Dwikarsa, Yahya and Basith, Abdul (2021) Benthic habitats classification using multi scale parameters of GEOBIA on orthophoto images of Karimunjawa waters. Communications in Science and Technology, 6 (1). 55 – 59. ISSN 25029258
Full text not available from this repository. (Request a copy)Abstract
Benthic as a source of food for marine life and an indicator of the quality of the marine environment habitats play an important role in coastal management. Hence, spatial information on benthic habitats is required for coastal management. The nature of benthic habitats requires high spatial resolution image for information extraction. Geographic object-based image analysis (GEOBIA) is an appropriate tool for working with high spatial resolution image. The irregular shape of the benthic habitats and their various dimensions, however, require the application of multi scale parameters for optimal segmentation of benthic habitats. The selection of scale parameter is an important part of image segmentation stage and determine the size of objects and in turn affects the results of classification accuracy. In addition, the selection of image classification algorithm applied to shallow water benthic habitat objects determine the success of the classification. Various combinations of scale parameter and classification algorithms are performed to get the optimal results indicated by classification accuracies. This study used orthophoto images processed from Unmanned Aerial Vehicle (UAV) mission intended to capture benthic habitats in the busiest coastal of Karimunjawa waters, around two Karimunjawa ports. Three classification algorithms, namely Support Vector Machine (SVM), Bayesian statistics, and K-Nearest Neighbors (KNN) are applied with combination of selected scale parameters, namely 100, 200, and 300 resulted from segmentation stage. The classified images are tested their accuracies based on field samples and Training Test Area (TTA) masks. The result showed that combination of SVM algorithm and a scale parameter of 300 produced the best accuracies in terms of overall, producer and user accuracies followed by Bayesian statistic and KNN algorithms. © 2021 Komunitas Ilmuwan dan Profesional Muslim Indonesia. All right reserved.
Item Type: | Article |
---|---|
Additional Information: | Cited by: 0; All Open Access, Gold Open Access |
Uncontrolled Keywords: | Benthic Habitats; GEOBIA; Multi scale parameters; Karimunjawa waters |
Subjects: | G Geography. Anthropology. Recreation > GE Environmental Sciences |
Divisions: | Faculty of Engineering > Geodetic Engineering Department |
Depositing User: | Sri JUNANDI |
Date Deposited: | 06 Oct 2024 23:14 |
Last Modified: | 06 Oct 2024 23:14 |
URI: | https://ir.lib.ugm.ac.id/id/eprint/8719 |