Related Experiment Video
Updated: Mar 11, 2026
11:37
RGB and Spectral Root Imaging for Plant Phenotyping and Physiological Research: Experimental Setup and Imaging Protocols
Published on: August 8, 2017
17.1K
Methods for Improving Image Quality and Reducing Data Load of NIR Hyperspectral Images.
Ferenc Firtha1, András Fekete2, Tímea Kaszab2
1Corvinus University of Budapest, Faculty of Food Science, Department of Physics and Control, Somlóiút 14-16, H-1118 Budapest, Hungary. ferenc.firtha@uni-corvinus.hu.
Sensors (Basel, Switzerland)
|November 24, 2016
Summary
Near Infrared Hyperspectral Imaging (NIRHSI) addresses data load and dead pixels. This study introduces a novel approach for real-time feature extraction and dead pixel compensation, improving NIRHSI applicability.
Area of Science:
- Multispectral Imaging
- Spectroscopy
- Data Science
Background:
- Near Infrared Hyperspectral Imaging (NIRHSI) offers combined spatial and spectral data but faces challenges.
- High data volumes (>50 MB/image) strain computer memory, limiting industrial use.
- NIR sensors have 'dead' pixels (approx. 1%), which are problematic for >100 wavelength hyperspectral images.
Purpose of the Study:
- To develop methods for reducing data load in hyperspectral imaging.
- To implement a systematic procedure for compensating 'dead' pixels in NIR sensors.
- To assess the feasibility of these approaches for predicting moisture content in carrot tissue.
Main Methods:
- Utilized sample-specific vector-to-scalar operators for real-time feature extraction.
- Developed a systematic procedure for compensating unserviceable ('dead') pixels in NIR detectors.
- Applied the developed methods to hyperspectral imaging experiments for moisture content prediction.
Main Results:
- Demonstrated a feasible approach for reducing data load in hyperspectral experiments.
- Successfully compensated for 'dead' pixels in NIR sensor data.
- Validated the approach's effectiveness in predicting moisture content in carrot tissue.
Conclusions:
- The proposed methods effectively reduce data load and compensate for dead pixels in NIRHSI.
- This approach enhances the practicality of NIRHSI for online industrial applications.
- The study successfully demonstrated the prediction of moisture content in carrot tissue using the developed techniques.
