KLASIFIKASI HABITAT PERAIRAN DANGKAL BERBASIS OBJEK DENGAN ALGORITMA SVM DAN KNN PADA CITRA WORLDVIEW 2 DAN CITRA SPOT 6 DI GUSUNG KARANG LEBAR

Esty Kurniawati (Universitas Maritim Raja Ali Haji)
Vincentius P. Siregar (Departemen Ilmu dan Teknologi Kelautan, IPB University, Bogor, Indonesia)
I Wayan Nurjaya (Departemen Ilmu dan Teknologi Kelautan, IPB University, Bogor, Indonesia)

JURNAL ENGGANO
Vol. 7 No. 1 (2022)

This study uses the Object Based Image Analysis (OBIA) approach for mapping shallow-water benthic habitats in Kepulau Seribu. This study aims to compare the capabilities of the classification techniques of Support Vector Machin algorithm and k-Nearest Neighbor on Worldview and SPOT Satellite Images. The selection of SVM and KNN algorithms in the classification process has an influence on the final results of image processing. The results show that the overall accuracy in the Worldview algorithm SVM image is 76% and KNN is 80%, while for SPOT imagery they are 73% and 77% respectively. The results of this study indicate that the SVM and KNN algorithms are able to map the shallow water benthic habitat well in Wordview 2 and SPOT 6 imagery.