基于叶面特征的烤烟中部叶采收成熟度量化初探OA北大核心CSTPCD

[Objective]This study aims to promote the realization of digital judgment of flue-cured tobacco harvesting maturity.[Methods]746 tobacco samples collected from the central part of Yunnan Tobacco 87 were classified according to four grades:under-ripe,still-ripe,moderately ripe and over-ripe,and the fuzzy DEMATEL analysis method was used to assign weights to the leaf surface characteristics of flue-cured tobacco in three aspects:agronomic characteristics,color characteristics and texture characteristics.After that,the partial least squares discriminant(PLS-DA)method was used to simplify the quantification model of the harvesting maturity of the flue-cured tobacco,and the robustness of the model was verified.[Results]The results showed that:(1)the final score for the under-ripe grade was determined as[0.0,81.0),and the final score for the under-ripe grade was determined as[0.0,81.0).(2)The final scores of under-ripe grades were determined as[0.0,81.0),under-ripe grades as[81.0,85.0),under-ripe grades as[85.0,91.0],and over-ripe grades as(91.0,100];(3)The results of the optimized maturity recognition model had a correct rate of 89.04％.[Conclusion]The maturity of flue-cured tobacco was quantified by collecting the leaf surface characteristics of fresh tobacco leaves.