Detection of Isoflavones Content in Soybean Based on Hyperspectral Imaging Technology

In this paper, the authors propose a hyperspectral imaging technology to detect the contents of soybean isoflavones. Based on the 40 varieties of soybeans produced in Heilongjiang province, they get the spectral reflection datum of soybean samples varied from the soybean's hyperspectral images which are collected by the hyperspectral imaging system, and apply high performance liquid chromatography method to determine the true value of the selected samples of isoflavones. The feature wavelengths for isoflavones content prediction were selected based on correlation analysis. The prediction model was established by using the method of BP neural network in order to realize the prediction of soybean isoflavones content analysis. The experimental results show that the ANN model could predict isoflavones content of soybean samples with of 0.9679, the average relative error is 0.8032%, and the mean square error is 0.110328, which indicates the effectiveness of the proposed method and provides a theoretical basis for the applications of hyerspectral imaging in non-destructive detection for interior quality of soybean.