| As an indispensable part of grassland resources,forage grass plays a very important role in the development of contemporary animal husbandry and environmental protection.High quality,high yield forage seeds varieties with high stress resistance are very challenging,so it is crucial to accelerate research on the identification and selection of high-quality forage seeds varieties.This paper uses terahertz time-domain spectroscopy(THz-TDS)to conduct two researches on forage seeds species with high yield and strong adaptability.Firstly,based on the experimental measurement of terahertz(THz)spectrum,the machine learning algorithm was used to study the differences between sadawang seeds species.It has enriched and improved forage species and identification methods for spectral identification of forage seeds by THz-TDS.More importantly,the relationship between the biological components of forage seeds and the transmission characteristics of THz waves was explored from the theoretical mechanism,and a multi-layer structure transmission model(MSP)was innovatively established to combine the differences of forage seeds with its optical properties and biological properties.Then the model is justified by experiments.The main research contents are as follows:(1)Based on terahertz time-domain spectroscopy technology combined with machine learning algorithm,the accurate identification of sadawang seeds species was successfully realized.On the basis of the terahertz spectroscopy experiment,the random forest model was established after processing the spectral experimental data of the samples.Next,in order to further improve the accuracy of the model,a principal component-random forest calculation model was proposed to measure five sadawang seeds species.The results show that the classification accuracy of the model can reach 91.20%.The above research provides a more accurate method for the rapid and effective identification of sadawang species and other similar forage species,and also further improves and develops the identification scope and method of using terahertz time-domain spectroscopy to study forage species and interspecific differences.(2)Based on the experimental measurement results of THz and the effective medium theory,a multi-layer structure transmission model was established.The effects of forage seeds species components on the transmission characteristics of THz waves were theoretically expounded,and an MSP theoretical model based on forage seeds species biological components and optical parameters was constructed.Then,the rationality and accuracy of the MSP model were successfully verified by taking sadawang seeds species and alfalfa seeds species as the research objects.(3)In order to further verify the accuracy of the model,the MSP model was applied to the identification of drought resistance and cold resistance of alfalfa seeds species.The results showed that,on the one hand,the MSP model could well verify the experimental results,with an accuracy rate of 94.433%.On the other hand,the calculation values ratio of protein and carbohydrate components based on the MSP model could effectively distinguish the cold tolerance and drought tolerance of alfalfa seeds..To sum up,this study firstly started with the identification of five varieties of sadawang,and explored the spectral differences of different varieties in the THz band,and combined with machine learning algorithms,in-depth quantitative analysis of the small differences between varieties.Secondly,in order to to further explore the reasons for the differences in the optical properties of grass species,this paper starts from the theory and establishes the propagation equation of THz waves in forage seeds species based on the information of seeds species biological components.The above work provides theoretical guidance for further research on high-quality forage seeds,and lays an experimental foundation and theoretical guidance method for the construction of high-quality forage seeds resource banks in my country. |
PDF Full Text Request