中文摘要：

用光学显微干涉法进行表面形貌测量时其深度测量范围的扩大和形貌测量精度的提高是一对矛盾。为此,本文设计出了一种基于波长轮换与相移扫描相结合的三波长表面形貌测量系统,并提出了一种基于椭圆拟合与相位差大小尺度相结合的相位提取与识别算法。将这种算法运用于多波长干涉图像的数据处理,有效地提高了形貌的整体测量精度,并拓展了深度测量范围。实验结果表明：在深度测量范围扩大近15倍的条件下,采用粗糙度国家基准校准的方波多刻线样板得到的表面粗糙度数据与校准数据的相对误差仅为4.12%,表明该系统在一定的深度范围内能够实现表面形貌的高精度测量。另外,针对该系统设计的多波长相位识别算法对环境噪声要求不高,可以支持系统的高噪声或在线测量。

英文摘要：

When optical microscopic interferometry is used to measure the surface topography, there is a contradiction between the enlargement of measuring interval and the improvement of measuring precision. Therefore, a three-wavelength interference measurement system for surface topography was designed based on wavelength switching and phase shift scanning. A data processing method for three wavelength interference images was proposed by using the phase extraction and recognition algorithm based on elliptic fitting and combined size scales of phase difference. The method was used to process multi-wavelength interformetric images, and it effectively improves the overall precision of surface topographic measurement and expands the measuring interval. Experimental results show that when the measurement interval is expanded to nearly 15 times, the relative measurement error of surface roughness of square wave specimen with multiple grooves is only 4.12% as compared to that of the data calibrated by China Metrology Institute. It concludes that the three wavelength interferometry realizes a high precision measurement for surface topography in a certain range. Moreover, the data processing method for multi-wavelength phase recognition has not higher demands for environment noise and can support system measurement in a higher noise environment in real time.