AUTOMATIC PALM COUNTING DENGAN METODE TEMPLATE MATCHING (STUDI KASUS DI UNIVERSITAS SAMUDRA)

Abstract

The oil palm land donated to Universitas Samudra is planned for the development of the campus area, including the construction of a number of buildings and supporting facilities. However, the process of identifying and mapping oil palm plants has been done manually, which is time-consuming, inefficient, and prone to errors. This problem underscores the need for faster and more accurate methods to support spatial data-based planning. This study aimed to calculate the number of oil palm plants in 2022 and 2023 at the University of Samudra using the template matching method with eCognition Developer software, as well as evaluate the accuracy of automatic detection results based on aerial images obtained using drones. The research was carried out using survey methods and descriptive analysis, involving primary data in the form of aerial imagery and field validation, as well as secondary data from the map of the oil palm plantation area of Samudra University. The results of the study show that the number of oil palm plants in 2022 based on automatic calculations was 2,060 trees, while the results of manual validation showed the actual number of 2,169 trees with a difference of 109 trees. In 2023, the automatic calculation detected 1,932 trees, while the actual number was 2,030 trees, with a difference of 98 trees. The accuracy level of automatic calculations in 2022 had an average accuracy of 98.56%, recall of 94.05%, and F1-score of 95.63%, higher than in 2023 with precision of 97.41%, recall of 92.73%, and F1-score of 94.98%. Then the template matching method is effectively used for oil palm tree detection and can support the planning of campus area development efficiently. The use of this technology is expected to be a model that can be implemented in various other educational institutions.

Keywords

• citra drone
• kelapa sawit
• template matching
• pemetaan otomatis