花生冻伤近红外光谱快速判别方法研究

Study on fast discrimination of peanut frostbite by near-infrared spectroscopy

中文关键词:  花生  冻伤  近红外光谱  特征波长选取  支持向量机

英文关键词:peanut  frostbite  near infrared spectroscopy  characteristic wavelength selection  support vector machine

中文摘要:

      目的 构建基于近红外光谱快速判别花生冻伤的模型。方法 采用移动窗口平均平滑(moving window average, WMA)、标准正态变量校正(standard normal variate correction, SNV)及一阶导数(first derivative, FD)的组合预处理方法提升光谱信号质量; 分别采用无信息变量消除法(elimination of uninformative variables, UVE)、竞争性自适应重加权法(competitive adaptive reweighted sampling, CARS)以及二者的联合算法筛选特征波长; 最后构建基于支持向量机分类算法(support vector machine classification, SVC)的花生冻伤分类模型。结果 使用UVE-CARS算法筛选特征波长效果最佳, 筛选出7个特征波长, 构建的判别模型准确率达95%。结论 该花生冻伤判别模型判别准确率高, 为花生冻伤快速、无损判别提供可行的技术方案, 并为基于滤光片式近红外技术的花生品质色选机的开发提供参考。

英文摘要:

      Objective To establish a model for rapid identification of peanut frostbite based on near infrared spectroscopy. Methods Moving window average (WMA), standard normal variate correction (SNV) and first derivative (FD) were used to improve the quality of spectral signal; elimination of uninformative variables (UVE), competitive adaptive reweighted sampling (CARS) and their combined algorithms were used to screen characteristic wavelengths; then, a peanut frostbite classification model based on the support vector machine classification algorithm (SVC) was constructed. Results UVE-CARS algorithm was the best method to filter the feature wavelength, 7 feature wavelengths were selected, and the accuracy of the model was 95%. Conclusion The identification accuracy of this model is high, which provides a feasible technical scheme for the rapid and non-destructive identification of peanut frostbite, and provides a reference for the development of peanut quality color sorter based on the filter near-infrared technology.