Our ChemWiz-ADK-FV (Fruits and Veggies) can be configured as handheld, process-line, lab mountable, or as a rugged case system and can be transported between field, processing, and lab inspection points. Our technology is non-contact and high-speed sampling ensures fast, accurate testing without the need for any wet chemistry.
Strawberry Spectroscopy: Calibration Based on the Picked Date
Strawberry Field Trip
Strawberry-bunch
Strawberry-photo-at-sign
Medallion-strawberry
Strawberry-up-close
Sensation-strawberry
Florida127-Srawberry
Strawberry-picking
Pink-strawberry-bunch
Felicity-strawberry
Brilliance-strawberry
Strawberry-bagged
Abstract
Strawberry quality is a major concern for commercial stores, yet determining freshness after transportation is challenging. This study employs UV-VIS spectra to determine the days elapsed since harvest. The methodology involves collecting spectra from multiple strawberries of various cultivars over nine days and training a K-Nearest Neighbors (KNN) machine learning algorithm. Various classification models are evaluated for freshness, considering their accuracy and applicability. While two-class models attain 95.19% accuracy, nine-class and three-class variants offer significant value. The nine-class version furnishes detailed age estimates crucial for scientific research or shelf-life determination, whereas coarser models, like the three-class or two-class versions, simplify decision-making, ideal for applications such as quality control or logistics. This quantitative approach complements qualitative evaluations, meeting diverse marketing needs in the strawberry industry.
The graphs below show the Two-class classifications (Before/at and after Day 4) Accuracy (%) column in Figure 2 of the paper:
Brilliance
FL-127
Medallion
Several classification calibration models are being created using different combinations of processing steps and wavelengths of the collected spectra. The processing steps experimented with are…
