Identifikasi Pesawat pada Citra Satelit Menggunakan Metode Convolutional Neural Network dengan Model Pre-trained YOLOv8

Ahmad, Hanif (2024) Identifikasi Pesawat pada Citra Satelit Menggunakan Metode Convolutional Neural Network dengan Model Pre-trained YOLOv8. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 06111940000101-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (9MB) | Request a copy

Abstract

Tangkapan citra satelit memperlihatkan pengertian lebih dalam terhadap berbagai sektor termasuk didalamnya wisata, pertanian, keamanan negara, dan keuangan. Dalam bidang aviasi citra satelit memiliki peran penting dalam berbagai aspek penerbangan pesawat, mulai dari perencanaan penerbangan hingga navigasi dan keselamatan. Identifikasi pesawat dalam citra satelit dapat digunakan untuk mengelola lalu lintas udara dan membantu mencegah tabrakan antar pesawat. Dalam upaya memenuhi kebutuhan pasar tersebut terbitlah layanan komersil seperti Keyhole dan DigitalGlobe yang mendagangkan tangkapan citra satelit tersebut. Jumlah data satelit yang ditangkap tiap harinya untuk memenuhi kebutuhan pasar tersebut kian membesar sampai titik dimana proses memilah tidak mungkin lagi dilakukan oleh seorang manusia, ada kebutuhan akan bantuan mesin spesifiknya visi komputer untuk membantu mengotomatiskan proses analisis. Metode Convolutional Neural Networks mendapati popularitas dalam tugas visi komputer dengan kemampuanya mempelajari representasi data visual yang kompleks. Pada Tugas akhir ini telah dilakukan penerapan metoda Convolutional Neural Networks dengan model pre-trained YOLOv8 sebagai solusi untuk identifikasi pesawat pada tangkapan satelit, menggunakan Hyperparameter optimal didapati model terbaik memberikan akurasi mAP50=0.77716.
============================================================
Satellite imagery provides a deeper understanding of various sectors including tourism, agriculture, national security, and finance. Satellite imagery plays an important role in many aspects of aviation, from flight planning to navigation and safety. Aircraft identification in satellite imagery can be used to manage air traffic and help prevent collisions between aircraft. In an effort to meet these market needs, commercial services such as Keyhole and DigitalGlobe have emerged to trade the imagery. The amount of satellite data captured each day to meet these market needs has grown to the point where the process of sorting it out is no longer possible for a human, there is a need for machine vision specifically to help automate the analysis process. The Convolutional Neural Networks method has gained popularity in computer vision tasks with its ability to learn complex visual data representations. In this final project, we will apply the Convolutional Neural Networks method with YOLOv8 pre-trained model as a solution for aircraft identification in satellite captures. Using the optimal hyperparameter obtained the best model provides an accuracy of mAP50=0.77716.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Convolutional Neural Networks, YOLO, Identifikasi, Satelit, Pesawat, Convolutional Neural Networks, YOLO, Identification, Satellite, Airplane.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Mathematics, Computation, and Data Science > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User:	Ahmad Hanif
Date Deposited:	08 Aug 2024 00:58
Last Modified:	25 Sep 2024 02:08
URI:	http://repository.its.ac.id/id/eprint/114390

Actions (login required)

View Item