Research On Raster Image Processing And Optimization Of Machining Path For NC Engraving

With the diversification of engraving products, the increasing demands for thoseengraving products which have personalization and the requirements of fast and real-timeresponse to the market, the numerical control(NC) engraving system which based on imageprocessing has become one direction of engraving system and is becoming more and morepopular. This paper take paper-cut pictures for example, and by using Matlab7.10.0(R2010a),we develop a NC engraving system which can process images and optimize the maching path.In this paper, the main works are as follows:1) Eight-connected contour tracing algorithm was adopted to track the contour in theimage after detecting the edges. By calculating the sum of current pixel’s eight-connected, ou-tliers and burrs can be distinguished from the boundary point,then the correct point on thecontour can be obtained.2) Using three-chain code sum to express the tangent of the boundary contour in theimage,and using three-chain code difference to detect the corners.In order to delete the flasecorners,neighbor maximum and adjacent corner methods were first used to eliminate the falsecorners. Then this paper simulate the algorithm and compares it to other algorithms, thesimulation result shows that, the algorithm has the highest rate of detecting the real corners.3) To make the boundary correct and avoid distortion,cubic spline interpolation anddouble circular-arc spline fitting methods was used to vectorize the boundary contour.4) Put forward a method to improve the nearest neighbor, and get the optimal contourstarting point,then ant colony algorithm was used to get the optimal machining path andreduce the empty run distances.Simulating the system which can process images and optimize the path and the resultshow that, comparing to harris、Freeman chain code corner detection algorithm,the system hasgreatly improved the rate of detecting the real corners, as well as the machining path.Thesystem is applied to the SCARA robot, The results are relative satisfied by using SCARArobot to draw three paper-cut pictures.