STM Topography Image Processing And Software Integration Application

Scanning tunneling microscope(STM)is an atomic-level surface imaging instrument,which has a wide range of applications in various disciplines due to its ultrahigh spatial resolution at the nanometer level.Its application in TERS technology combines ultra-high spatial resolution at the nanometer scale with the chemical recognition capability of Raman spectroscopy,and has become an emerging nanoscale surface analysis technology.In order to better perform surface analysis,we hope that the obtained STM topography image is of high quality,complete,free from noise and defects.However,the scanning tunnel microscope is a very precise system.Only slight external vibration or noise interference and the hysteresis of the control system itself will cause defects in the topography image,so that the topography features of the sample itself are masked.In this paper,based on the analysis of the current status of STM topography image processing,in order to solve the problem of STM topography image repair,it is necessary to manually identify the type of defect and manually locate the defect location.Periodic noise and other defects propose a fully automatic image repair method that combines deep learning with traditional image processing.This method first uses deep convolutional neural networks to automatically identify the types of defects in the STM shape image to be repaired,and then Automatically locate and repair defects according to the defect types identified by the network.The automatic STM shape image defect repair algorithm proposed in this paper can achieve automatic repair of STM image defects by self-identification,self-positioning and self-processing.The entire repair process does not require any manual intervention.In terms of software integration and application,this paper intends to build a comprehensive TERS cloud service system with comprehensive functions,convenient interaction,openness and scalability.The entire software system includes STM instrument control,STM data post-processing,Raman instrument control,Raman data post-processing,laser spot position monitoring,TERS joint acquisition and joint analysis and other functions.In terms of architecture,the system uses a browser/server model and a separate front-end and back-end architecture,breaking the traditional single-machine application model.In addition,the system cloudizes the atlas data postprocessing and analysis functions,making the atlas data post-processing not limited to the instrument itself,making it more convenient and versatile.At the same time,the system adopts a Docker containerized deployment solution,so that services with different language architectures can be conveniently deployed without being restricted by specific machines and operating environments.