Study On Morphological Feature Extraction And Geometric Modeling Of Rough Surfaces

The evaluation and analysis of the rough surface geometry is one of the important research topics in the field of tribology and surface engineering.Influenced by a variety of deterministic and random factors,the surface of any parts will form a unique geometric shape.The surface of real parts consists of a series of uneven micro-convex body.Contact stress,deformation,temperature,fatigue damage,adhesion,furrow and other effects on the rough surface have a direct relationship with the height,shape and spatial distribution of micro-bulge.The microscopic geometrical morphology of the rough surface affects the contact characteristics of the real surface,which has an impact on the properties of friction,sealing and lubrication of the whole machine.Study of the geometrical appearance characteristics of the rough surface and representation and modeling of it have important engineering significance.The main contents of this paper:20 samples of milling and turning surfaces were prepared under the same processing technology parameters.The geometrical morphology data of rough surfaces were measured by STIL Micromesure2,and the datum of the surface was filtered by wavelet transform.Angular second moment/energy,entropy,contrast ratio,uniformity and correlation of the samples were analyzed and compared by gray-level co-occurrence matrix method,and then the similarity of texture among surface under the same processing parameters was determined.Based on the two-dimensional discrete wavelet transform,the milling and turning specimen was decomposed by multi-scale and multi-directional decomposition,and then the statistical law of the wavelet coefficients was found and the surface modeling in different scales and directions was realized.The surface reconstruction of the milling specimen will be realized by using the two-dimensional discrete wavelet inverse transformation.The surface three-dimensional characterization parameters were used to verify it;In this paper,the influence of Contourlet transform on the multi-dimensional and multi-directional decomposition of the surface of the turning specimen was studied.By looking for the statistical law of the Contourlet coefficient,the surface reconstruction in different scales and directions was realized.Contourlet transform was used to realize the surface modeling at full scale,and characterization parameters was used to verify it.Finally,we study that how to make the surface after modeling have a given bearing rate curve.Research results:Through the feature parameters analysis of the gray level co-occurrence matrix,the results show that the geometrical morphology of the rough surface that obtained under the same machining method and machining parameters has high similarity in the high frequency and low frequency characteristics,which provides theoretical basis for computer modeling and following analysis of rough surface;' Through the milling test specimen modeling research,it found out that two-dimensional discrete wavelet transform can realize multi-scale decomposition of the rough surface and is beneficial to capture straight-line texture under the fixed direction of rough surface.In the meanwhile,there is high similarity in the before and after modeling surface by the three dimensional characterization parameter validation.Through the analysis of surface modeling process of turning sample,it indicates that Contourlet transform is more flexible in direction of decomposition and good at capturing the characteristics of the curve texture.By the three dimensional characterization parameter validation,the before and after modeling surface exists high similarity;There exists a simple algebraic relations between the rough surface support rate curve and the surface height cumulative probability distribution curve.Based on the normalization method of the histogram of the gray image,it is possible to establish rough surface model with a specified bearing rate Curve.