CD Skripsi

Prediksi Kandungan Beta Karoten Pada Tandan Buah Segar Kelapa Sawit Menggunakan Pencitraan Multispektral

---

The quality of Crude Palm Oil (CPO) can be determined through the assessment of beta-carotene content in Fresh Fruit Bunches (FFB) of oil palm. This study aims to utilize an LED-based multispectral imaging system to predict the beta-carotene content in FFB. The imaging system employs eight LED wavelengths (680 nm, 700 nm, 750 nm, 780 nm, 810 nm, 850 nm, 880 nm, and 900 nm) as light sources. A Python-based program is used for image acquisition and processing, as well as object detection modeling. Recorded multispectral images are processed to obtain the relative reflectance intensity values for each wavelength. Support Vector Machine (SVM) method is used to predict the beta-carotene content in FFB. Beta- carotene levels measured using Soxhlet extraction range from 141 to 286 ppm. The Principal Component Analysis (PCA) method is applied to recognize reflectance intensity patterns in multispectral images for two ripeness categories based on beta- carotene values. Support Regression Matrix is employed to analyze the performance of the predictive model. The study results indicate that the relative reflectance intensity of ripe FFB is 17.74% higher than unripe FFB across all LED wavelengths. PCA successfully separates the two ripeness levels of FFB with a cumulative variance of 97.41% (PC1 = 91.20% and PC2 = 6.21%). The SVM model achieves a prediction accuracy of 95%, with a Mean Absolute Error (MAE) of 0.04, Mean Squared Error (MSE) of 0.007, and a coefficient of determination (R²) of
0.97. The results demonstrate the potential of LED-based multispectral imaging for predicting beta-carotene content in FFB
Keywords: Multispectral imaging, beta-carotene content, Principal Component Analysis (PCA), Support Vector Machine (SVM)

---

Ketersediaan

Informasi Detail

Judul Seri

-

No. Panggil

2003114233

Penerbit

Pekanbaru : Universitas Riau FMIPA Fisika., 2024

Deskripsi Fisik

-

Bahasa

Indonesia

ISBN/ISSN

-

Klasifikasi

2003114233

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