中国草食动物科学2025，Vol.45Issue(3)：62-68,7.DOI:10.3969/j.issn.2095-3887.2025.00.017
全株玉米青贮营养成分含量近红外光谱预测模型的建立
Establishment of Near Infrared Spectroscopy Prediction Model for Nutrient Content of Whole Corn Silage
摘要
Abstract
The study aimed to use near infrared reflectance spectroscopy(NIRS)to establish a near-infrared prediction model for the nutrient content of whole corn silage,and to provide theoretical basis and technical support for more effi-cient and reasonable utilization of corn silage in actual production.A total of 190 corn silages were collected from 5 provinces in northwestern China to measure contents of DM,Ash,CP,NDF,ADF,EE and starch.According to the Kennard-stone method,all samples were randomly divided into a calibration set and a validation set at a ratio of 4∶1.Abnormal spectra were removed through principal component analysis,and prediction models for each nutrient compo-nent were constructed using mathematical and stoichiometry methods,followed by external validation.The results indi-cated that the 1-VR and RSQ of ADF content prediction model were 0.946 7 and 0.90,respectively,which could be used for accurate analysis in the whole corn silage.The 1-VR of DM,CP and Ash component prediction models were 0.747 0,0.634 4 and 0.777 3,and the RSQ were 0.71,0.68 and 0.70 respectively,which could be used for prediction in actual production.The 1-VR and RSQ of NDF content prediction model were 0.289 4 and 0.39,respectively,indicat-ing low prediction accuracy,and needing further optimization.The 1-VR values of EE and starr prediction models were 0.181 0 and 0.170 5,respectively,indicating that prediction models were unavailable.关键词
全株玉米青贮/营养成分/近红外预测模型Key words
whole corn silage/nutritional component/prediction model分类
畜牧业引用本文复制引用
潘发明,郭涛,李飞,郝生燕,刘佳..全株玉米青贮营养成分含量近红外光谱预测模型的建立[J].中国草食动物科学,2025,45(3):62-68,7.基金项目
甘肃省科技重大专项计划项目(25ZDNG001) （25ZDNG001）
甘肃省科技计划项目(21CX1NA363) （21CX1NA363）
甘肃省林业和草原科技创新项目(GYCX2020032) （GYCX2020032）
