Efficient texture analysis of binary images

Ohser, Joachim; Steinbach, Bernd; Lang, Christian

doi:10.1046/j.1365-2818.1998.00390.x

Cited by 18 publications

(20 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“… Minkowski densities obtained with the method of Ohser et al[ OSL98 ] (filter size v = 10 , aspect factor ζ= 1 ). The point of a real pattern which is not on the curve corresponds to which is visually different from the other real patterns.…”

Section: Results and Applicationsmentioning

confidence: 99%

“…We have used a squared Euclidean distance map [ST94] of the cracks pattern and, as Vogel et al we have recovered the different erosions through a segmentation of this map, considered as a grey level image, according to different thresholds whose values correspond to the different values of r . As before, an estimation of the Minkowski densities was obtained by the method proposed by Ohser et al [OSL98].…”

Section: Results and Applicationsmentioning

confidence: 99%

“… Comparison of the Minkowski functions of real cracks patterns (a,b,c), simulated cracks patterns (d,e,f) and arbitrary geometrical patterns (g,h,i). The Minkowski densities: the area fraction (a,d,g), the line length (b,e,h), the integral of curvature (c,f,i) are estimated by the method proposed by Ohser et al[ OSL98 ] (filter size v = 10 , aspect factor ζ= 1 ) and they are represented as functions of the signed squared distance, corresponding to the level of erosion r. A positive value of r is related to the erosion of white regions, and a negative value to the erosion of black regions (i.e. the cracks).…”

Section: Results and Applicationsmentioning

confidence: 99%

See 2 more Smart Citations

A Dynamic Model of Cracks Development Based on a 3D Discrete Shrinkage Volume Propagation

Valette¹,

Prévost²,

Lucas³

et al. 2007

Computer Graphics Forum

View full text Add to dashboard Cite

We attempt to model and visualize the main characteristics of cracks produced on the surface of a desiccating crusted soil: their patterns, their different widths and depths and their dynamics of creation and evolution. In this purpose we propose a method to dynamically produce three-dimensional (3D) quasi-static fractures, which takes into account the characteristics of the soil. The main originality of this method is the use of a 3D discrete propagation of 'shrinkage volumes' with respect to 2D precalculated paths. In order to get realistic cracks, we newly propose to take into account a possibly inhomogeneous thickness of the shrinking layer by using a watershed transformation to compute these paths. Moreover, we use the waterfall algorithm in order to introduce in our simulation a hierarchy notion in the cracks appearance, which is therefore linked with the initial structure of the surface. In this paper, this method is presented in detail and a validation of the cracks patterns by a comparison with real ones is given.

show abstract

Section: Results and Applicationsmentioning

confidence: 99%

Section: Results and Applicationsmentioning

confidence: 99%

Section: Results and Applicationsmentioning

confidence: 99%

See 1 more Smart Citation

A Dynamic Model of Cracks Development Based on a 3D Discrete Shrinkage Volume Propagation

Valette¹,

Prévost²,

Lucas³

et al. 2007

Computer Graphics Forum

View full text Add to dashboard Cite

show abstract

“…This number can be calculated very efficiently using binary operations, avoiding the explicit calculation of G (cf. Ohser et al, 1998). The approach is illustrated in Fig.…”

Section: Configuration Counts By Filteringmentioning

confidence: 99%

Directional analysis of digitized planar sets by configuration counts

Jensen

Kiderlen

2003

Journal of Microscopy

View full text Add to dashboard Cite

Key words . Digitization, length density, mean normal measure, oriented rose of directions, stationary random set. SummaryThe oriented rose of normal directions of a random set Z can be used to quantify its anisotropy. A method for estimating this quantity from a digitization of Z in a sampling window, i.e. its pixel image, is presented. The image is analysed locally by considering pixel squares of size n × n . This allows us to count the number of different types of n × n configurations in the pixel image. We show that it suffices to restrict attention to the socalled informative configurations. The number of informative configurations increases only polynomially in n . An algorithm to find these informative configurations is presented. Furthermore, estimators of the oriented rose based on counts of observed informative configurations are derived. The procedure is illustrated by a simulated example and an analysis of a microscopic image of steel.

show abstract

“…While the former method is highly efficient with a low directional and high lateral resolution, the latter one features a high directional resolution and a low lateral resolution. Ohser et al (1998) pointed out the mutual exclusion of increasing both, directional and lateral resolution for this class of algorithms. In the present paper, estimating the rose of directions is not considered, since in contrast to the methods that we will use here, it does not allow for a local analysis of fibre systems.…”

Section: Introductionmentioning

confidence: 99%

Estimating Fibre Direction Distributions of Reinforced Composites From Tomographic Images

et al. 2016

View full text Add to dashboard Cite

Fibre reinforced composites constitute a relevant class of materials used chiefly in lightweight constructions for example in fuselages or car bodies. The spatial arrangement of the fibres and in particular their direction distribution have huge impact on macroscopic properties and, thus, its determination is an important topic of material characterisation. The fibre direction distribution is defined on the unit sphere, and it is therefore preferable to work with fully three-dimensional images of the microstructure as obtained, e.g., by computed micro-tomography. A number of recent image analysis algorithms exploit local grey value variations to estimate a preferred direction in each fibre point. Averaging these local results leads estimates of the volume-weighted fibre direction distribution. We show how the thus derived fibre direction distribution is related to quantities commonly used in engineering applications. Furthermore, we discuss four algorithms for local orientation analysis, namely those based on the response of anisotropic Gaussian filters, moments and axes of inertia derived from directed distance transforms, the structure tensor, or the Hessian matrix. Finally, the feasibility of these algorithms is demonstrated for application examples and some advantages and disadvantages of the underlying methods are pointed out.

show abstract

Efficient texture analysis of binary images

Cited by 18 publications

References 8 publications

A Dynamic Model of Cracks Development Based on a 3D Discrete Shrinkage Volume Propagation

A Dynamic Model of Cracks Development Based on a 3D Discrete Shrinkage Volume Propagation

Directional analysis of digitized planar sets by configuration counts

Estimating Fibre Direction Distributions of Reinforced Composites From Tomographic Images

Contact Info

Product

Resources

About