Combined use of Processed Images by Wavelet and Neural Fuzzy İnference System to Estimate EC Parameter of the Karun River

Nowadays, monitoring of river quality information is one of the most important issues in water resources engineering because of the direct relationship of water quality with environmental health and quality of life. Today, traditional methods of river monitoring are receiving less attention due to the fact that they are costly and time-consuming for the researcher. Instead, the recent, low-cost methods are favorable to many researchers in this filed. Different methods have always been considered for river monitoring, but the application of spectral indicators and remote sensing technologies to control and monitor the water quality of rivers and reservoirs is very cost-effective and could be a good alternative to traditional methods. Since it is time-saving and less costly, it would be a good indicator for the whole region and a good alternative to manual methods (Bonansea et al., 2015). Although satellite imagery has been widely used in estimating water quality indices (Onderka and Pekárová, 2008), the complexity of hydrological systems and the presence of noise in images can increase the calculation error. Wavelet transform and intelligent models are among the most efficient methods that can significantly increase computation accuracy by filtering and noise reduction. Good research has been done on the use of wavelet transform in image processing (Graps, 1995) and fuzzy inference system to estimate water quality parameters (Solgi et al., 2017). In this study, using wavelet transform, Landsat 8 images were processed, then the processed images were considered as inputs of ANFIS model.