Near-Infrared Spectroscopy Detection of Cotton/Polyester Content Based on Dropout Deep Belief Network

Table 1. Statistical results of each component content in cotton-polyester blended fabric
View table
Table 1. Statistical results of each component content in cotton-polyester blended fabric
Component Mass fraction /%
Maximum Minimum Mean
Cotton 99.40 5.13 52.16
Polyester 94.87 0.60 47.84

Table 2. Model performance comparison of different DBN structures

View table

Table 2. Model performance comparison of different DBN structures

DBNstructure	Meanerror	Reconstruction error
DBNstructure	Meanerror	GRBM1	GRBM2	GRBM3
102-90-50	0.1664	0.2152	0.1176
102-80-50	0.1676	0.2001	0.1351
102-70-40	0.2161	0.2569	0.1752
102-60-30	0.1860	0.2184	0.1535
102-50-30	0.1036	0.1055	0.1016
102-80-40-20	0.1328	0.1521	0.1369	0.1095
102-70-30-10	0.1803	0.2291	0.1704	0.1414
102-60-20-10	0.1319	0.1504	0.1429	0.1023
102-50-20-10	0.0794	0.0973	0.0789	0.0619

Table 3. Comparison of prediction results of different models
View table
Table 3. Comparison of prediction results of different models
Model Cotton Polyester
R_P RMSEP R_P RMSEP
PLS 0.9128 0.2723 0.8957 0.2978
BP-ANN 0.9618 0.1473 0.9709 0.1462
DBN 0.9864 0.1055 0.9839 0.1133
Dropout-DBN 0.9927 0.0792 0.9903 0.0869

Tools

Get Citation

Copy Citation Text

Zhuo Wei, Wenwen Li, Min Lin, Wensong Jiang, Xinqi Zhou. Near-Infrared Spectroscopy Detection of Cotton/Polyester Content Based on Dropout Deep Belief Network[J]. Acta Optica Sinica, 2021, 41(17): 1730002

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites