基于数学形态学的表面原子熔融相的STM图像识别算法  

A mathematical-morphology-based algorithm for identifying surface atomic molten phases

作　　者：樊甫江 时国栋 吴升清 王中寅 桂飞林 商波[3] 魏征 FAN Fu-Jiang;SHI Guo-Dong;WU Sheng-Qing;WANG Zhong-Yin;GUI Fei-Lin;SHANG Bo;WEI Zheng(College of Materials Science and Engineering,Chongqing University,Chongqin 400045,China;College of Mechanical and Vehicle Engineering,Chongqing University,Chongqing 400045,China;School of Chemistry and Chemical Engineering,Chongqing University,Chongqing 400045,China)

机构地区：[1]重庆大学材料科学与工程学院,重庆400044 [2]重庆大学机械与运载工程学院,重庆400044 [3]重庆大学化学与化工学院,重庆400044

出　　处：《原子与分子物理学报》2021年第6期15-20,共6页Journal of Atomic and Molecular Physics

基　　金：国家自然科学基金(11774039);重庆市大学生创新创业训练计划资助项目(S201910611305)。

摘　　要：高温扫描隧道显微镜(HT-STM)可以实时原位地捕捉到表面原子的熔化相变过程.在这一原位变温实验中,快速可靠地识别出每帧STM图像中的熔融相十分关键.传统的手工统计方法存在效率低下、随意性大等问题.我们发展出一套基于数学形态学的算法,来自动快速地识别.与人工方法相比,该算法消除了人为主观误差,使确定的边界更加准确、光滑,处理效率提高了266倍.High-temperature scanning tunneling microscope(HT-STM) can capture the melting transition of surface atom in real-time. In such in-situ HT-STM experiments, it is crucial to reliably identify the molten phase region in HT-STM images as fast as possible. The traditional manual counting method is inefficient and very casual. We developed an automatic image processing method based on mathematical-morphology, to overcome these problems. It is 266 times faster than the manual method, eliminates the possible human errors, and identifies the molten regions’ boundaries more accurately and smoothly.

关 键 词：图像处理 STM图像 数学形态学 阈值分割 

分 类 号：O469[理学—凝聚态物理]