Analisis Pengaruh Ground Control Point Terhadap Ketelitian Koordinat Pada Pemetaan Berbasis Unmanned Aerial Vehicle

Abstract

Teknologi Unmanned Aerial Vehicle (UAV) dengan sensor GPS standar berperan penting dalam pemetaan kehutanan beresolusi tinggi, namun masih memerlukan orthorektifikasi untuk memperbaiki distorsi geometrik. Penelitian ini bertujuan menganalisis pengaruh jumlah Ground Control Point (GCP) terhadap akurasi posisi serta membandingkan ketelitian orthophoto tanpa GCP, orthophoto dengan GCP optimum, dan orthophoto hasil koreksi menggunakan citra Sentinel-2C di Hutan Pendidikan Gunung Walat. Data diperoleh di Blok Toso seluas 16 ha menggunakan DJI Mavic 2 Pro dengan variasi 3, 5, dan 7 GCP yang divalidasi menggunakan 3 Independent Control Points (ICP), serta dianalisis menggunakan RMSE dan uji t berpasangan. Hasil menunjukkan bahwa penggunaan GCP dapat meningkatkan akurasi dibandingkan tanpa GCP. Jumlah GCP optimum adalah 5 titik dengan nilai RMSE sebesar 2,395 cm. Perbandingan antara metode 5 GCP dan Sentinel-2C menunjukkan perbedaan signifikan pada sumbu Y dengan rata-rata kesalahan sebesar 11,653 m. Terjadinya perbedaan signifikan pada sumbu Y disebabkan karena terdapat kesalahan yang terakumulasi pada sumbu searah jalur terbang.

Unmanned Aerial Vehicle (UAV) technology equipped with standard GPS sensors that rely on satellites for airborne positioning plays an important role in forest mapping due to its high spatial resolution. However, it still requires an orthorectification process to correct geometric distortions. This study aims to analyze the effect of variations in the number of Ground Control Point (GCP) on positional accuracy and to compare the accuracy of orthophotos without GCPs, orthophotos using the optimum GCP configuration, and orthophotos corrected using Sentinel-2C imagery in the Gunung Walat Educational Forest. Data were acquired in the 16-ha Toso Block using a DJI Mavic 2 Pro. The experimental scheme applied 3, 5, and 7 GCP configurations validated using three Independent Control Points (ICPs), combined with RMSE analysis and paired T-test statistical analysis. The results showed that the use of GCPs significantly improved positional accuracy compared to orthophotos without GCPs. The optimum number of GCPs was 5, producing the lowest RMSE value of 2.395 cm. The comparison between the 5 GCP method and Sentinel-2C-based correction showed a significant difference on the Y-axis, with an average positional error of 11.653 m. The significant difference in the Y-axis is likely caused by the accumulation of positional errors along the direction of the UAV flight path.