AFM virtual system imaging prediction method based on VGG16

• Main Authors: XU HONGMEI, WANG ZUOBIN, SONG JUNG-HUN, LIU XUWEI, YANG JINXIN, MA DA, DONG LITONG
• Format: Patent
• Language: Chinese; English
• Published: 15.03.2024
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention belongs to the technical field of convolutional neural network image prediction and virtual simulation, and particularly relates to a VGG16-based AFM virtual system imaging prediction method. According to the method, the pre-trained VGG16 model is used as a feature extractor, the model structure is optimized by integrating a residual encoder and batch normalization, and meanwhile, the training efficiency is improved by using an Adam optimizer. A vibration equation of a probe micro-cantilever beam in a tapping mode is established according to an Euler-Bernoulli beam theory, a finite element method is used for simulating vibration response in Unity 3D, a virtual sample is placed in a virtual AFM for scanning to obtain four-quadrant light spot data, the four-quadrant light spot data is used as an input quantity of a prediction model, information of U and V channels is predicted, and finally, the four-quadrant light spot data is obtained. And finally, the model prediction accuracy is evaluated by us