Modeling method of tobacco leaf conventional chemical component general model based on near infrared spectrum

• Main Authors: MU JIRUI, LI YONGSHENG, PENG YUHAN, LI XIAORUN, LIAO FU, HE WENMIAO, SHI YIDAN, DAI LU, HUANG HUI, ZHANG LILI, WU JIZHONG, LI SHITOU, WANG HUI
• Format: Patent
• Language: Chinese; English
• Published: 08.04.2022
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIRCHEMICAL OR PHYSICAL PROPERTIES; MEASURING; PHYSICS; TESTING

The invention provides a modeling method of a tobacco leaf conventional chemical component general model based on a near infrared spectrum. The method comprises the following steps: collecting near infrared spectrum data and producing area information of a tobacco leaf sample under the same condition, and measuring the content of conventional chemical components; and establishing a tobacco leaf producing area prediction model. Secondly, setting a group of expected model target evaluation parameters, carrying out partial least square regression analysis modeling by utilizing sample chemical component content data and preprocessed spectral data, and if the obtained model does not meet the requirements, rejecting the samples participating in modeling by taking the production places as units to obtain a main model which meets the requirements and is suitable for the most production places; for samples which do not participate in modeling of the main model, a model transfer algorithm is used for establishing trans