Classification of crop flours based on protein contents using near infra-red spectroscopy and principle component analysis

Indonesia has abundant crops that can be used as carbohydrate sources. Those crops are made into flour to prolong the shelf life, as well to ease for handling and cooking. Crops as carbohydrate sources are usually high energy but low protein. The aim of this research was to classify flours made of v...

Full description

Saved in:
Bibliographic Details
Published in:IOP conference series. Earth and environmental science 2019-11, Vol.355 (1), p.12002
Main Authors: Yuliyanda, I, Masithoh, R E, Khuriyati, N, Saputro, A D
Format: Article
Language:English
Subjects:
Carbohydrates
Cassava
Chemical analysis
Classification
Cooking
Crops
Fiber optics
Flour
Potatoes
Principal components analysis
Proteins
Shelf life
Spectroscopic analysis
Spectroscopy
Spectrum analysis
Wavelengths
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Indonesia has abundant crops that can be used as carbohydrate sources. Those crops are made into flour to prolong the shelf life, as well to ease for handling and cooking. Crops as carbohydrate sources are usually high energy but low protein. The aim of this research was to classify flours made of various crops using near infra-red spectroscopy (NIRS) and principle component analysis (PCA). The samples used in this study were six types of flour made of banana, breadfruit, taro, arrowroot, purple sweet potato, and modified cassava (mocaf). The reflectance data were taken using the NIRFlex N500 Fiber Optic Solids Cell at wavelengths of 1000-2500 nm. The spectral obtained were pre-processed and analyzed using The Unscrambler X version 10.5.1. Three pre-processing methods were used, i.e. 1st Savitzky Golay Derivative, Normalization, and Standard Normal Variate (SNV). PCA was able to classify flours based on types of crops. The best transformation was SNV which was able to classify all groups of samples with 100% success rate. PCA model was also able to differentiate low and high protein level of samples aligned with the chemical analysis.
ISSN:1755-1307
1755-1315
DOI:10.1088/1755-1315/355/1/012002