Researches On Segmentation And Recognition Methods Of Pulp Fibre Images

Image Segmentation and Image recognition is one of the most pivotal techniques of digital image processing and also is the foundation of object search, it has been widely attached importance to for a long time in theory research and practical use, there are many method for Image Segmentation and Image recognition, some of the methods have use value, but all of them have their own advantages in automatic degree,universal property,veracity,highly active property and the capacity of resisting disturbance of the Image Segmentation methods. The methods we often use are: the region-based Image Segmentation, the edge-based Image Segmentation and the edge-detection-based Image Segmentation and so on. The methods for Image recognition we often use are: the statistic law,the syntax law and the nerve network law, more often we have to use some of them in the same time to have a good segmentation and recognition of the image.The basic characteristic of the image is the edge, the edge is the concourse of the pixel whose around pixel has step change or rooftop change, the change is the most important characteristic that Image Segmentation depended on, it also is the important information source of texture characteristic and the base of shape characteristic and template matching. The mathematical morphology is a good image edge detection method, matching with median wave filtering method; it can detect the edge of pulp fiber image perfectly.The track and search of pulp fiber object image is a pivotal technique of digit image processing. How can we find the expecting object among the large number objects is the main problem the image search has to solve? How can we find the expecting fiber in the least time and least cost, to advance search rapid and efficiency and reduce time and money is the problem that image optimization search has to solve.In this passage, we discuss the search of object image based on Image Segmentation and Image recognition, and bring forward a method for the search method of expecting fiber in pulp fiber image based on the hausdorff distance method, and reach a good result.