Perancangan Aplikasi Object Detection Guna Mengidentifikasi Kematangan Buah Melon Apollo Menggunakan Metode YOLOv8

Akbar, Muhammad Irfan (2026) Perancangan Aplikasi Object Detection Guna Mengidentifikasi Kematangan Buah Melon Apollo Menggunakan Metode YOLOv8. Bachelor thesis, Institut Teknologi Kalimantan.

	Text 04221028_cover.pdf Restricted to Repository staff only until 4 October 2028. Download (334kB) | Request a copy
	Text 04221028_statement_of_authenticity.pdf Restricted to Repository staff only until 4 October 2028. Download (214kB) | Request a copy
	Text 04221028_publishing_agreement.pdf Restricted to Repository staff only until 4 October 2028. Download (262kB) | Request a copy
	Text 04221028_approval_sheet.pdf Restricted to Repository staff only until 4 October 2028. Download (220kB) | Request a copy
	Text 04221028_preface.pdf Restricted to Repository staff only until 4 October 2028. Download (312kB) | Request a copy
	Text 04221028_abstract_id.pdf Restricted to Repository staff only until 4 October 2028. Download (332kB) | Request a copy
	Text 04221028_abstract_en.pdf Restricted to Repository staff only until 4 October 2028. Download (323kB) | Request a copy
	Text 04221028_table_of_content.pdf Restricted to Repository staff only until 4 October 2028. Download (410kB) | Request a copy
	Text 04221028_illustrations.pdf Restricted to Repository staff only until 4 October 2028. Download (321kB) | Request a copy
	Text 04221028_tables.pdf Restricted to Repository staff only until 4 October 2028. Download (290kB) | Request a copy
	Text 04221028_notations.pdf Restricted to Repository staff only until 4 October 2028. Download (296kB) | Request a copy
	Text 04221028_chapter_1.pdf Restricted to Repository staff only until 4 October 2028. Download (426kB) | Request a copy
	Text 04221028_chapter_2.pdf Restricted to Repository staff only until 4 October 2028. Download (552kB) | Request a copy
	Text 04221028_chapter_3.pdf Restricted to Repository staff only until 4 October 2028. Download (1MB) | Request a copy
	Text 04221028_chapter_4.pdf Restricted to Repository staff only until 4 October 2028. Download (4MB) | Request a copy
	Text 04221028_conclusions.pdf Restricted to Repository staff only until 4 October 2028. Download (335kB) | Request a copy
	Text 04221028_bibliography.pdf Restricted to Repository staff only until 4 October 2028. Download (371kB) | Request a copy
	Text 04221028_enclosure.pdf Restricted to Repository staff only until 4 October 2028. Download (430kB) | Request a copy
	Text 04221028_paper.pdf Restricted to Repository staff only until 4 October 2028. Download (519kB) | Request a copy
	Text 04221028_presentation.pdf Restricted to Repository staff only until 4 October 2028. Download (1MB) | Request a copy
	Text 04221028_Form. TA-020.pdf Restricted to Repository staff only until 4 October 2028. Download (272kB) | Request a copy

Abstract

Penentuan kematangan buah melon Apollo selama ini masih dilakukan secara visual dan subjektif sehingga berpotensi menimbulkan kesalahan penilaian. Guna mengatasi permasalahan tersebut, sebuah aplikasi object detection berbasis Android dirancang menggunakan metode YOLOv8n untuk mengidentifikasi empat tingkat kematangan melon secara otomatis dan realtime. Dataset citra dua dimensi dikumpulkan langsung dari lahan perkebunan di KM 3,5 Kota Balikpapan, lalu dianotasi menggunakan Roboflow, serta dilengkapi dengan sistem penyimpanan riwayat deteksi melalui Firebase Firestore. Berdasarkan hasil evaluasi, performa yang baik berhasil ditunjukkan oleh model YOLOv8n dengan perolehan nilai mAP@50 keseluruhan sebesar 0,862, serta rata-rata precision dan recall masing-masing sebesar 0,82 dan 0,72. Secara spesifik, nilai Average Precision (AP) diperoleh sebesar 0,974 (unripe), 0,896 (underripe), 0,846 (ripe), 0,597 (overripe), dan 0,995 (unknown). Rendahnya nilai AP pada kelas overripe diidentifikasi sebagai akibat keterbatasan dataset serta faktor subjektivitas pelabelan kualitatif. Pada pengujian variasi pencahayaan terhadap 40 data uji, tingkat keberhasilan deteksi dapat dicapai pada 29 data dengan rata-rata akurasi sebesar 72,5% dan nilai confidence 77,2%. Meskipun kesalahan klasifikasi akibat intensitas cahaya ekstrem masih ditemukan pada kelas underripe dan overripe, pelokalan objek menggunakan bounding box tetap dapat dilakukan secara stabil pada berbagai kondisi. Melalui seluruh rangkaian pengujian tersebut, metode YOLOv8n dinyatakan efektif untuk diterapkan pada sistem identifikasi kematangan buah melon Apollo berbasis object detection. Kata kunci: Android, Kematangan Buah, Melon Apollo, Object Detection, YOLOv8n

Item Type:	Thesis (Bachelor)
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions:	Jurusan Teknologi Industri dan Proses > Teknik Elektro
Depositing User:	Muhammad Irfan Akbar
Date Deposited:	15 Jul 2026 03:44
Last Modified:	15 Jul 2026 03:44
URI:	http://repository.itk.ac.id/id/eprint/27288

Actions (login required)

View Item