REAL-TIME CUTTING PATH DETERMINATION USING MACHINE VISION-BASED INCREMENTAL PATTERN TRACKING

In this article a fundamentally new approach to cutting patterned brea dths of lace fabric from a mesh backing web is presented. A line scan camera-based vision system employing a special tracking algorithm is u sed to detect pattern registration of lace fabric incrementally, allow ing real-time determination of the local cutting path in accordance wi th a pre-defined reference map. This approach achieves high tracking s peeds and a high level of robustness. The determined cutting path data is used to target a CO2 laser beam actively to achieve separation of lace ribbon at speeds of 1 m/sec. The new machine which utilizes this vision guidance system provides levels of production that have been hi therto unattainable with conventional mechanically based lace cutting machines. (C) 1996 Academic Press Limited