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.
作者
樊甫江
时国栋
吴升清
王中寅
桂飞林
商波
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)
出处
《原子与分子物理学报》
CAS
北大核心
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)。
关键词
图像处理
STM图像
数学形态学
阈值分割
Image processing
STM image
Mathematical morphology
Threshold segmentation
