Identifikasi Varietas Jagung Dari Data Citra Satelit Menggunakan Metode Spectral Unmixing (Studi Kasus: Kabupaten Ngawi)

Rivani, Amalia Putri (2023) Identifikasi Varietas Jagung Dari Data Citra Satelit Menggunakan Metode Spectral Unmixing (Studi Kasus: Kabupaten Ngawi). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 03311940000026-Undergradute_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 September 2025. Download (11MB)

Abstract

Kabupaten Ngawi di Jawa Timur merupakan wilayah dengan sekitar 40% lahannya digunakan untuk pertanian, termasuk budidaya jagung. Jagung adalah sumber karbohidrat yang penting bagi industri pangan dan permintaannya terus meningkat seiring pertumbuhan penduduk. Untuk meningkatkan potensi hasil pertanian jagung, penting untuk memantau lahan pertanian dengan menggunakan teknologi penginderaan jauh. Namun, untuk melakukan identifikasi varietas jagung, diperlukan pemantauan lahan jagung yang spesifik. Oleh karena itu, penelitian ini bertujuan untuk mengidentifikasi lahan jagung dan sebaran varietas tanaman jagung dikabupaten Ngawi menggunakan data citra satelit Landsat-9 dengan metode Random Forest dan Spectral Unmixing. Pada penelitian ini, dilakukan pengukuran lapangan menggunakan Spektrometer Optics USB4000 untuk mengidentifikasi endmember dari setiap varietas tanaman jagung. Selanjutnya, fase pertumbuhan jagung dapat diketahui dengan menggunakan citra Sentinel-2 menggunakan algoritma NDVI dan NDWI. Hasil yang diperoleh dari penelitian ini adalah luasan lahan jagung dan citra fraksi endmember varietas jagung serta luasan varietas jagung dominan. Dimana, klasifikasi lahan jagung dan non-jagung didapat overall accuracy sebesar 97,6% , kappa 84% yang termasuk kategori akurasi sangat baik dan didapatkan luas lahan jagung seluas 64,756 km2. Hasil sebaran diterapkan nilai threshold fase generatif akhir seluas 64.364 km2. Sebaran varietas dominan yang terdeteksi dari hasil pengolahan adalah NK Sumo, NK Perkasa, NK Wirosableng, dan Varietas lain (selain Bisi-18, NK Sumo, NK Perkasa, dan NK Wirosableng) dengan luasan sebesar 0,00088 km2; 7,800 km2 ;22,662 km2 ; 33,900 km2. Namun, setelah dilakukan validasi menggunakan beberapa titik sampel yang tersebar didapatkan hasil yang berbeda dengan hasil pengolahan data yang dilakukan. Perbedaan hasil yang diperoleh tersebut dapat didasarkan oleh beberapa faktor diantaranya adalah ketersediaan pustaka spektral dan resolusi spasial citra yang digunakan.Oleh karena itu, dilakukan uji akurasi menggunakan RMSe dari nilai Residual Erorr piksel pada Citra Landsat-9 menghasilkan nilai pada rentang 0,019–0,042 dengan nilai rata-rata yaitu sebesar 0,031. Hasil menunjukan pemisahan nilai fraksi campuran menggunakan metode Spectral Unmixing memiliki error sebesar ± 3,1%, hasil kesalahan error tersebut tergolong akurasi yang sangat baik yang dapat memberikan nilai persentase fraksi endmember dari setiap varietas lahan jagung di Kabupaten Ngawi.
========================================================================================================================================
Ngawi Regency in East Java is an area where approximately 40% of its agricultural land is used for farming, including corn cultivation. Corn is an important source of carbohydrates for the food industry, and its demand continues to increase with population growth. To enhance the potential agricultural yield of corn, it is important to monitor farmland using remote sensing technology. However, specific monitoring of corn fields is required to identify corn varieties. Therefore, this research aims to identify corn fields and the distribution of corn varieties in Ngawi Regency using Landsat-9 satellite imagery with the Random Forest and Spectral Unmixing methods. Field measurements are conducted using the USB4000 Optical Spectrometer to identify endmembers of each corn variety. Furthermore, the corn growth stages can be determined using Sentinel-2 imagery using the NDVI and NDWI algorithms. The results obtained from this research include the extent of corn distribution, the fractional endmember images of corn varieties, and the dominant corn variety areas. The classification of corn and non-corn land achieved an overall accuracy of 97.6%, with a kappa of 84%, indicating excellent accuracy. The total area of corn land was found to be 64.756 km2. The distribution results applied a threshold value for the final generative phase, covering an area of 64.364 Km2. The dominant detected corn varieties from the processing results are NK Sumo, NK Perkasa, NK Wirosableng, and other varieties (excluding Bisi-18, NK Sumo, NK Perkasa, and NK Wirosableng), with areas of 0.00088 km2, 7.800 km2, 22.662 km2, and 33.900 km2. However, after validation using several scattered sample points, different results were obtained compared to the processed data. The differences in the obtained results can be attributed to various factors, including the availability of spectral libraries and the spatial resolution of the utilized imagery. Hence, accuracy testing was conducted using the RMSe form residual error pixel of Landsat 9 imagery, ranging from 0.019 to 0.042, with an average RMSe value of 0.031. The results indicate that the separation of mixed fraction values using the Linear Spectral Unmixing method has an error of approximately ± 3.1%. This level of error is considered highly accurate and provides the percentage values of endmember fractions for each corn variety in Ngawi Regency.

Item Type:	Thesis (Other)
Additional Information:	RSG 621.367 8 Riv i-1 2023
Uncontrolled Keywords:	Corn, Ngawi Regency, Spectral Indices, Spectral Unmixing Jagung, Kabupaten Ngawi, Spektral Indeks, Spectral Unmixing
Subjects:	G Geography. Anthropology. Recreation > G Geography (General) > G70.212 ArcGIS. Geographic information systems. G Geography. Anthropology. Recreation > G Geography (General) > G70.217 Geospatial data G Geography. Anthropology. Recreation > G Geography (General) > G70.5.I4 Remote sensing G Geography. Anthropology. Recreation > GA Mathematical geography. Cartography > GA102.4.R44 Cartography--Remote sensing
Divisions:	Faculty of Civil, Planning, and Geo Engineering (CIVPLAN) > Geomatics Engineering > 29202-(S1) Undergraduate Thesis
Depositing User:	Amalia Putri Rivani
Date Deposited:	02 Aug 2023 02:42
Last Modified:	19 Dec 2023 04:03
URI:	http://repository.its.ac.id/id/eprint/101510

Actions (login required)

View Item