Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus)
The quality of pineapple shows spatial differences within the fruit, which is related to the appearance of flowers in the inflorescence. Therefore, for supporting uniform quality of fresh-cut products, the postharvest industry requires non-contact imaging units to analyse the quality of fresh-cut pi...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Published: |
International Society for Horticultural Science
2022
|
| _version_ | 1811137591140941824 |
|---|---|
| author | Mollazade, K. Hashim, N. |
| author_facet | Mollazade, K. Hashim, N. |
| author_sort | Mollazade, K. |
| collection | UPM |
| description | The quality of pineapple shows spatial differences within the fruit, which is related to the appearance of flowers in the inflorescence. Therefore, for supporting uniform quality of fresh-cut products, the postharvest industry requires non-contact imaging units to analyse the quality of fresh-cut pineapple. In this study, chemical properties of fresh-cut pineapples (n=60) were evaluated by push-broom hyperspectral imaging setup in the range 380-1680 nm. A slice was cut from the middle of each fruit along the stem axis. Immediately, spectral hypercube of the slice was recorded. Then, using a cork borer, 10 cylindrical-shaped samples were extracted from different positions of each slice. From the cylinders (n=600) the chemical properties (soluble solids content, acidity, moisture and carotenoids contents) were analysed. Image registration was carried out to locate the samples (ROIs) in the hyperspectral images. The average spectra of the points in each ROI was considered as the spectral signature of each sample. Principal component analysis was used to compress the preprocessed spectra and consequently to reduce the size of input vector for modeling. Supervised multivariate regression analysis was applied to develop quantitative prediction models for the chemical properties of pineapples. Results showed expected differences of chemical properties between and within the fruit slices. Furthermore, hyperspectral imaging coupled with multivariate methods demonstrated potential for non-contact evaluation of fresh-cut pineapple produce. |
| first_indexed | 2024-09-25T03:36:44Z |
| format | Article |
| id | upm.eprints-101759 |
| institution | Universiti Putra Malaysia |
| last_indexed | 2024-09-25T03:36:44Z |
| publishDate | 2022 |
| publisher | International Society for Horticultural Science |
| record_format | dspace |
| spelling | upm.eprints-1017592024-05-02T06:39:22Z http://psasir.upm.edu.my/id/eprint/101759/ Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus) Mollazade, K. Hashim, N. The quality of pineapple shows spatial differences within the fruit, which is related to the appearance of flowers in the inflorescence. Therefore, for supporting uniform quality of fresh-cut products, the postharvest industry requires non-contact imaging units to analyse the quality of fresh-cut pineapple. In this study, chemical properties of fresh-cut pineapples (n=60) were evaluated by push-broom hyperspectral imaging setup in the range 380-1680 nm. A slice was cut from the middle of each fruit along the stem axis. Immediately, spectral hypercube of the slice was recorded. Then, using a cork borer, 10 cylindrical-shaped samples were extracted from different positions of each slice. From the cylinders (n=600) the chemical properties (soluble solids content, acidity, moisture and carotenoids contents) were analysed. Image registration was carried out to locate the samples (ROIs) in the hyperspectral images. The average spectra of the points in each ROI was considered as the spectral signature of each sample. Principal component analysis was used to compress the preprocessed spectra and consequently to reduce the size of input vector for modeling. Supervised multivariate regression analysis was applied to develop quantitative prediction models for the chemical properties of pineapples. Results showed expected differences of chemical properties between and within the fruit slices. Furthermore, hyperspectral imaging coupled with multivariate methods demonstrated potential for non-contact evaluation of fresh-cut pineapple produce. International Society for Horticultural Science 2022 Article PeerReviewed Mollazade, K. and Hashim, N. (2022) Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus). Acta Horticulturae. art. no. 1353. 79 - 85. ISSN 0567-7572; ESSN: 2406-6168 https://www.ishs.org/ishs-article/1353_11 10.17660/actahortic.2022.1353.11 |
| spellingShingle | Mollazade, K. Hashim, N. Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus) |
| title | Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus) |
| title_full | Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus) |
| title_fullStr | Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus) |
| title_full_unstemmed | Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus) |
| title_short | Hyperspectral imaging for predicting the chemical composition of fresh-cut pineapple (Ananas comosus) |
| title_sort | hyperspectral imaging for predicting the chemical composition of fresh cut pineapple ananas comosus |
| work_keys_str_mv | AT mollazadek hyperspectralimagingforpredictingthechemicalcompositionoffreshcutpineappleananascomosus AT hashimn hyperspectralimagingforpredictingthechemicalcompositionoffreshcutpineappleananascomosus |