高光谱成像
人工智能
预处理器
计算机科学
特征(语言学)
数据预处理
模式识别(心理学)
机器学习
语言学
哲学
作者
Zhiyong Zou,Qingsong Wu,Jian Wang,Lijia Xu,Man Zhou,Zhiwei Lü,Yong He,Yuchao Wang,Bi Liu,Yongpeng Zhao
标识
DOI:10.1016/j.saa.2022.121785
摘要
Abstract Eating repeatedly used hotpot oil will cause serious harm to human health. In order to realize rapid non-destructive testing of hotpot oil quality, a modeling experiment method of fluorescence hyperspectral technology combined with machine learning algorithm was proposed. Five preprocessing algorithms are used to preprocess the original spectral data, which realizes data denoising and reduces the influence of baseline drift and tilt. The characteristic bands extracted from the spectral data show that the best characteristic bands for the two-classification model and the six-classification model are concentrated between 469–962 nm and 534–809 nm, respectively. Using the PCA algorithm to visualize the spectral data, the results show the distribution of the six types of samples intuitively, and indicate that the data can be classified. Based on the modeling analysis of the characteristic bands, the results show that the best two-classification models and the best six-classification models are MF-RF-RF and MF-XGBoost-LGB models, respectively, and the classification accuracy reaches 100%. Compared with the traditional model, the error is greatly reduced, and the calculation time is also saved. This study confirms that the use of fluorescence hyperspectral technology combined with machine learning algorithms provides a new theory and method for hotpot oil quality detection.
科研通智能强力驱动
Strongly Powered by AbleSci AI