Image Segmentation Applied Research, The Protection Of The Bamboo Slips

As an elementary and important part of digital image processing technique, image segmentation bears a broad prospect of application and researches. An image can be segmented according to some characteristic into some regions that have some meanings and will not be superposed by a method of image segmentation. Image segmentation is a key step from image processing to image analysis. From the viewpoint of image techniques, image segmentation is between image processing and image analysis. It is not like a common image processing such as image acquisition, image transforming, image enhancement and so on. And it is a basically important stage of image analysis. Because of its obvious importance, it has become one hotspot of research in the domain of image techniques. In the same time, the quality of image segmentation would directly affect the following result of image processing, therefore the image segmentation technique is regarded as the bottleneck of image processing and also has important meaning.In this paper , mainly studied the Bamboo Slips pictures, focused on gray picture splitting problems. The research work includes following items:1. After massive reading of references, a fully and systematic understanding of picture splitting algorithm and its qualities evaluation is achieved. Detailed research and summary of basic theory and deduction of classical picture splitting algorithm is done.2. Analysis, classification and conclusion of current gray picture splitting algorithm is carried out. Pros and cons of Different methods is clarified, provided reference for their application in different situations.3. When processing pictures, firstly pre-processing of color Bamboo Slips picture is carried out: making it have only two values and change to gray pictures. Linear and non-linear method is adopted to eliminate the noise. Boundary detection algorithm is used to prepare object for splitting.4. When splitting picture, three methods are carried out. They are based on threshold value, area and Maximum Between-Class Variance(OTSU)segmentation methods. And some parameters are changed in experiment to achieve better result.