Abstract:

Jacquard warp knitted patterns with net effect have complex curves, leading to the time-consuming design of jacquard patterns. In order to simplify the design process, an automatic identification approach was developed. After scanning fabric and obtaining images and pre-processing, Laws texture energy measurement, iterative threshold segmentation and edge smoothness algorithm were comprehensively used to get the fabric pattern and then generate notation of jacquard design. Test results show that the method can effectively remove the noise for the image recognition caused by the mesh fabric, rapidly and accurately complete the automatic recognition of the required patterns, and then generate the picture of design. Recognition accuracy can reach 90%, which is of practical value to improve the design efficiency of such fabric.