Otsu 'segmentation Method Research And Application

Image segmentation is one of the most important issues in image processing technology and computer vision research. For example, in computer vision research, image understanding including target detection, feature extraction and target recognition, etc. And image understanding relies on the quality of image segmentation. Image segmentation is the basic premise of pattern recognition and visual analysis, and segmentation accuracy directly affect the quality and efficiency of subsequent image analysis, identification and interpretation, therefore, image segmentation has the extremely vital significance in image processing.Threshold segmentation method is the most commonly used image segmentation method. In essence, threshold segmentation method is based on the gray histogram information of target image to obtain optimal threshold. Otsu method is always considered as the most useful method among all the automatic selection methods. According to the gray-scale characteristics of image, image is divided into two parts, target and background, by using one-dimensional Otsu method. As of the simplicity and effectiveness of this method, under certain conditions, it isn't affected by the contrast and brightness changes of image; this method is widely applied in real-time image processing systems.In this paper, we introduce the purpose, significance, research status and prospects of digital image segmentation research. This paper also introduces basic theory, commonly used segmentation method and classification of digital image segmentation. Otsu theory is the main research object of this paper. First, this paper introduces the traditional one-dimensional Otsu method, and an improved one-dimensional method based on single-threshold Otsu is proposed, which comprehensive considers a kind of one-dimensional single-threshold Otsu method and a kind of multi-threshold Otsu method. Experiments result and the corresponding analyzing show that this method is more effective compared with traditional Otsu method. Secondly, multi-dimensional Otsu method is studied in this paper. we generalize one-dimensional Otsu method to two-dimension, and proposed a new two-dimensional Otsu method. Related experiments and results analyzing indicate that the new proposed method is effective. Finally, main work and further research direction are summarized in this paper.