Fathan Musyaffa Abdul Jabbar, - and Lili Somantri, - and Silmi Afina Aliyan, - (2025) INTEGRASI CITRA MULTISPEKTRAL DAN SYNTHETIC APERTURE RADAR (SAR) UNTUK PENGEMBANGAN METODE KLASIFIKASI TUTUPAN LAHAN PADA WILAYAH PERI URBAN KABUPATEN BANDUNG BARAT. S1 thesis, Universitas Pendidikan Indonesia.
Abstract
Perubahan tutupan lahan di wilayah peri urban Kabupaten Bandung Barat berkembang semakin kompleks akibat urbanisasi yang pesat. Teknologi penginderaan jauh citra multispektral memiliki keterbatasan dalam kondisi atmosfer cuaca tropis dan kesamaan spektral antar kelas lahan sehingga tidak mampu mengklasifikasikan tutupan lahan secara akurat. Penelitian ini bertujuan untuk mengembangkan metode klasifikasi tutupan lahan dengan mengintegrasikan citra multispektral dan Synthetic Aperture Radar (SAR) menggunakan machine learning algoritma Random Forest yang ditingkatkan melalui deep learning berupa parameter tuning, validasi dilakukan pada model dan validasi lapangan sebanyak 36 dan 90 titik Ground Control Points (GCP). Hasil penelitian ini menunjukan bahwa pemetaan tutupan lahan berbasis integrasi citra multispektral dan SAR terbukti mampu memperbaiki kesalahan akibat efek atmosfer dan kesamaan nilai spektral. Model integrasi citra multispektral dan SAR terbukti unggul dibandingkan model konvensional berdasarkan akurasi keseluruhan yang meningkat dari 88,4% (Kappa 0,86) menjadi 94% (Kappa 0,93). Validasi lapangan tahap awal dengan 36 titik GCP menunjukkan lonjakan akurasi dari 61,1% (Kappa 0,53) menjadi 83,3% (Kappa 0,80). Validasi lanjutan menggunakan 90 titik GCP semakin menegaskan keunggulan model terintegrasi, dengan akurasi mencapai 93,3% (Kappa 0,92), jauh di atas model konvensional yang hanya mencapai 68,8% (Kappa 0,62). Sehingga pengkategorian Wilayah Peri Urban KBB menggunakan citra terintegrasi dapat dianggap lebih benar berdasarkan keunggulan dari tiga tahap validasi yang dilakukan yaitu luas kenampakan kedesaan mencapai 11.110,42 Ha dan kenampakan kekotaan hanya 12.066,31 Ha. Penelitian ini membuktikan bahwa integrasi SAR dan multispektral lebih efektif dalam mengurangi bias klasifikasi, meningkatkan akurasi pemetaan, dan dapat diterapkan dalam perencanaan tata ruang serta manajemen lingkungan secara lebih presisi. Land cover change in the peri urban area of West Bandung Regency is growing more complex due to rapid urbanization. Multispectral image remote sensing technology has limitations in tropical weather atmospheric conditions and spectral similarity between land classes, making it unable to classify land cover accurately. This research aims to develop a land cover classification method by integrating multispectral imagery and Synthetic Aperture Radar (SAR) using Random Forest algorithm machine learning enhanced through deep learning in the form of parameter tuning, validation is done on the model and field validation of 36 and 90 Ground Control Points (GCP). The results of this study found that land cover mapping based on the integration of multispectral and SAR imagery proved to be able to correct errors due to atmospheric effects and similarity of spectral values. The integration model of multispectral and SAR imagery proved superior to the conventional model based on the overall accuracy which increased from 88.4% (Kappa 0.86) to 94% (Kappa 0.93). first field validation with 36 GCP points showed a jump in accuracy from 61.1% (Kappa 0.53) to 83.3% (Kappa 0.80). Further validation using 90 GCP points further confirmed the superiority of the integrated model, with accuracy reaching 93.3% (Kappa 0.92), far above the conventional model which only reached 68.8% (Kappa 0.62). So that the categorization of the KBB Peri Urban Area using integrated imagery can be considered more correct based on the advantages of the three stages of validation carried out, namely the area of rural appearance reaching 11,110.42 Ha and urban appearance only 12,066.31 Ha. This research proves that the integration of SAR and multispectral is more effective in reducing classification bias, increasing mapping accuracy, and can be applied in spatial planning and environmental management more precisely.
![[img]](http://repository.upi.edu/style/images/fileicons/text.png) |
Text
S_SIG_2102374_Title.pdf Download (1MB) |
|
Text
S_SIG_2102374_Chapter1.pdf Download (194kB) |
|
Text
S_SIG_2102374_Chapter2.pdf Restricted to Staf Perpustakaan Download (822kB) | Request a copy |
|
Text
S_SIG_2102374_Chapter3.pdf Download (868kB) |
|
Text
S_SIG_2102374_Chapter4.pdf Restricted to Staf Perpustakaan Download (6MB) | Request a copy |
|
Text
S_SIG_2102374_Chapter5.pdf Download (119kB) |
|
Text
S_SIG_2102374_Appendix.pdf Restricted to Staf Perpustakaan Download (1MB) | Request a copy |
| Item Type: | Thesis (S1) |
|---|---|
| Additional Information: | https://scholar.google.com/citations?user=5GYjV4QAAAAJ&hl=en ID SINTA Dosen Pembimbing: Lili Somantri 5995390 Silmi Afina Aliyan 6749474 |
| Uncontrolled Keywords: | Penginderaan jauh, klasifikasi tutupan lahan, multispektral, SAR, Random Forest, wilayah peri urban, machine learning, deep learning. Remote sensing, land cover classification, multispectral, SAR, Random Forest, peri urban area, machine learning, deep learning. |
| Subjects: | G Geography. Anthropology. Recreation > G Geography (General) G Geography. Anthropology. Recreation > GA Mathematical geography. Cartography G Geography. Anthropology. Recreation > GF Human ecology. Anthropogeography |
| Divisions: | Fakultas Pendidikan Ilmu Pengetahuan Sosial > Sains Informasi Geografi |
| Depositing User: | Fathan Musyaffa Abdul Jabbar |
| Date Deposited: | 22 Apr 2025 08:20 |
| Last Modified: | 22 Apr 2025 08:20 |
| URI: | http://repository.upi.edu/id/eprint/132336 |
Actions (login required)
 |
View Item |