Prediction of Degreening Velocity of Broccoli Buds Using Hyperspectral Camera Combined with Artificial Neural Networks

Abstract

Developing a noninvasive technique to estimate the degreening (loss of green color) velocity of harvested broccoli was attempted. Loss of green color on a harvested broccoli head occurs heterogeneously. Therefore, hyperspectral imaging technique that stores spectral reflectance with spatial information was used in the present research. Using artificial neural networks (ANNs), we demonstrated that the reduction velocity of chlorophyll at a site on a broccoli head was related to the second derivative of spectral reflectance data at 15 wavelengths from 405 to 960 nm. The reduction velocity was predicted using the ANNs model with a correlative coefficient of 0.995 and a standard error of prediction of 5.37 × 10^-5 mg·g^-1·d^-1. The estimated reduction velocity was effective for predicting the chlorophyll concentration of broccoli buds until 7 d of storage, which was established as the maximum time for maintaining marketability. This technique may be useful for nondestructive prediction of the shelf life of broccoli heads.

Keywords: Brassica oleracea var. italica; chlorophyll; mathematical model; nondestructive analysis; shelf life; spectroscopy; statistical analysis; vegetable.

Figure 1

Overall view and description of the components in the hyperspectral camera system (JFE Techno-Research Corporation, Tokyo, Japan) [28]. A, Sample; B, Sample stage; C, Lens; D, Spectrograph; E, 12 bit CCD camera; F, 150 W Xe lamp; G, 150 W tungsten halogen lamp; H, 250 mm illumination rod; I, 17 inch monitor; J, desktop computer; K, 40 mm × 220 mm white reference.

Figure 2

Relationship between hue angle and chlorophyll concentration in broccoli buds. The open circles and full line are the actual values and single regression line (C = 0.0104H° − 0.6613). Bias was 2.12 × 10⁻⁴ mg·g⁻¹, standard error of calibration 2.79 × 10⁻² mg·g⁻¹, relative percent difference 4.48, and correlative coefficient of validation 0.962.

Figure 3

(a) Raw and (b) second derivatives of spectral reflectance in the range of 380 and 1000 nm from 11 broccoli heads (110 regions of interest).

Figure 4

Correlative coefficients between the chlorophyll degradation velocity and reflectance at 5 nm wavelength bands between 380 and 1000 nm.

Figure 5

Validation for predicting the chlorophyll degradation velocity of broccoli heads at 60 regions of interest by artificial neural networks. The open circles and full line are the actual values and single regression line. Bias was −4.82 × 10⁻⁷ mg·g⁻¹·d⁻¹, standard error of prediction 5.37 × 10⁻⁵ mg·g⁻¹·d⁻¹, and correlative coefficient of validation was 0.995. This level of correlation was significant at the 99.9% level calculated using the F-test.

Figure 6

Change in mass loss of broccoli heads. The mean ± SE of six observations has been plotted.

Figure 7

Relationship between measured and predicted chlorophyll concentrations after 3 and 7 d of storage. The circles and squares denote the plots for days 3 and 7, respectively. The solid lines denote the linear regression lines for day 3 (Bias 3.31 × 10⁻³ mg·g⁻¹, standard error of prediction 5.03 × 10⁻³ mg·g⁻¹·d⁻¹, correlative coefficient of validation 0.795; significant at the 99.9% level calculated using the F-test). The dashed lines denote the linear regression lines for day 7 (Bias 1.37 × 10⁻² mg·g⁻¹, standard error of prediction 4.33 × 10⁻³ mg·g⁻¹·d⁻¹, correlative coefficient of validation 0.884; significant at the 99.9% level calculated using the F-test).