Abstract: As the seed vigor was gradually brought to the public attention， the rapid detection of seed vigor without destroying has been a research hot spot. The rice seeds with different aging degrees were used as experimental materials， PCA-SVM method and hyperspectral imaging technology were compared to study the difference in rice seed vigor. The hyperspectral image data of two rice varieties were collected in 400-1 000 nm range. In order to obtain principal component images， PCA was used to determine the feature band. SVM was adopted to build the model of seed vigor. The results showed that the prediction accuracy could reach 100%. A new way of rapid detection of seed vigor was proposed by hyperspectral imaging technology.

Key words: seed vigor, hyperspectral, support vector machine, principal component analysis