Object based classification of benthic habitat using Sentinel 2 imagery by applying with support vector machine and random forest algorithms in shallow waters of Kepulauan Seribu, Indonesia

Abstract

Abstract. Hartoni, Siregar VP, Wouthuyzen S, Agus SB. 2021. Object based classification of benthic habitat using Sentinel 2 imagery by applying with support vector machine and random forest algorithms in shallow waters of Kepulauan Seribu. Biodiversitas 23: 514-520. Benthic habitats have very high complexity and are home to many types of aquatic organisms. Benthic habitats have various functions, including habitat for flora and fauna, sediment traps, nursery areas, and foraging areas for aquatic fauna that are susceptible to damage due to human activities or natural factors. Therefore, more accurate spatial information is needed. The purpose of this study was to examine the ability of object-based classification techniques for mapping shallow waters benthic habitats using Sentinel 2A imagery. The two classification algorithms used are support vector machine (SVM) and random forest (RF). The input image layer (IIL) used for classification is the natural color band (Band 432). The results showed that the SVM and RF classification algorithms could classify eight classes of benthic habitats. The overall accuracy (OA) of the SVM algorithm is 65%, while the RF accuracy is 67%, with kappa values of 0.59 and 0.60, respectively. The significant test applied to Sentinel 2 images with SVM and RF algorithms for benthic habitats has a Z test value of-0.41. These results indicate that the classification results between the SVM and RF algorithms are not significantly different.

Identificação de Garimpos de Ouro Embarcados por meio de Algoritmos de Classificação em Imagens Sentinel-2

“…Ao comparar a abordagem pixel a pixel versus orientada a objeto, é perceptível a diferença nos resultados para o SVM (Tabela 5) quando aprimorou todas as métricas de avaliação. Keshtkar et al (2017) e Noi e Kappas (2017) obtiveram resultados semelhantes em que o classificador SVM apresentou as melhores métricas para identificação de diferentes tipos de águas, melhorando a performance, inclusive, quando utilizado método orientado a objeto (Hartoni et al, 2022). Os resultados para KNN e RF foram variados, pois apesar de terem sido mantidos os scores F1 médios, verificou-se que o precision e recall foram mais aleatórios, melhorando para alguns alvos e piorando para Os classificadores pixel a pixel cometeram muitos erros de comissão entre garimpos embarcados e solo, gerando classificações com erros sistêmicos (Figuras 10).…”

Monitoring the Spatial and Interannual Dynamic of Zostera noltei

“…Sentinel-2 data have been widely used to map different habitats and several studies have demonstrated their reliability in mapping and monitoring changes in marine biocenoses such as coral (Hedley et al 2012 ), mangrove (Pham et al 2019b , 2020 ) and seagrass beds Topouzelis et al 2016 ; Traganos and Reinartz 2018b ; Ha et al 2020 , 2021 ; Wicaksono et al 2021 ; Nur et al 2021 ; Ivajnšič et al 2022 ; Hartoni et al 2022 ). Many authors have used vegetation indices as the input for RF classification to map plant communities in terrestrial wetlands (Fletcher 2016 ) and coastal wetlands (Zoffoli et al 2020 ; Martínez Prentice et al 2021 ; Benmokhtar et al 2021 ), and compared to many machine learning algorithms and support vector machine techniques, the RF algorithm has produced promising results in terms of classifying seagrass (Zhang et al 2013 ; Traganos and Reinartz 2018b ; Ha et al 2020 ).…”

Integrated Supervised Classification of LULC in Identifying Musang King Durian Illegal Farming Location

“…Therefore, the extraction of LULC changes and NDVI for vegetation discrimination were obtained by OBIA to specify the existence of illegal MKD farming activities. This is because OBIA entails segmenting images into homogeneous areas and identifying object characteristics using spatial and contextual features (Hartoni et al 2022).…”