| Summary: | Sedimentation at the river mouth always occurs in Kuala Perlis disrupting the ferries and boat users, especially during the low tide. This study is aimed to compare the best method for predicting river sediment deposition between K-Means unsupervised image classification machine learning and Modified Normalized Difference Water Index (MNDWI) in analyzing the areas most influenced by depositing river sediments from the clustered images. Quantification of Landsat 8 OLI satellite images were applied using ENVI software in the study area to detect sedimentation utilizing image data band correlation deposited river sediment through the unsupervised classifier algorithm and selection of spectral bands for MNDWI. The determination of determinant bands from analysis of correlation coefficient resulted in NIR bands for their lowest R2 coefficient that ranged from 0.5 to 0.7. Then the selected K-Means classification method has been taken for further clustered image analysis compared to the MNDWI method. From the analysis through stage statistic, visual observation and previous studies review, the river sediment deposition at the river mouth was significantly increased from the year 2019 to the year 2021. These results were supported by the 14 percent of the increase for riverbed regions subjected to sediment deposition with R2 0.916. The location of Kuala Perlis itself exacerbated the problem of dumping sediment returned to the river mouth in a brief period, which is also reliant on the wave flow. This study exhibits importance results for the future development of Kuala Perlis and local communities nearby. © 2023 IEEE.
|
|---|