Texture-based segmentation with Gabor filters, wavelet and pyramid decompositions for extracting individual surface features from areal surface topography maps

Senin, Nicola; Leach, Richard; Pini, Stefano; Blunt, Liam

doi:10.1088/0957-0233/26/9/095405

Cited by 18 publications

(4 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, simple thresholding is not a stable method when surfaces have a stochastic content and can produce many insignificant features that can cause problems for many characterisation parameters, such as the number of defects and the density of features [ 55 ]. Alternatively, morphological segmentation into hills or dales is the only partitioning operation currently endorsed by the ISO specification standards on surface texture metrology [ 56 ]. More sophisticated thresholding techniques may allow determination of valuable results of complex surfaces.…”

Section: Methodsmentioning

confidence: 99%

Thresholding Methods for Reduction in Data Processing Errors in the Laser-Textured Surface Topography Measurements

Podulka

2022

Materials

View full text Add to dashboard Cite

There are many factors influencing the accuracy of surface topography measurement results: one of them is the vibrations caused by the high-frequency noise occurrence. It is extremely difficult to extract results defined as noise from the real measured data, especially the application of various methods requiring skilled users and, additionally, the improper use of software may cause errors in the data processing. Accordingly, various thresholding methods for the minimization of errors in the raw surface topography data processing were proposed and compared with commonly used (available in the commercial software) techniques. Applied procedures were used for the minimization of errors in the surface topography parameters (from ISO 25178 standard) calculation after the removal and reduction, respectively, of the high-frequency noise (S-filter). Methods were applied for analysis of the laser-textured surfaces with a comparison of many regular methods, proposed previously in the commercial measuring equipment. It was found that the application of commonly used algorithms can be suitable for the processing of the measured data when selected procedures are provided. Moreover, errors in both the measurement process and the data processing can be reduced when thresholding methods support regular algorithms and procedures. From applied, commonly used methods (regular Gaussian regression filter, robust Gaussian regression filter, spline filter and fast Fourier transform filter), the most encouraging results were obtained for high-frequency noise reduction in laser-textured details when the fast Fourier transform filter was supported by a thresholding approach.

show abstract

Section: Methodsmentioning

confidence: 99%

Thresholding Methods for Reduction in Data Processing Errors in the Laser-Textured Surface Topography Measurements

Podulka

2022

Materials

View full text Add to dashboard Cite

show abstract

“…The void regions created by spatter removal are then filled by weighted interpolation of valid neighbours [25]. The resulting topography is sub jected to a Gaussian pyramid decomposition [29,30] (four levels, dyadic decomposition, sigma of the Gaussian kernel: 2 pixels) to obtain a smoothed representation of the overall weld track shape. The fourth level of the pyramid decomposi tion is extracted and fitted to a smooth surface by weighted, nonparametric local regression to a second order polynomial surface (LOESS fitting [31]), as shown in figure 7(a).…”

Section: Identification and Characterisation Of The Weld Tracksmentioning

confidence: 99%

Feature-based characterisation of signature topography in laser powder bed fusion of metals

Senin

Thompson

Leach

2018

Meas. Sci. Technol.

Self Cite

View full text Add to dashboard Cite

The use of state-of-the-art areal topography measurement instrumentation allows for a high level of detail in the acquisition of topographic information at micrometric scales. The three-dimensional geometric models of surface topography obtained from measured data create new opportunities for the investigation of manufacturing processes through characterisation of the surfaces of manufactured parts. Conventional methods for quantitative assessment of topography usually only involve the computation of texture parameters; summary indicators of topographyrelated characteristics that are computed over the investigated area. However, further useful information may be obtained through characterisation of signature topographic formations, as more direct indicators of manufacturing process behaviour and performance. In this work, laser powder bed fusion of metals is considered. An original algorithmic method is proposed to isolate relevant topographic formations and to quantify their dimensional and geometric properties, using areal topography data acquired by state-of-the-art areal topography measurement instrumentation.

show abstract

“…Examples of models used for describing textures are Markov random field models [ 37 , 38 ] and fractal models [ 39 , 40 ]. Finally, filter-based methods, also known as transform-based methods, describe textures through frequency analysis based on Fourier transforms [ 41 , 42 ], Gabor filters [ 43 , 44 , 45 ] and wavelets [ 46 , 47 , 48 ].…”

Section: Introductionmentioning

confidence: 99%

A Class-Independent Texture-Separation Method Based on a Pixel-Wise Binary Classification

Soares

Côco

Ciarelli

et al. 2020

Sensors

View full text Add to dashboard Cite

Texture segmentation is a challenging problem in computer vision due to the subjective nature of textures, the variability in which they occur in images, their dependence on scale and illumination variation, and the lack of a precise definition in the literature. This paper proposes a method to segment textures through a binary pixel-wise classification, thereby without the need for a predefined number of textures classes. Using a convolutional neural network, with an encoder–decoder architecture, each pixel is classified as being inside an internal texture region or in a border between two different textures. The network is trained using the Prague Texture Segmentation Datagenerator and Benchmark and tested using the same dataset, besides the Brodatz textures dataset, and the Describable Texture Dataset. The method is also evaluated on the separation of regions in images from different applications, namely remote sensing images and H&E-stained tissue images. It is shown that the method has a good performance on different test sets, can precisely identify borders between texture regions and does not suffer from over-segmentation.

show abstract

Texture-based segmentation with Gabor filters, wavelet and pyramid decompositions for extracting individual surface features from areal surface topography maps

Cited by 18 publications

References 29 publications

Thresholding Methods for Reduction in Data Processing Errors in the Laser-Textured Surface Topography Measurements

Thresholding Methods for Reduction in Data Processing Errors in the Laser-Textured Surface Topography Measurements

Feature-based characterisation of signature topography in laser powder bed fusion of metals

A Class-Independent Texture-Separation Method Based on a Pixel-Wise Binary Classification

Contact Info

Product

Resources

About