IDENTIFIKASI LAHAN SAWAH DENGAN SUPPORT VECTOR MACHINE BERBASISKAN PARAMETER SPEKTRAL DAN PARAMETER FISIK (STUDI KASUS: KABUPATEN BANDUNG)

• Main Author: yendi maulana, Fadillah
• Format: Theses
• Language: Indonesia
• Subjects: Teknik (Rekayasa, enjinering dan kegiatan berkaitan)
• Online Access: https://digilib.itb.ac.id/gdl/view/87079
• Institution: Institut Teknologi Bandung

This research aims to identify rice fields in Bandung Regency using the support vector machine method based on spectral parameters and physical parameters. Mapping rice fields usually requires several satellite images over different periods to obtain typical phenotypic information on the rice fields. To minimize time and costs, this research optimizes the support vector machine classification using satellite images at one time by adding physical characteristic aspects of the rice fields. The data used include NDVI, NDWI, LST, SMI, BSI, SAVI, and physical data such as slope, elevation, soil pH, total nitrogen, and clay content. The classification results show that the use of a combination of spectral and physical parameters has the potential to improve classification accuracy compared to using only spectral data, but the improvement is not significant and further understanding is needed regarding the selection of physical data used, considering that not all physical data used in this research improve classification results and some other physical data provide lower results compared to using only spectral data. In the best scenario, the overall accuracy reaches 98% with a Kappa of 0.87. This method has the potential to improve the effectiveness and efficiency of rice field mapping, which is important for agricultural management and food security. Further research is needed on the selection of physical data and better sampling strategies so that the application of this methodology can support efforts to monitor the distribution of rice fields sustainably in Bandung Regency.