CD Skripsi

Deteksi Tandan Buah Kelapa Sawit Normal Dan Cacat Bertumpuk Menggunakan Computer Vision Berbasis Deep Learning

---

The manual classification process of fresh fruit bunches (FFB) in oil palm has limitations, particularly with stacked objects, due to subjective assessment and low efficiency. This research develops an automated system based on computer vision using the YOLOv8 algorithm to detect and classify normal and defective FFB (empty bunches, rotten, long stalks, and long thorns) stacked on a conveyor. Additionally, average RGB intensity analysis is performed to differentiate the visual characteristics between normal and defective FFB. Image acquisition is conducted using an RGB camera, while the system modeling utilizes two algorithm variants, YOLOv8l-Seg and YOLOv8x-Seg, capable of real-time detection. A total of 50 FFB samples are used, consisting of five classes. The results indicate that the computer vision-based system effectively detects normal and defective FFB under stacked conditions. Both algorithms produce outputs in the form of bounding boxes, object segmentation, class labels, and confidence scores in real-time. The RGB intensity analysis shows that the average RGB values for the empty bunch class are higher, with values R: 82.9, G: 84.5, B: 82.8, compared to the normal class values R: 75.2, G: 50.7, B: 42.2, long stalks R: 63.5, G: 44.0, B: 38.3, long thorns R: 70.9, G: 53.6, B: 46.0, and rotten R: 79.0, G: 79.8, B: 77.7.

Keywords: Computer Vision, Deep Learning, RGB Intensity, Sorting and Grading, Stacked Oil Palm Fresh Fruit Bunches, Normal and Defective, YOLOv8 Algorithm.

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

2103111643

Penerbit

Pekanbaru : Universitas Riau FMIPA Fisika., 2025

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

2103111643

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

FISIKA

Info Detail Spesifik

-

Pernyataan Tanggungjawab

ERA

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar