Interference Image Processing Algorithm In Measurement Of Surface Topography

Modern optical interferometry combines the modern science and technology such as photoelectric detection, computer technology and image processing technology. Optical interferometry is developing rapidly, but the physical properties of the measurement are characterized by interference images. Therefore, optical interference image processing is extremely important. Optical interference image processing algorithm is proposed for higher requirements because of the emerging of optical interferometry and new technology. This paper focuses on the interference image processing algorithm, It studied the phase extraction algorithm, phase identification algorithm and surface 3D reconstruction of the interference images in the surface topography measurement. The main contents of this paper include:1、Starting from the basic principle of interference, the basic theory and phase extraction algorithm of the interference are studied. Based on this, a phase extraction algorithm based on least square ellipse fitting is proposed. Under the condition of open loop control of PZT to realize phase shift, only 40 single wavelength interference images are acquired in one cycle, it can get the driving phase sequence and driving step sequence of PZT by using ellipse fitting technology; on this basis, the initial phase of any point on the image was calculated with a certain accuracy.2、Based on the phase extraction, a new image data processing algorithm based on phase difference is proposed, which can effectively expand the range of measurement range of the morphology height. The algorithm uses the phase difference measurement scale and accuracy to gradually correct and refine phase data, finally, the highest accuracy of the single wave looks is used to calculate the morphology height, in the expansion of the high measurement range, it guarantees the final phase identification accuracy.3、On the basis of phase extraction and recognition, the 3D morphology of the local surface of the template is restored, and the 3D morphology of the surface is reconstructed by the image mosaic technology. Based on the research of spatial filtering algorithm and the noise feature of the reconstructed morphology, a threshold median filtering method is proposed, which removes the edge noise in the reconstructed morphology, and preserves the original valid height data maximally. Finally, the surface 3D morphology of the template is evaluated and error analysis is carried out.