| Summary: | This study proposes a new method to distinguish between the different qualities of dry alfalfa. This method uses scanning electron microscopy (SEM) to obtain grayscale images, and. then, using an equalized histogram, the gray-level co-occurrence matrix (GLCM) extracts 14 texture features. The texture feature vector is processed by principal component analysis (PCA) and linear discriminant analysis (LDA) to reduce data redundancy and extract the best features. Finally, a back propagation neural network (BPNN) algorithm, an artificial neural network (ANN) and a least square support vector machine (LS-SVM) classification model are established to evaluate the classification effect. The results show that LDA is more effective in transforming the original data. In addition, LDA-based classification results are better than PCA-based classification results, and the recognition rate is 100% accurate. In contrast, the reliability and potential of extracting the main information based on LDA are shown. Based on these conclusions, it is possible to identify various types of alfalfa after different drying methods.
|
|---|