Cigar tobacco leaf air-curing process stage identification method based on incremental learning

Abstract

The invention belongs to the technical field of tobacco leaf air-curing, discloses a cigar tobacco leaf air-curing process stage identification method based on incremental learning, and provides a fusion model for identifying a tobacco leaf air-curing process stage by adopting an incremental learning mode based on an SGD logic classification algorithm. According to the model, data can be subjected to preprocessing and feature selection in combination with airing data characteristics collected by an airing room, and through an incremental training learning mode, the accuracy of judging the airing process stage is gradually improved, the tobacco airing process is optimized, the working pressure of tobacco growers is relieved, and the economic benefits of tobacco are improved. According to the method, a large amount of effective information is filled for tobacco leaf data features, the subsequent model prediction accuracy is improved, the problem of a large amount of noise in air-curing data is solved, and the model training efficiency is improved; the air-curing process stage is rapidly judged in real time, and meanwhile, the model learns again by utilizing subsequent data increment, so that the tobacco air-curing process is improved, and remote, intelligent and accurate air-curing is realized.

Description

technical field [0001] The invention belongs to the technical field of air-curing tobacco leaves, and in particular relates to a process stage identification method for air-curing cigar tobacco leaves based on incremental learning. Background technique [0002] At present, air-curing of cigar tobacco leaves refers to the process of gradually drying the harvested tobacco leaves and gradually changing the chemical components of the tobacco leaves. It is an important process for the formation of the appearance and internal quality of cigar tobacco leaves. From a physiological point of view, air curing is mainly a metabolic process of starvation, a process in which dehydration and drying of tobacco leaves and changes in internal chemical substances are coordinated. Among them, as the part with the highest unit price in cigars, the prepared wrapper is dark brown and uniform in color, the tobacco leaves are thin, have sufficient oil content, good elasticity, thin and flat veins, a...

AI Technical Summary

Problems solved by technology

[0006] (1) The existing tobacco leaf image processing method in the air-curing process of tobacco leaves imports a variety of data characteristics into the model, and the number of noise and data anomalies also increases, resulting in poor data quality

[0007] (2) The existing research on the identification of tobacco leaf drying process all use one-time learning and training algorithms. These algorithms perform generally when the data is noisy, and it is difficult to apply them to actual production.

[0009] At present, due to the problem of data and modeling ideas, the one-time learning and training algorithm performs generally in the data set with complicated data, and the performance of the model is limited.

Images