Abstract:In the era of artificial intelligence, the development of domain-specific intelligent agents for rice will further enhance the intelligent analysis level of its quality inspection data processing. Targeting the need for rice origin identification, this research integrates Raman spectral characterization with support vector machine learning algorithms, connects to the iFlytek Spark general-purpose large model to provide knowledge question-answering services, adopts the Flask framework to develop a unified front-end interface, and integrates the identification and question-answering functional modules in the form of tabs. It can achieve accurate identification of rice origins while providing knowledge question-answering services in aspects such as rice culture, characteristics, and identification. This solution realizes an automated and intelligent processing path for the task of rice origin identification. It only requires collecting Raman spectral data of samples and inputting them into the intelligent agent, which then invokes the model to complete subsequent analysis tasks and output the results.