Research On The Key Technologies And Measuring System For Large-scale Microstructure Surface Topography With White Light Interference

Surface topography is an important quality feature of manufactured products and scientific samples,which directly affects their functional characteristics.The measurement of surface topography is of great significance to the quality analysis and function optimization.Some special surfaces,such as IC surface,MEMS surface,and microarray surface are micro in scale,but macro in structure and function.With characteristics of surface discontinuity and high precision,these surfaces require measurement with large-scale,high precision,and high efficiency.Based on white light interference measurement,key techniques and system for large-scale structured surface measurement are researched and developed in this thesis.A novel white light interferometry that combines the continuous rapid vertical scanning and synchronous external triggering photographing is put forward.This solves the low efficiency problem in the traditional sampling method step scanning and static photographing.A white light interference signal model corresponding to the novel sampling method is established,which provides theoretical foundation for surface topography recovery algorithm.A surface topography recovery algorithm is proposed based on the envelope matching and phase discrimination of interference signals.The relationship between the envelope correlation of interference signals and the relative height difference is analyzed.The criterion of matching phases in the same period is proposed under the optimal correlation position of the envelope.The calculating method of the surface height from multi-period wrapped phases is established.This algorithm fundamentally changes the traditional method which uses the aplanatic position as the absolute datum,and can effectively realize accurate surface topography recovery with reduced signal contrast due to dynamic acquisition.A 3D image stitching method based on an interferometric quality evaluation model is proposed to eliminate the negative effects of image noises.Firstly,an interference quality evaluation model for classification and definition of the effective heights and noises is established.Then,a quality matrix representing the effective point distribution is derived.Finally quality matrix is used to optimize the algorithms for feature extraction,correlation matching,and height fusion.This successively eliminates noise-caused stitching errors and improves the stitching accuracy.An evaluation method of microstructure is proposed to achieve efficient and accurate evaluation for large-scale microstructure surfaces based on maximum information entropy feature extraction.Firstly,effective points set and sensitive boundary points set for surface evaluation are extracted using interference quality evaluation model to shield image noises.Secondly,the distribution areas of effective features for extraction are categorized according to the gradient and height attributes.Thirdly,the threshold of each feature area is determined by the entropy value judgment condition of the maximum information entropy.Fourthly,the evaluation datum of the extracted features is used as the ideal element for the sensitive boundary points.Finally,large-scale surface feature extraction is obtained based on the probability of the distance between the boundary points and the ideal element.Based on the research of key technologies and the application of mature engineering methods,a white light interference measuring system is constructed,and a large number of testing experiments are carried out to verify its accuracy,rapidity and stability according to the national standards for scientific instrument verification.