BENTHIC HABITAT CLASSIFICATION OF SHALLOW WATER USING WORLDVIEW-2 IMAGERY WITH IN-SITU AND DRONE DATA
| Main Authors: | Sugara, Ayub, Siregar, Vincentius Paulus, Agus, Syamsul Bahri |
|---|---|
| Format: | Article info application/pdf eJournal |
| Bahasa: | eng |
| Terbitan: |
Department of Marine Science and Technology, Faculty of Fisheries and Marine Science, IPB University
, 2020
|
| Subjects: | |
| Online Access: |
http://journal.ipb.ac.id/index.php/jurnalikt/article/view/26448 http://journal.ipb.ac.id/index.php/jurnalikt/article/view/26448/19764 |
Daftar Isi:
- Aplikasi citra Worldview-2 dengan data groud truth habitat masih memiliki kekurangan yaitu membutuhkan waktu yang lama, akses yang terbatas, biaya yang tinggi serta faktor resiko. Teknik survei menggunakan drone dapat mengurangi keterbatasan tersebut. Penelitian ini bertujuan untuk melakukan klasifikasi dan uji akurasi hasil klasifikasi habitat perairan dangkal di Pulau Lancang dan Pulau Sebaru Besar dari citra Worldview-2 dengan inputan data ground truth habitat (GTH) dan virtual ground truth (VGT) serta mengeksplorasi resolusi spasial citra drone pada ketinggian yang berbeda. Hasil overall accuraccy diperoleh untuk 7 kelas habitat di Pulau Lancang dengan data GTH dan VGT masing-masing sebesar 65,5% dan 60,6%. Sedangkan di Pulau Sebaru Besar masing-masing sebesar 67,5% dan 64,6%. Perbandingan akurasi hasil klasifikasi didapat selisih 4,9% di Pulau Lancang dan 2,9% di Pulau Sebaru Besar. Hasil uji signifikansi metode GTH dan VGT di Pulau Lancang berbeda nyata dengan nilai Z sebesar 2,0851, sedangkan di Pulau Sebaru Besar tidak berbeda nyata dengan nilai Z sebesar 0,5255, sehingga pemetaan habitat bentik dengan metode VGT dapat digunakan sebagai alternatif pengamatan di lapangan secara in-situ, namun ini masih memerlukan penelitian lanjutan.
- The Worldview-2 imagery application with groud truth habitat data still has shortcomings that require a long time, limited access, high costs and risk factors. Surveying techniques using drones can reduce these limitations. This study aims to classify and test the accuracy of shallow water habitat classification results in Lancang Island and Sebaru Besar Island from Worldview-2 imagery with ground truth habitat (GTH) and virtual ground truth (VGT) input data and explore the spatial resolution of drone images at altitude different. Overall accuracy results were obtained for 7 habitat classes on Lancang Island with GTH and VGT data of 65.5% and 60.6%, respectively. Whereas in Sebaru Besar Island they were 67.5% and 64.6%, respectively. Comparison of the accuracy of the classification results obtained 4.9% selisi on Lancang Island and 2.9% on Sebaru Besar Island. Significance test results of the GTH and VGT methods on Lancang Island were significantly different with a Z value of 2.0851, while on Sebaru Besar Island it was not significantly different from the Z value of 0.5255, so that benthic habitat mapping with the VGT method could be used as an alternative in-situ field observation, however this still requires further research.