Analisis Delineasi Garis Pantai dengan Metode Object-Based Image Analysis di Kabupaten Malang Menggunakan Citra Satelit Planetscope: Analysis of Shoreline Delineation Using Object-Based Image Analysis Method in Malang Regency with Planetscope Satellite Imagery

Saphira, Shada
Setyawan, Fahreza Okta
Fuad, M. Arif Zainul
Rijal, Seftiawan Samsu

JFMR (Journal of Fisheries and Marine Research)
Vol. 9 No. 1 (2025): JFMR on March

Salah satu fenomena alam yang rentan terjadi di wilayah pantai adalah perubahan garis pantai, karena sifat pantai yang dinamis, mengakibatkan wilayah pantai selalu mengalami perubahan. Oleh karena itu, delineasi garis pantai penting untuk dilakukan secara berkala agar garis pantai dapat dipantau secara terus menerus untuk mencegah terjadinya abrasi maupun akresi, juga delineasi garis pantai merupakan hal yang esensial dalam pengelolaan wilayah pesisir berkelanjutan. Penelitian ini berfokus pada delineasi garis pantai berdasarkan metode penginderaan jauh menggunakan Citra Satelit Resolusi Tinggi (CSRT) di Kabupaten Malang, dimana perubahan garis pantai terjadi wilayahnya. Delineasi garis pantai dilakukan dengan menggunakan metode integrasi dari model deep learning, Convolutional Neural Network (CNN), dan Object Based Image Analysis (OBIA). Penelitian ini bertujuan untuk mendapatkan dan mengevaluasi garis pantai dari metode yang digunakan. Didapatkan garis pantai hasil metode OBIA Kabupaten Malang di bagian barat 34.49 km, tengah 51.21 km, dan timur 67.81 km. Nilai overall accuracy yang didapatkan sebesar 83.62% di barat, tengah 80.79%, dan timur 85.06%. Jarak antara garis pantai hasil ekstraksi dan tracking dibandingkan dan didapatkan nilai RMSE di barat sebesar 4.91 meter, tengah 6.62 meter, dan timur 3.68 meter. Hasil tersebut membuktikan bahwa integrasi metode CNN dan OBIA berhasil menciptakan garis pantai dengan akurasi yang baik.   One of the natural phenomena commonly occurring in coastal areas is shoreline change, driven by the coast’s dynamic nature, resulting in coastal areas always experiencing changes. Therefore, delineation of the shoreline is important to be carried out periodically so that the shoreline can be monitored continuously to prevent abrasion and accretion, also delineation of the shoreline is essential in the management of sustainable coastal areas. This research focuses on shoreline delineation based on remote sensing method using High Resolution satellite imagery in Malang Regency, where shoreline changes occur. Shoreline delineation is performed using an integrated method of deep learning models, Convolutional Neural Network (CNN), and Object Based Image Analysis (OBIA). This research aims to obtain and evaluate the shoreline of the method used. The shoreline obtained from OBIA method in Malang Regency is 34.49 km in the west, 51.21 km in the central, and 67.81 km in the east. The overall accuracy value achieved in the western part is 83.62%, the central part is 80.79%, and the eastern part is 85.06%. The distance between the extracted and the tracking shoreline was compared and an overall average RMSE value of 4.91 meters in the western part, 6.62 meters in the middle part, and 3.68 meters in the eastern part. These results prove that the integration of CNN and OBIA methods can successfully create shoreline with good accuracy.