Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination
The discrimination power of a hyperspectral imaging system for image segmentation or object detection is determined by the illumination, the camera spatial–spectral resolution, and both the pre-processing and analysis methods used for image processing. In this study, we methodically reviewed the alt...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IM Publications Open
2020-12-01
|
| Series: | Journal of Spectral Imaging |
| Subjects: | |
| Online Access: | https://www.impopen.com/download.php?code=I09_a16 |
| _version_ | 1818622063528640512 |
|---|---|
| author | Carolina Blanch-Pérez del Notario Carlos López-Molina Andy Lambrechts Wouter Saeys |
| author_facet | Carolina Blanch-Pérez del Notario Carlos López-Molina Andy Lambrechts Wouter Saeys |
| author_sort | Carolina Blanch-Pérez del Notario |
| collection | DOAJ |
| description | The discrimination power of a hyperspectral imaging system for image segmentation or object detection is determined by the illumination, the camera spatial–spectral resolution, and both the pre-processing and analysis methods used for image processing. In this study, we methodically reviewed the alternatives for each of those factors for a case study from the food industry to provide guidance in the construction and configuration of hyperspectral imaging systems in the visible near infrared range for food quality inspection. We investigated both halogen- and LED-based illuminations and considered cameras with different spatial–spectral resolution trade-offs. At the level of the data analysis, we evaluated the impact of binning, median filtering and bilateral filtering as pre- or post-processing and compared pixel-based classifiers with convolutional neural networks for a challenging application in the food industry, namely ingredient identification in a flour–seed mix. Starting from a basic configuration and by modifying the combination of system aspects we were able to increase the mean accuracy by at least 25 %. In addition, different trade-offs in performance-complexity were identified for different combinations of system parameters, allowing adaptation to diverse application requirements. |
| first_indexed | 2024-12-16T18:19:13Z |
| format | Article |
| id | doaj.art-256fb5106fb44acca27a3e6e46715743 |
| institution | Directory Open Access Journal |
| issn | 2040-4565 2040-4565 |
| language | English |
| last_indexed | 2024-12-16T18:19:13Z |
| publishDate | 2020-12-01 |
| publisher | IM Publications Open |
| record_format | Article |
| series | Journal of Spectral Imaging |
| spelling | doaj.art-256fb5106fb44acca27a3e6e467157432022-12-21T22:21:36ZengIM Publications OpenJournal of Spectral Imaging2040-45652040-45652020-12-0191a1610.1255/jsi.2020.a16Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discriminationCarolina Blanch-Pérez del Notario0Carlos López-Molina1https://orcid.org/0000-0002-0904-9834Andy Lambrechts2Wouter Saeys3https://orcid.org/0000-0002-5849-4301Imec, Kapeldreef 75, 3001, Leuven, Belgium and KU Leuven, Department of Biosystems, MeBioS, Kasteelpark Arenberg 30, 3001, Leuven, BelgiumCarlos López-MolinaImec, Kapeldreef 75, 3001, Leuven, BelgiumKU Leuven, Department of Biosystems, MeBioS, Kasteelpark Arenberg 30, 3001, Leuven, BelgiumThe discrimination power of a hyperspectral imaging system for image segmentation or object detection is determined by the illumination, the camera spatial–spectral resolution, and both the pre-processing and analysis methods used for image processing. In this study, we methodically reviewed the alternatives for each of those factors for a case study from the food industry to provide guidance in the construction and configuration of hyperspectral imaging systems in the visible near infrared range for food quality inspection. We investigated both halogen- and LED-based illuminations and considered cameras with different spatial–spectral resolution trade-offs. At the level of the data analysis, we evaluated the impact of binning, median filtering and bilateral filtering as pre- or post-processing and compared pixel-based classifiers with convolutional neural networks for a challenging application in the food industry, namely ingredient identification in a flour–seed mix. Starting from a basic configuration and by modifying the combination of system aspects we were able to increase the mean accuracy by at least 25 %. In addition, different trade-offs in performance-complexity were identified for different combinations of system parameters, allowing adaptation to diverse application requirements.https://www.impopen.com/download.php?code=I09_a16system parametershyperspectralilluminationpre- and post-processingclassification accuracyconvolutional neural networksspatial–spectral resolution |
| spellingShingle | Carolina Blanch-Pérez del Notario Carlos López-Molina Andy Lambrechts Wouter Saeys Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination Journal of Spectral Imaging system parameters hyperspectral illumination pre- and post-processing classification accuracy convolutional neural networks spatial–spectral resolution |
| title | Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination |
| title_full | Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination |
| title_fullStr | Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination |
| title_full_unstemmed | Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination |
| title_short | Hyperspectral system trade-offs for illumination, hardware and analysis methods: a case study of seed mix ingredient discrimination |
| title_sort | hyperspectral system trade offs for illumination hardware and analysis methods a case study of seed mix ingredient discrimination |
| topic | system parameters hyperspectral illumination pre- and post-processing classification accuracy convolutional neural networks spatial–spectral resolution |
| url | https://www.impopen.com/download.php?code=I09_a16 |
| work_keys_str_mv | AT carolinablanchperezdelnotario hyperspectralsystemtradeoffsforilluminationhardwareandanalysismethodsacasestudyofseedmixingredientdiscrimination AT carloslopezmolina hyperspectralsystemtradeoffsforilluminationhardwareandanalysismethodsacasestudyofseedmixingredientdiscrimination AT andylambrechts hyperspectralsystemtradeoffsforilluminationhardwareandanalysismethodsacasestudyofseedmixingredientdiscrimination AT woutersaeys hyperspectralsystemtradeoffsforilluminationhardwareandanalysismethodsacasestudyofseedmixingredientdiscrimination |