Prediksi Parameter Kematangan Buah Melon Menggunakan Spektroskopi Near Infra-red
| Main Authors: | Ahmad, Usman; Departemen Teknik Mesin dan Biosistem, Fakultas Teknologi Pertanian, Institut Pertanian Bogor, Kampus IPB Dramaga, Bogor 16680, Sabihah, .; Program Studi Teknik Mesin dan Biosistem, Fakultas Teknologi Pertanian, Institut Pertanian Bogor, Kampus IPB Dramaga, Bogor 16680 |
|---|---|
| Format: | Article info application/pdf eJournal |
| Bahasa: | eng |
| Terbitan: |
Institut Pertanian Bogor
, 2018
|
| Subjects: | |
| Online Access: |
http://journal.ipb.ac.id/index.php/JIPI/article/view/24463 http://journal.ipb.ac.id/index.php/JIPI/article/view/24463/15967 |
Daftar Isi:
- Prediction of ripeness level of melon fruit is conducted manually by looking at the colors and nets on the skin, or the sound of a knocking on the melon fruit. This manual method produces an inconsistent degree of accuracy that cannot be applied in harvesting. The use of spectroscopy in the near infra-red region (NIR) is a way of improving consistency and accelerating the process of estimating the fruit ripeness level without damaging the product. The objective of this research was to study the ripeness parameters of the Golden-Apollo variety of melon using NIR spectroscopy method. Parameters of fruit ripeness studied were moisture content, total soluble solids (TSS), and fruit firmness. The material used was melon fruits with four different harvesting ages. The equipment used was NIRFlex N-500 Solid Optical Fiber spectrometer that worked at 1.000-2.500 nm wavelengths. The method used in calibration and validation of NIR spectrum data and reference data was partial least square (PLS) with pre-processing of spectral data i.e., normalization, first and second derivatives of savitzky-golay, and multiplicate-scatter correction. The results showed that the best predictive model obtained to predict ripeness level of melon fruit was to use the ripeness parameters of total soluble solids and firmness of meat. The moisture parameter resulted in a poor predictive model that could not be used to predict a melon ripeness level with a good result.