Quality evaluation of Hanyuan Zanthoxylum bungeanum Maxim. Using computer vision system combined with artificial neural network: A novel method

Figures & data

Figure 1. The procedure designed for predicting the chemical attributes of HZB by computer vision system and artificial neural networks.

Table 1. Colour parameters, VOC, and IMC of HZB samples.

Sample no	L*	a*	b*	VOC (mL)	TALC (%)	IMC (%)
1–40	90.30 ± 2.87	150.90 ± 2.09	141.56 ± 1.37	0.97 ± 0.25	15.30 ± 4.39	1.73 ± 0.45
40–80	82.57 ± 1.44	147.74 ± 0.90	140.10 ± 0.38	0.80 ± 0.15	11.01 ± 2.82	2.88 ± 0.50
80–120	88.15 ± 4.69	146.88 ± 1.61	139.78 ± 1.32	0.53 ± 0.10	5.71 ± 3.76	3.80 ± 0.65

Data were expressed as Mean ± SD; Colour parameters including CIE L*, a*, and b* were determined by CVS; VOC, TALC, and IMC means the content of volatile oil (VOC), total alkylamides (TALC) and impurity (IMC), respectively.

Table 2. Result of method validation of colour measurement (CIE L*, a*, and b*) with RSD for Grade I, Grade II, and Grade III, respectively.

Grade no	RSD
	L*	a*	b*
Grade I	2.02%	0.55%	0.62
Grade II	1.18%	0.40%	0.20%
Grade III	3.56%	0.63%	0.70%

Table 3. Result of method validation of TALC measurement.

Method validation	RSD
Precision	0.81%
Stability	1.15%
Repeatability	1.10%
Recovery	2.21%

Table 4. The relationships between HZB colour and the content of TALC, VOC, and IMC are presented in the correlation matrix (presenting Pearson correlation coefficients).

Matrix	a*	b*	VOC content	TALC content	IMC content
a*	—
b*	0.687	—
VOC content	0.761	0.627	—
TALC content	0.676	0.522	0.683	—
IMC content	−0.811	−0.559	−0.707	−0.848	—

Figure 2. Correlations between the predicted and experimental values of VOC (A), TALC (B), and IMC (C) after trained.