南果梨果实硬度近红外无损检测模型的建立
Establishment of firmness value nondestructive detecting models on ‘Nanguo’ pears by NIRS technology
-
摘要: 实验比较了不同分区波段建模模型的性能,结果表明全波段643.26~954.15nm光谱建立的硬度模型优于其它两个分段光谱建立的模型;利用剔除异常样品和主成分分布图法对模型进行优化,建立了鞍山、海城两产区混合的硬度模型,相关系数(R)为0.970,均方根误差(RMSEC)为0.124;预测残差分布结果表明所建立的南果梨硬度模型性能较稳定,满足实际应用要求,模型适用于预测范围在2~15kg/cm2的南果梨硬度。Abstract: By comparing the models that were established by different spectra band, the results showed that the firmness model establishing under the 643.26~954.15nm spectrum was superior to the other two models. Methods of eliminating unusual samples and main components analyze were used to optimize the models. The firmness model results:correlation coefficient of calibration and RMSEC were 0.970 and 0.124. The prediction residual distribution results indicated that the models were stable and fulfilled the requirements of practical application, the optimal predictive scope of firmness model was 2~15kg/cm2.