| abstract |
The invention provides an optical characterization method, model and application of a single-layer molybdenum disulfide sample based on machine learning. First, the optical imaging of the molybdenum disulfide sample is used to extract the suspicious monolayer ROI area through graphic processing; then, the difference vector between the pixel value of the suspicious monolayer ROI local area and the eigenvalue photographed by the silicon wafer under the optical microscope is calculated, and the Raman characterization is used to distinguish single-layer and few-layer samples, the residual glue is determined by visual observation, and the target value is established according to the number of layers; the difference vector is averaged and the standard value is used as the feature value, and the data is composed with the target value. Finally, by reducing the dimensionality of the dataset and classifying the dataset through machine learning algorithms, the best single-layer representation model is obtained. Based on this model, single-layer molybdenum disulfide samples can be quickly distinguished by optical imaging, which greatly saves the time spent in finding single-layer molybdenum disulfide. |