UTILIZATION OF UNMANNED AERIAL VEHICLE (UAV) FOR THE DETECTION OF MARINE DEBRIS WITH SUPERVISED CLASSIFICATION METHODS ON THE COAST OF CINTA BEACH, TANJUNG PASIR, TANGERANG REGENCY

• Main Author: Dwi Marco, Jovany
• Format: Final Project
• Language: Indonesia
• Subjects: Geologi, hidrologi & meteorologi
• Online Access: https://digilib.itb.ac.id/gdl/view/67461
• Institution: Institut Teknologi Bandung

The increase in social and economic aspects due to tourism activities at Cinta Beach is followed by an increase in the amount of marine debris distributed to the sea and coast. Marine debris is waste that comes from land, water bodies, and coasts flowing into the sea or waste that comes from activities in the sea. Research related to the distribution of marine debris is still manually using direct transects, which is time consuming. In this study, Unmanned Aerial Vehicle (UAV) technology was used as an alternative to mapping the distribution of marine debris. The purpose of this study to detect the distribution of marine debris on the Cinta Coast, Tanjung Pasir, Tangerang Regency using two supervised classification methods, namely Maximum Likelihood Classification (MLC) and Minimum Distance Classification (MDC). The data used in this study are 150 meters of aerial photo data and 100 meters of direct transect data that was taken on June 15th, 2021 and also data on oceanographic parameters such as tides, wind and currents.Then the results of the classification are compared with the results of direct transects and analyzed the characteristics of oceanographic parameters that related to the distribution of marine debris. The results of the guided classification using the Maximum Likelihood Classification (MLC) method produced the largest area, namely sand covering an area of 10,776.4 m2 followed by plastic waste with an area of 65.5 m2 and wood with an area of 5.1 m2. While the Minimum Distance Classification (MDC) method produces a sand area of 6,919.65 m2, wood with an area of 4,854.11 m2 and plastic waste with an area of 65.3 m2. In this study, an accuracy test calculation was carried out on the classification results using the confusion matrix table as a validation test. There are differences in the accuracy results of the two methods. The Maximum Likelihood Classification (MLC) method is superior with an overall accuracy value of 95.65%, while in the Minimum Distance Classification (MDC) it is only 81.59%. Maximum Likelihood Classification (MLC) is the best method for mapping the distribution of marine debris and this method corresponds to the results of direct transects.