基于B样条的TERS形貌成像背景扣除方法

在针尖增强拉曼光谱(TERS)形貌成像过程中，由于针尖与扫描台无法绝对平行、样品电子密度骤变处针尖快速升降以及扫描控制系统响应时间特性差等综合原因的影响，往往使形貌图中带有倾斜或边界面卷曲的成像背景。成像背景对样品形貌的识别和分析带来十分不利的影响，而背景扣除就是解决该问题的重要手段，也是形貌成像预处理的重要组成部分。背景扣除的原理一般是通过拟合背景的方法来扣除成像中的背景。传统的背景扣除方法是利用多项式拟合的方法对成像进行逐行的基线校正，但是该方法在处理形貌成像时常常会由于过拟合而造成样品形貌的失真，同时容易在图片上留下明显的线条纹理。针对传统方法的缺点，本文提出采用B样条曲面拟合方法，直接对样品形貌图进行曲面背景拟合，发挥B样条低阶光滑的优点，能够有效克服传统方法的缺陷。在实验中，同时利用传统方法和该方法对金单晶和合成金片的形貌图进行背景扣除，实验结果表明，两种方法都能够扣除样品形貌图中的成像背景，但与传统方法相比，所提出的方法不会造成样品形貌的失真，且不会留下线条纹理，获得了更加良好的背景扣除效果，为进一步分析样品形貌特征提供了更准确可靠的信息，是一种更加有效的TERS形貌成像背景扣除算法。

Background Subtraction Method of TERS Topography Imaging Based on B-Spline

In the topography imaging of Tip-Enhanced Raman Spectroscopy, the difficulty of keeping the tip and the scan stage absolutely parallel, the fast change of the height of the tip due to the quick variety of the electron density, and the long response time of the scanning control system result in imaging background with inclined or boundary curly. Imaging background would bring a negative impact to the identification and analysis of the sample’s topographic image，while background subtraction is an important means to solve this problem, which is also an important part of topographic image’s preprocessing. The general principle of background subtraction is to fit the topographic image’s background firstly and then subtract it from the image. The traditional fitting method is polynomial fitting line by line, but this method will always destroy the sample’s actual spatial topography by over-fitting, and it is prone to leave clear line texture on the image. In this paper, the traditional method was improved, and the B-spline surface fitting method was used to approach the background of topographic image through constant iteration. The advantage of B-spline surface, namely low-order and smoothness, can help the method overcome the disadvantages of traditional polynomial method. In the experiments, the topographic image of Au (111) single-crystal and synthetic golden slice were detected, and then the proposed method and the traditional method were applied to perform background subtraction. Experimental results showed that the proposed method can eliminate the topographic image’s background effectively without destroying the sample’s actual spatial topography by over-fitting and leaving line texture on the image. Therefore, the proposed method provided more accurate and reliable information for the further analysis of topographic images, and it’s a more effective background subtraction method of TERS topography imaging.