evaluated by using different statistics including combined error, producer’s accuracy, user accuracy and omission and commission errors. Consequently, the results of this study have shown that the ETM+ sensor was generally more accurate than OLI sensor. All in all the Modified Normalised Difference Water Index and the Automated Water Extraction Index shadow was the most accurate indices. Also Automated Water Extraction Index no shadow index had the lowest accuracy for the river’s detecting process
