Carpet Texture Measurement Using Image Analysis

Abstract

Image analysis techniques have been applied to the objective measurement of carpet texture. By converting an image into a form that highlights the regions of largest intensity variation, small differences in texture between carpets, due to either wear treatment or construction, can be reliably detected. This has been demonstrated using sets of samples that have received controlled wear treatments in the Hexapod tumbler tester. The optimum conditions for texture measurement are discussed, in particular the requirements for lighting and camera focusing. The influences of sample orientation and substrate color are also stressed. While the emphasis in this study is on plain shade carpets, possible methods of measuring the pile texture of colored pattern carpets are also discussed. Image analysis is considered to have a promising future as a scientific tool in studying carpet performance, as an objective means of carpet grading, and for product optimization and quality control in carpet manufacture.

Learning to rank: a ROC-based graph-theoretic approach

“…However, most researchers applied image analysis techniques to classify carpet wear. Extensive research has been conducted in this field using many different image analysis algorithms with the intention to quantify tuft definition, tuft geometry, tuft placement, periodicity, texture, pile-lay orientation and roughness (see for example (Wood & Hodgson 1989, Xu 1994, Pourdeyhimi et al 1994, Sette et al 1995, Van Steenlandt et al 1996). Some of the algorithms which have been used successfully on a limited set of carpet samples are grey value histogram analysis, local intensity-variation filters, statistical measures, edge detection filters, template matching and classifier systems.…”

“…Another problem in the development of such a global system is the reproducibility of the images (Wood & Hodgson 1989). Lighting arrangement, camera setup, focussing and level of digitization noise must all be considered with the greatest care to produce accurate (and reproducible) results.…”

Analysing wear in carpets by detecting varying local binary patterns

“…2,3 Manufacturers however require a more objective assessment because the human assessment is often too subjective. Much effort has been devoted to automate this rating process, initially on techniques extracting texture features from digital images, [4][5][6][7][8][9][10][11][12] and lately combining features from digital and range images,, [13][14][15][16] where a range image represents the depth information of the surface of a carpet. The range images are obtained with a 3D scanner designed specifically for scanning carpets.…”

Quantifying Appearance Retention in Carpets Using Geometrical Local Binary Patterns

“…Several studies based on computer vision have been aimed to quantify the AR grades objectively [5][6][7][8][9]. They still are not good enough to meet the required discrimination of the AR grades imposed by standards on a sufficient wide of carpets [10].…”