Research Of Intelligent Image Processing Oriented To NC Engraving

With the development of engraving technique and competition,the engraving technology which has become one direction of engraving system is accepted based on image processing. The paper introduced intelligent control algolithm,and the image preprocessing,border detecting, border connection and contour tracking in image engraving system are researched according to NC engraving principle based on digital image processing.The topic is of great significance.The demands and procedures in engraving image preprocessing were analysed.The image-repair technology was discussed. A repair matrix was obtained based on distress image's handwork-repair,and then it could be used to repair full width image. This method could solve the problem perfectly that engraving machine can't recognize the images with poor quality. A new de-noise algorithm and a modified sharpening algorithm are presented,which are used to reinforce the image's quality. Therefore, the image contour was obtained for engraveing.A genetic algorithm of image edge detection was presented.Based on the handwork-repair training results,the optimal detecting template was obtained with genetic algorithm. Consequently identical images' edge detection was achieved successfully.The experiment results which indicates that perfect edge feature could be obtained shown that this algorithm was markedly improved in the point sharpness, peak value Signal-to-Noise Ratio and computing speed.Aimed at the variance characteristic and noise effect of the image border, one fuzzy edge linking algorithm was proposed,which mostly base on the edge-point feature to realize edge connection.The edge-point feature was established by weighting the unknown point's neighbourhood pixel,and the character of the gradation distribution of edge-point' neighborhood. The results show that the method can connect the obscure edges and disturbed edges effectivly,and the method could satisfy the demand of contour closing.The contour feature extraction was realized through object point's 8-neighborhood coding, and the contour's gray was defined, at the same time the relationship between contour dots was found, and the contour extraction and track algorithm were designed.The inner contour trackalgorithm was proposed based on segment coding, aimed at the problem that traditional contour tracking could only track outer contour. The test shows that this algorithm provides efficient solutions against the island in image engraving and isolated closing stroke.The NC engraving image processing module was realized with Visual C++, based on system functional requirement. The NC machine tool and articulated robot experiments were done based on the intelligent image process.The experiment result show that the practical method could satisfy the needs of engraving image processing.