Pemetaan Spesies Mangrove Di Kawasan Lindung Pamurbaya Menggunakan Citra Sentinel-2A Dan Metode Linear Spectral Unmixing

Safitri, Dwi Sugma (2024) Pemetaan Spesies Mangrove Di Kawasan Lindung Pamurbaya Menggunakan Citra Sentinel-2A Dan Metode Linear Spectral Unmixing. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5016201090-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (9MB) | Request a copy

Abstract

Hutan mangrove dapat menjaga ekosistem mulai dari tingkat lokal hingga global, termasuk penstabilan garis pantai, perlindungan dari badai, dan berfungsi sebagai habitat penting, meskipun kemampuan ini berbeda antar spesies. Oleh karena itu, diperlukan peningkatan eksplorasi dan penelitian mangrove hingga ke tingkat spesies. Pemetaan jenis spesies mangrove dapat dilakukan dengan memanfaatkan teknologi penginderaan jauh, salah satunya dengan menggunakan metode Linear Spectral Unmixing (LSU). Metode ini dapat membedakan jenis spesies mangrove berdasarkan karakteristik spektral dan keberadaan suatu objek murni (endmember) di setiap piksel citra. Penelitian ini bertujuan untuk menerapkan metode analisis LSU (Linear Spectral Unmixing) berbasis pustaka spektral pada citra satelit Sentinel-2A sebagai alternatif metode pemetaan spesies mangrove konvensional. Metode LSU mendefinisikan nilai spektral Avicennia Marina, Rhizopora Apiculata, Rhizopora Mucronata, dan Sonneratia Alba menggunakan panjang gelombang tengah dari band B2 (blue), B3 (green), B4 (red), B5 (red edge 1) dan B8 (NIR) pada Sentinel-2A. Analisis LSU diterapkan pada Kawasan Lindung Pantai Timur Surabaya (Pamurbaya) sebagai wilayah penelitian. Hasil penelitian menunjukkan nilai karakteristik spektral yang unik. Nilai reflektansi sedikit lebih tinggi pada panjang gelombang 500—600 nm dan lebih rendah pada 750- 770 nm. Sebagian besar kawasan lindung Pamurbaya didominasi oleh spesies mangrove A. Marina dilanjutkan dengan R. Mucronata, dan R. Apiculata. Namun, terdapat pola distribusi spasial yang berbeda untuk setiap spesies. Berdasarkan nilai Root Mean Square Error (RMSE), pengolahan LSU menghasilkan kesalahan ±3,7% pada setiap piksel. Validasi lapangan (ground-truthing) membantu memvalidasi pola distribusi, hal ini dikaitkan dengan faktor lingkungan, seperti substrat penunjang dan akses air. Kesimpulan dari penelitian ini adalah bahwa analisis LSU efektif untuk pemetaan spesies mangrove skala luas menggunakan data satelit multispektral. Namun demikian, validasi lapangan tetap diperlukan untuk memastikan akurasi pemetaan.
==================================================================================================================================
Mangrove forests safeguard ecosystems from local to global scales, encompassing shoreline stabilization, storm protection, and critical habitat provision. However, these capabilities vary among mangrove species, necessitating in-depth exploration and research at the species level. Mapping mangrove species can be achieved using remote sensing technologies, such as the Linear Spectral Unmixing (LSU) method. LSU differentiates mangrove species based on spectral characteristics and the presence of pure objects (endmembers) within each image pixel. This study aims to implement LSU analysis using a spectral library on Sentinel-2A satellite imagery as an alternative to conventional mangrove species mapping methods. LSU defines the spectral values of Avicennia marina, Rhizophora apiculata, Rhizophora mucronata, and Sonneratia alba using the center wavelengths of bands B2 (blue), B3 (green), B4 (red), B5 (red edge 1), and B8 (NIR) on Sentinel-2A. LSU analysis was applied to the Pamurbaya East Surabaya Coastal Protected Area as the study site. The results revealed unique spectral characteristic values. Reflectance values were slightly higher at 500-600 nm and lower at 750-770 nm. A. marina dominated the Pamurbaya protected area, followed by R. mucronata and R. apiculata. However, each species exhibited distinct spatial distribution patterns. Based on Root Mean Square Error (RMSE) values, LSU processing produced an error of ±3.7% per pixel. Field validation (ground-truthing) supported the distribution patterns, which were linked to environmental factors such as supporting substrates and water access. The study concludes that LSU analysis is effective for large-scale mangrove species mapping using multispectral satellite data. However, field validation remains crucial for ensuring mapping accuracy.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Linear Spectral Unmixing (LSU), Mangrove, Pamurbaya, Pustaka Spektral, Sentinel-2A, Linear Spectral Unmixing (LSU), Mangrove, Pamurbaya, Sentinel-2A, Spectral Library
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) > G70.217 Geospatial data G Geography. Anthropology. Recreation > G Geography (General) > G70.5.I4 Remote sensing
Divisions:	Faculty of Civil Engineering and Planning > Geomatics Engineering > 29202-(S1) Undergraduate Thesis
Depositing User:	Dwi Sugma Safitri
Date Deposited:	22 Jul 2024 02:09
Last Modified:	22 Jul 2024 02:09
URI:	http://repository.its.ac.id/id/eprint/108576

Actions (login required)

View Item