Research on the Method of Hyperspectral and Image Deep Features for Bacon Classification

The deep learning method for hyperspectral image and spectral curves image features is proposed in this paper. The convolutional neural network (CNN) model has demonstrated as a universal representation for image feature extraction, especially in the application of image classification. In this paper, deep features of hyperspectral images are extracted using CNN, and the cross entropy is used as the optimization target. Then, the features of spectral curves are used into the image features. Finally, the fusion features are considered as input data, and classified by support vector machine (SVM), which realizes the classification and retrieval of meat. The method is getting our Bacon classification accuracy rate to 99.2%, The experimental results show the feasibility and effectiveness of the proposed method.