Analisis Sebaran Spasial Genangan Banjir dengan Citra Sentinel-1 Menggunakan Metode Manual dan Automatic Thresholding Pada Area Persawahan

Azizah, Luthfia (2024) Analisis Sebaran Spasial Genangan Banjir dengan Citra Sentinel-1 Menggunakan Metode Manual dan Automatic Thresholding Pada Area Persawahan. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6016231022-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (4MB) | Request a copy

Abstract

Banjir merupakan bencana alam yang terjadi hampir di seluruh wilayah Indonesia baik wilayah urban maupun rural. Pemanfaatan data SAR dengan menggunakan Sentinel-1 dalam pemetaan banjir telah banyak dilakukan, namun pemetaan banjir terus dikembangkan. Metode pemetaan banjir dengan citra SAR Sentinel-1 GRD yang paling umum digunakan yaitu metode thresholding secara global dan memisahkan kelas genangan dan non-genangan secara langsung. Namun beberapa penelitian menunjukkan bahwa penggunaan global thresholding kurang baik untuk digunakan pada citra dengan ukuran yang besar seperti citra Sentinel-1. Sehingga dalam penelitian ini akan dilakukan perbandingan antara pemetaan sebaran genangan banjir di wilayah persawahan Kabupaten Lamongan dengan menggunakan metode ambang batas manual berupa change detection dan otsu thresholding yang diperoleh dan diterapkan pada keseluruhan citra, dengan metode ambang batas otomatis Water-S1yang berupa pemodelan parameter dan penentuan probabilitas posterior dari setiap piksel hingga menghasilkan peta probabilitas permukaan air. Hasilnya menunjukkan bahwa kedua metode dapat mengidentifikasi genangan banjir pada area sawah yang hasil presentasenya terhadap total luas sawah secara berturut-turut yaitu 51% dan 18% dan untuk Change Detection dan Water-S1. Pola hasil visualisasi banjir yang diberikan menunjukkan bahwa hasil banjir dari Water-S1 lebih mengelompok jika dibandingkan dengan hasil banjir dari change detection yang lebih menyebar hampir ke seluruh wilayah Kabupaten Lamongan. Hasil uji akurasi berupa perbandingan dengan Peta Bahaya Banjir BNPB menunjukkan bahwa metode ke dua yaitu Water-S1 lebih akurat dengan nilai overall accuracy sebesar 90% dan change detection sebesar 79%.
============================================================
Floods are natural disasters that occur almost throughout Indonesia, both in urban and rural areas. The use of SAR data with Sentinel-1 for flood mapping has been widely conducted, but flood mapping continues to be developed. The most commonly used method for flood mapping with Sentinel-1 GRD SAR images is the global thresholding method, which directly separates the flooded and non-flooded classes. However, several studies have shown that the use of global thresholding is less suitable for large images such as Sentinel-1 images. Therefore, this study will compare the flood inundation mapping in rice field areas at Lamongan Regency using the manual thresholding methods of change detection and Otsu thresholding, which are obtained and applied to the entire image, with the automatic thresholding method Water-S1, which involves parameter modeling and determining the posterior probability of each pixel to produce a water surface probability map. The results show that both methods can identify flood inundation in rice field areas, with the percentage of flooded area relative to the total rice field area being 51% and 18% for Change Detection and Water-S1, respectively. The flood visualization patterns provided indicate that the flood results from Water-S1 are more clustered compared to the flood results from change detection, which are more dispersed almost throughout the Lamongan Regency. Accuracy tests, comparing with the BNPB Flood Hazard Map, show that the second method, Water-S1, is more accurate with an overall accuracy of 90%, compared to 79% for change detection.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Banjir, SAR Sentinel-1 GRD, Thresholding, Water-S1 ================================================== Flood, SAR Sentinel-1 GRD, Thresholding, Water-S1
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) > G70.5.I4 Remote sensing G Geography. Anthropology. Recreation > GB Physical geography > GB1399.9 Floods
Divisions:	Faculty of Civil Engineering and Planning > Geomatics Engineering > 29101-(S2) Master Thesis
Depositing User:	Luthfia Azizah
Date Deposited:	31 Jul 2024 04:50
Last Modified:	31 Jul 2024 04:50
URI:	http://repository.its.ac.id/id/eprint/110083

Actions (login required)

View Item