The design of a moderate gain embroidered antenna using the resonant cavity concept is presented. The patch as the main radiator and the partially reflective surface (PRS) structure as the superstrate are realized using an industrial embroidery machine. The PRS consists of periodic cross-slots (9 × 9 unit cells) in order to have a durable embroidered structure. The groundplane is made of a conductive woven fabric. The fabricated antenna has a relatively broad 1 dB-gain-bandwidth (10%) and a directive pattern (3 dB beam-width of 40° at the center frequency). Despite the higher losses in the conductive thread due to the resonant cavity effect, this antenna has a maximum realized gain of 6.8 dBi, which is typically 3 to 6 dB more than other embroidered single patch antennas.
With the improvement of mechanization, both large-scale computerized embroidery machines and small-sized home embroidery machines have gained rapid development, which demonstrates development of computerized embroidery technology. Currently, it’s convenient to capture images while it’s hard to convert images into embroidery files. The latter job needs to be finished by professional software and expert pattern makers. Current home embroidery machines don’t have matched pattern and making software and converting images into embroidery files automatically still need to be explored. This thesis makes outline capturing in order to set a good foundation for making of pattern files by stitching filling based on the outline that can be recognized by embroidery machines.
Border detection plays a very important role in image processing. At present, border detection of gray images has gained remarkable achievement while that of coloured images still needs improvement. Gray images can be seen as two-dimensional and plane while coloured images can be seen as three-dimensional and stereo. When detecting border of coloured images, we can divide them into three two-dimensional gray images and thus apply border detection method of gray images into coloured images
At present, the border detection of coloured images algorithm is mainly used in RGB space. However, the RGB coloured models cannot meet people’s interpretation of colours and the three components of R, G and B are mutually related rather than independent, which means change of any one will lead to change of the other two. Therefore, it’s hard to control them and et a good detection result. However, the HIS coloured model is different. It can eliminate the impact of strength component on colour information (chrome and saturation) contained in coloured images and description of colour is natural and direct2. Therefore, this thesis focuses on outline capturing of coloured images .
Commentaires