3D pattern for knitted objects

Abstract

This study attempted to develop a theoretical basis for knitting three-dimensional (3D) shapes. This generic and scientific study aimed at converting 3D data cloud into a 3D knitted shape. An approach of 3D data cloud conversion of a freeform surface based on its geometry, instead of finite elements in many present researches, was introduced. In order to preserve the target 3D shape in knitting, a series of experiments was carried out to reconstruct the 3D form from its data cloud. The relationship between 3D scanned data and two-dimensional (2D) knitting structure was examined. This study involved two parts, 3D data cloud conversion and 3D shape knitting. In order to develop a complete 3D shape knitting theory on freeform shapes, surfaces which would be examined in this study were categorised into three types by means of the Theory of Gaussian Curvature. As a freeform shape can be zoned into different atlases according to their curvatures, surfaces with zero, positive and negative Gaussian curvature were contemplated and taken to experiments separately in the study. All of them underwent a process of capturing 3D data cloud by 3D scanning, flattening the 3D form into a 2D plane as in cartography and reconstructing the 3D form by aligning knitting loops onto the plane. The scope of the study focused on weft knitting with double-bed flatbed knitting machines, so as to develop and establish a complete basis of theory proposed with fundamental knitting techniques and components. Ground on the knitting constrains in flatbed weft knitting, a knitting mechanism in 2D, knitted fabrics grow only in length upon a width depending on the number of needles to be selected, but not in depth in 3D. Discussions on how a 2D knitting mechanism was capable of creating a 3D shape by specific loop alignment methods were raised.