不同气候生态型籼稻糙米粗蛋白含量光谱估测模型研究  

Spectral Estimation Model of Unpolished Rice Protein Content in Indica Rice with Different Climate Ecotypes

作　　者：田容才 周昆[2] 高志强[1] 卢俊玮 Tian Rongcai;Zhou Kun;Gao Zhiqiang;Lu Junwei(Agronomy College of Hunan Agricultural University, Changsha 410128;Hunan Rice Research Institute, Hunan Academy of Agricultural Sciences, Changsha 410125)

机构地区：[1]湖南农业大学农学院,长沙410128 [2]湖南省农业科学院水稻研究所,长沙410125

出　　处：《中国粮油学报》2021年第11期162-169,共8页Journal of the Chinese Cereals and Oils Association

基　　金：国家重点研发计划(2017YFD0301506)。

摘　　要：为探究气候生态型存在差异的籼稻品种光谱响应规律及其对粗蛋白含量估测模型的影响,本研究利用2019年长江中下游籼稻联合区实验,采集了不同气候生态型籼稻籽粒反射光谱及糙米粗蛋白含量数据,分析了稻谷原始及一阶微分光谱与糙米粗蛋白含量的相关关系,建立了基于最优光谱指数、全波长和特征波长的籼稻粗蛋白含量的PLSR、PCR和SMLR估测模型,并用R2、RMSE评价模型精度。研究发现,籼稻籽粒光谱反射率随着粗蛋白含量的升高而降低,呈现出中籼稻>晚籼稻>早籼稻的规律;在以原始及一阶微分任意两波长构建的DSI、NDSI和RSI最优光谱指数模型中,PLSR模型效果较好,建模集R2C、RMSEC分别为0.841、0.507%,验证集R2V、RMSEV分别为0.810、0.542%;在全波长模型中,建模效果表现为PLSR>SMLR>PCR,以原始光谱建立的PLSR模型效果最好,建模集R2C、RMSEC分别为0.867、0.464%,验证集R2V、RMSEV分别为0.856、0.472%;特征波长模型中,一阶微分构建的模型优于原始光谱,尤以基于一阶微分光谱建立的PLSR模型稳定性更好,建模集R2C、RMSEC分别为0.842、0.506%,验证集R2V、RMSEV分别为0.823、0.523%。结果表明PLSR模型适用于不同气候生态型籼稻品种间的粗蛋白含量光谱估测。To explore the spectral response law of indica rice varieties with different climate ecotypes and its influence on the estimation model of crude protein content,we used the trail of indica rice joint area in the middle and lower reaches of the Yangtze River in 2019.We collected the grain spectral reflectance and the crude protein content data of unpolished rice that with different climate and ecological types,and analyzed the relationships between the original and first derivative spectral reflectance of indica rice and the corresponding crude protein content in the unpolished rice.The PLSR,PCR,SMLR estimation model based on the optimal spectral index,full wavelengths and characteristic wavelengths of the grain crude protein content of indica rice were established,and the determination coefficient(R2)and root mean square error(RMSE)were used to evaluate the accuracy of the models.The results showed that the spectral reflectance decreased with the increase of crude protein content,roughly showing the rule of medium indica>late indica>early indica.In the DSI,NDSI and RSI optimal spectral index models that constructed with any two wavelengths of original and first derivative reflectance,the PLSR model had a good effect.The calibration set R2c and RMSEc were 0.841 and 0.507%,and the validation set R2v and RMSEv were 0.81 and 0.542%,respectively.Among the full-wavelength model,the effect was shown as PLSR>SMLR>PCR.The PLSR model established based on the original spectrum had the best effect,with the calibration set R2c and RMSEc were 0.867 and 0.464%,and the validation set R2v and RMSEv were 0.856 and 0.472%,respectively.During the characteristic wavelength models,the model constructed by the first-order differential was better than the original spectrum,especially the PLSR model based on the first derivative spectral reflectance was more stable.The training set R2c and RMSEc were 0.842 and 0.506%,and the testing set R2v and RMSEv were 0.823 and 0.523%,respectively.It indicated that the PLSR model was suitable for spe

关 键 词：籼稻 粗蛋白含量 光谱 生态型 偏最小二乘回归 

分 类 号：TS210[轻工技术与工程—粮食、油脂及植物蛋白工程]