A Large Area Of ​​the Automatic Extraction Of The Upper Surface Of Joints And Vectorization

From the practical application requirements to the automatic upper sewing process with the technological background of image processing , this dissertation focus on the following key technologies related to the upper image identification and vectoring process, including the pretreatments of upper image, BMP discerning and vectoring, drawing file format transforming based on OOP.Firstly, aiming at noise removing of the upper image in the input step and following requirements ,several pretreatment algorithms of image are introduced and realized. Algorithm for image thinning processing and its computing program are fully developed.Secondly, on the result of the image thinning process, methods for the image vectoring is under investigated, its algorithm are proposed in detail. Freeman chain code is adopted in vector tracing. On the basis of short straight-line fitting, this dissertation uses the least-square method of curve matching to discern circle and straight line. At the same time, compensating strategy is introduced for the elimination of distortion phenomena during the process of vectorization by wiping off short straight lines and reducing line segments.Finally , according to DXF's file format, the vector quantization data files are transformed to DXF's files. The graphics can be edited, modified and output in the platform of AutoCAD.The research of discerning and vectorization of image has been studied for quite a long-time, but its application in uppers sewing only in the step of underway. This dissertation is an elementary study on its application and also exists lots of shortages, so further research is needed.