摘要
通过将扫描电镜(SEM)所有光学系统调到同轴,选择合适的加速电压及工作距离,同时消除图像的像散,得到分辨率和衬度都适中的SEM图像。以小波分析为基础,对SEM图像进行分析与处理。设计模极大值去噪、阈值去噪及相关性去噪3种去噪方式,对SEM原始图像进行去噪处理,并分别计算峰值信噪比、均方根误差、归一化相关系数及运算时间4种参数,对比去噪后图像与原始图像的相似程度,选择相关性去噪方法为最优去噪方法。利用小波塔式分解将去噪后的SEM图像进行特征增强处理。最后用Maltab2015a对所有结果进行了仿真对比,证明该方法不仅确保SEM图像的质量,而且为后续的图像分析奠定良好的基础。
The quality of scanning electron microscope(SEM)image plays a decisive role in the study of material microstructure,phase structure and crystal structure.By adjusting all optical systems of SEM to be coaxial,selecting appropriate accelerating voltage and working distance,and eliminating image astigmatism,SEM image with moderate resolution and contrast is obtained.SEMimageis analyzed and processed by wavelet analysis.Three denoising methods of modulus maximum denoising,threshold denoising and correlation denoising are designed and the four parameters of peak signal to noise ratio,mean square error,normalized correlation and operation time are calculated,respectively,to compare the similarity between the denoised image and the original imageand select the correlation denoising method as the optimal denoising method.The denoised SEM image is enhanced by wavelet tower decomposition.Finally,all the results are simulated and compared with Maltab2015a.It is proved that this method not only ensures the quality of SEM image,but also lays a good foundation for subsequent image analysis.
作者
李晓瑜
LI Xiaoyu(Key Laboratory of Electromagnetic Processing of Materials,Ministry of Education,Northeastern University,Shenyang 110004,China)
出处
《实验室研究与探索》
CAS
北大核心
2022年第5期26-29,共4页
Research and Exploration In Laboratory
基金
国家自然科学基金面上项目(62073064)
辽宁省自然科学基金面上项目(2021-MS-084)。
关键词
小波分析
扫描电镜
图像去噪
图像增强
wavelet analysis
scanning electron microscope(SEM)
image denoising
image enhancement
