• Optical Instruments
  • Vol. 45, Issue 4, 71 (2023)
Shaobo WANG1, Jiangkun ZHANG1, Qingbiao CHENG1, Ning SHEN1..., Jie LIU2 and Jie FENG1,*|Show fewer author(s)
Author Affiliations
  • 1College of Physics and Electronic Information, Yunnan Normal University, Kunming 650000, China
  • 2Yunnan Provincial Museum, Kunming 650000, China
  • show less
    DOI: 10.3969/j.issn.1005-5630.2023.004.010 Cite this Article
    Shaobo WANG, Jiangkun ZHANG, Qingbiao CHENG, Ning SHEN, Jie LIU, Jie FENG. Spectral reflectivity based tea concentration prediction for tea dyeing of rice paper[J]. Optical Instruments, 2023, 45(4): 71 Copy Citation Text show less
    Images of two kinds of tea dyed rice papers using distilled water and tap water
    Fig. 1. Images of two kinds of tea dyed rice papers using distilled water and tap water
    The average spectral reflectance of two kinds of tea dyed rice papers using distilled water and tap water
    Fig. 2. The average spectral reflectance of two kinds of tea dyed rice papers using distilled water and tap water
    Images of tea dyed rice paper of four brands
    Fig. 3. Images of tea dyed rice paper of four brands
    Spectral curves of 300 samples of each of the four brands of ricepaper
    Fig. 4. Spectral curves of 300 samples of each of the four brands of ricepaper
    Results of PLS model estimation for four brands of rice paper
    Fig. 5. Results of PLS model estimation for four brands of rice paper
    Results of BP neural network model for four brands of rice paper
    Fig. 6. Results of BP neural network model for four brands of rice paper
    The characteristic spectral wavelengths of four brands of rice paper were screened by SPA
    Fig. 7. The characteristic spectral wavelengths of four brands of rice paper were screened by SPA
    茶叶使用量Paper1Paper2
    25 g35 g25 g35 g
    蒸馏水500 mLL*=72.207 a*=3.172 b*=19.341 L*=66.760 a*=4.547 b*=20.257 L*=71.186 a*=4.775 b*=22.072 L*=66.597 a*=5.490 b*=22.957
    自来水500 mLL*=72.732 a*=3.171 b*=19.328 L*=67.067 a*=4.611 b*=20.656 L*=71.890 a*=4.764 b*=22.052 L*=66.763 a*=4.919 b*=21.937
    色差0.32320.39460.66640.5350
    Table 1. Chromatism of two kinds of tea dyed rice papers using distilled water and tap water
    样本训练集测试集
    预测正确率/%R2RMSE预测正确率/%R2RMSE
    Paper190.290.90262.525088.130.87262.7988
    Paper293.610.91782.112391.730.90662.3162
    Paper395.550.95061.701093.350.93191.8980
    Paper496.330.97240.921295.690.95211.0121
    平均93.950.93591.814992.230.91582.0063
    Table 2. PLS model results of four brands of rice paper
    样本训练集测试集
    预测正确率/%R2RMSE预测正确率/%R2RMSE
    Paper195.640.92941.002895.020.92110.9626
    Paper297.540.95310.920696.610.94880.9937
    Paper396.980.98640.804896.920.95770.8657
    Paper498.270.99190.707998.060.98860.7156
    平均97.110.96520.859096.650.95410.8844
    Table 3. BP neural network model results of four brands of rice paper
    模型样本训练集测试集
    预测正确率/%R2RMSE预测正确率/%R2RMSE
    SPA-PLSPaper190.870.86892.224191.690.89942.0906
    Paper294.470.93462.183294.470.93462.1834
    Paper396.740.98371.081796.700.98221.1391
    Paper497.750.98950.763698.470.99310.5170
    平均94.960.94421.563295.330.95231.4826
    SPA-BPPaper197.600.98171.178098.220.98191.1618
    Paper298.320.99450.642098.160.99351.0628
    Paper398.770.99050.604098.360.99210.6122
    Paper498.490.99380.675398.840.99650.5363
    平均98.300.99010.774398.400.99100.8433
    Table 4. Results of SPA-PLS and SPA-BP neural network estimation models for four brands of rice paper
    Shaobo WANG, Jiangkun ZHANG, Qingbiao CHENG, Ning SHEN, Jie LIU, Jie FENG. Spectral reflectivity based tea concentration prediction for tea dyeing of rice paper[J]. Optical Instruments, 2023, 45(4): 71
    Download Citation