Investigating the Applicability of a Snapshot Computed Tomography Imaging Spectrometer for the Prediction of Brix and pH of Grapes

Abstract

In this paper, a recently developed snapshot hyperspectral imaging (HSI) system based on Computed Tomography Imaging Spectroscopy (CTIS) is utilized to determine Brix and pH values in Sheegene 20 table grapes through Partial Least Squares Regression (PLSR) modeling. The performance of the CTIS system is compared with that of a state-of-the-art line scan HSI system by imaging 100 grapes across both platforms. Reference measurements of Brix and pH values are obtained directly using a refractometer and a pH meter, as these parameters are essential for assessing the quality of table and wine grapes. The findings indicate that the spectra captured by the CTIS camera correlate well with the reference measurements, despite the system's narrower spectral range. The CTIS camera's advantages, including its lower cost, portability, and reduced susceptibility to motion errors, highlight its potential for promising in-field applications in grape quality assessment.

Figures (10)

• Figure 1: Flowchart of the data pipeline from data acquisition to PLSR models for the snapshot CTIS and line scan system. For each grape, a line scan datacube, snapshot CTIS image, pH measurement and ° Brix measurement are acquired. The datacubes are preprocessed and the mean standardized spectrum (expressed by SNV) of each grape is determined. PLSR models are created from the mean spectra and the measured response variables (pH value and ° Brix number).
• Figure 2: Experimental setup for hyperspectral imaging of grapes with both line scan (Buteo) and snapshot (CTIS) HSI systems.
• Figure 3: Data preprocessing pipeline for line scan hyperspectral images: (a) Mean raw datacube, (b) reflectance datacube cropped to region with grapes, (c) SAM score map for datacube and mean reflectance spectrum of a grape, (d) isolated grape datacubes by SAM score and thresholding, and (e) final grape datacubes after removal of reflections. All images are illustrated using the viridis colormap.
• Figure 4: Data preprocessing pipeline for snapshot hyperspectral images. (a) Cropped and dark frame subtracted CTIS image, (b) mean reconstructed datacube (along spectral dimension), and (c) masked reconstructed datacube from which the mean spectrum is calculated. All images are illustrated using the viridis colormap.
• Figure 5: (a) Transmission of optical bandpass filters from 600-850 nm used for the acquisition of training data for the U-Net neural network. (b) Experimental setup for data acquisition of CTIS and multispectral images for ANN training. (c) Filter placement inside the filter wheel without the front cover.