Rice Identification and Classification System Based on SVM Algorithm

Abstract

Based on the spectral sensing and analysis technology, we designed and developed a rapid identification system for rice varieties based on the spectral characteristics analysis of rice of different varieties and origins, and integrated original parts such as light source, miniature spectral sensor, hardware development board, printer and display. The spectral data were filtered and classified using a support vector machine (SVM, support vector machine) classifier, and the recognition accuracy reached 98.41% when using the portable spectrometer in the device. The system includes two parts: software algorithm design and hardware device development. The final system realized the acquisition, transmission, processing and storage of spectral data. The operator interface of the rice type identification device was also designed and developed for real-time display of the spectral data measured by the microspectrometer and the rice type identification results.