CD Skripsi

Identifikasi Tandan Buah Sawit Berisi Dan Kosong Setelah Proses Perontokan Berbasis Computer Vision Dan Algoritma Yolo Untuk Estimasi Oil Losses

---

The identification of full and empty palm fruit bunches after threshing plays an important role in improving the oil extraction rate (OER) in palm oil mills. Undetected filled palm fruit bunches after threshing can reduce extraction efficiency and lead to oil loss cases. Manual assessment is often slow and inaccurate, so a faster and more effective technology is needed. This study uses computer imaging and the YOLO (You Only Look Once) algorithm to identify filled and empty palm fruit bunches. In this study, two versions of the YOLO algorithm, namely YOLOv8l and YOLOv8-Seg, were applied to a dataset of 526 FFB images, divided into 263 filled and 263 empty bunch samples. The YOLO model was tested using videos of conveyor lines in the mill and the model performance was evaluated using confusion matrix. The model evaluation results show that YOLOv8l has a detection accuracy of 83.8% with a detection speed of 8.5 ms, which is higher and faster than YOLOv8-Seg which reaches 82.8% with a detection speed of 9.6 ms. These results show that the accuracy and detection speed obtained are effective for identifying palm fruit bunches. The results also show that the intensity of the RGB values of the filled bunches is higher than that of the empty bunches.
Keywords: Computer Vision; Palm oil fruits; Unstripped bunches; YOLOv8; RGB intensity 

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

2003114223

Penerbit

Pekanbaru : Universitas Riau FMIPA Fisika., 2024

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

2003114223

Tipe Isi

-

Tipe Media

-

Tipe Pembawa

-

Edisi

-

Subjek

FISIKA

Info Detail Spesifik

-

Pernyataan Tanggungjawab

Mutia

Versi lain/terkait

Lampiran Berkas

Komentar

---

Anda harus masuk sebelum memberikan komentar