การประยุกต์ใช้ข้อมูลการรับรู้จากระยะไกลในการจำแนกแนวปะการัง : กรณีศึกษา เกาะกระดาน จังหวัดตรัง และเกาะไหง จังหวัดกระบี่

Abstract

The objective of this study was to assess the effectiveness of remote sensing techniques for coral reef classifications at Kradan Island, Trang province and Ngai Island, Krabi province. The Landsat 8 OLI image of 2018 was preprocessed with radiometric correction, atmospheric correction, and water column correction. Coral reefs were classified using three classification methods included Maximum Likelihood Classification (MLC), Minimum Distance Classification (MDC) and Mahalanobis Distance Classification (MHC) to compare the accuracy of coral reefs classification from remote sensing. Coral reefs were divided into four classes including live coral, dead coral, sand, and sea. The classified images were validated with ground control points obtained from field survey. The results showed that three classification methods could be able to classify the components and the extent of coral reefs. In kradan island, the overall accuracy and kappa coefficients of the MLC, MDC and MHC were 72.73% (0.64), 63.64% (0.51), and 72.73% (0.64), respectively. For the accuracy of coral reef classifies in Ngai Island was found that the MLC, MDC and MHC method have the overall accuracy and kappa coefficients were 65.22% (0.55), 52.17% (0.38), and 65.22% (0.55), respectively. Thus, the MLC and MHD were the most accurate for coral reefs classification.