Research Article
Quantitative Detection of Quartz Sandstone SiO2 Grade Using Polarized Infrared Absorption Spectroscopy with Convolutional Neural Network Model
Figure 8
Prediction of the PCA-CNN model: (a) original data; (b) first-order derivative; (c) reciprocal logarithm; and (d) MSC. The black and red lines show the 1 : 1 relationship and the model trend. The reciprocal logarithm-PCA-GRNN model falls within the largest number of points within the 95% confidence band.
(a) |
(b) |
(c) |
(d) |