基于颜色分布信息的烟叶烟碱含量预测模型评估与比较

打开文本图片集

关键词：颜色分布；烟碱预测；回归模型； 近邻；无损检测中图分类号：TP391.4；TP181 文献标识码：A 文章编号：2096-4706（2025）08-0132-07

Abstract： This paper uses optical imaging technology to establish an image dataset of abatch oftobacco leaves with knownnicotinecontent，anduses theneuralnetworkmodelU2-Nettoaccuratelydetecttobaccoleaftargets.Byextractingthe color distributioninformationofthetobaccoleaf targets，four typicalMachineLeamingalgorithmsofRF，XGBoost，MP，and KNNareusedtomaketheregressonpredictionforthenicotinecontentof tobacco leaves，respectively.Theresults indicate that theKNNmodelcanefectivelyutilizecolordistributioninformationtoaccuratelypredictthenicotinecontentoftbaccoleaves. The value of determination coefficient is as high as 9 7 . 4 6 % ，theMSE isaslow as O.020 2，and the MAE is as low as 0.075 6，indicatingasigncantcorelationbetweentobaccoleafcolordistributioniformationandnicotinecontent，andprovidingan effective nondestructive detection method for nicotine.

Keywords： color distribution; nicotine prediction;regressionmodel; K-Nearest Neighbor;nondestructive testing

0 引言

烟草作为一种重要的农作物，广泛应用于烟草制品的生产，对于全球经济和人们的日常生活有着重要的影响。（剩余6969字）