Image Segmentation Methods Based On Adaptive Background Region Choice

Image segmentation is the process of divided into which contains real-world objects or regions have strong correlation integral part of the image. Image segmentation is the image processing and analysis of important issues, but also is a classic dilemma in the computer vision research area. In the area of studying on intelligent automatic technique, image segmentation has been of great importance to scientific research personnel, and also explored a number of classical image segmentation methods, but its development in the segmentation effect and the time still has much shortage, and is considered a bottleneck in computer vision. To date, no single image segmentation method is applicable to all images, a class of image segmentation methods are all applied to it.Thresholding segmentation algorithm is one of the most classic, simple and popular method of image segmentation. According to the image of the overall information or partial information to select one or more thresholdings, thus the image is divided into two types of the background area or multiple targets and background area. Now has made a lot of the threshold algorithm. According to the number of thresholding, the methods can be divided into single-threshold, dual-threshold and multi-threshold image segmentation methods. According the different roles of the threshold range, the threshold method is divided into the global threshold and the local threshold too. After refer to a lot of literature found that there a lot of threshold segmentation algorithm divided partitions of the edge is not clear enough, or can not be a good division of multi-objective. In this paper, the advantages of threshold algorithm is proposed which is a image segmentation method based on adaptive background region of choice, and for this algorithm is done in less than a certain amount of improvement. The main works of the dissertation can be organized as follows:1 Firstly this article introduces the research background and significance, the basis concept of image segmentation and theoretical. Concluded the progress in image segmentation research in recent years, summarized each algorithm for the advantages and shortcomings.2 Introduces the threshold method of image segmentation principle detailed and several common methods of image segmentation. Proposed the image segmentation method based on adaptive background region select. This method has the advantage that is presented ideas of hierarchical statistical pixel information; take full advantage of pixel spatial information, the image divide into layer but without losing the integrity. By a large number of experiments the algorithm proved that clarity and accuracy of image segmentation, so have a high degree of effectiveness.3 Describes the principles and application of fuzzy theory and analyze the advantages and shortcomings of fuzzy thinking.4 Introduced fuzzy threshold image segmentation and fuzzy clustering segmentation method, learn fuzzy thinking in the application of image segmentation and study the fuzzy algorithm of the difficulty in image segmentation, propose own improved. For the image at different levels, one layer of pixel information will constraint the next layer, under the shortcomings, have made the use of fuzzy ideas to be got the reconcile threshold, so that each layer to maintain the integrity and at the same time not be too constrained by the background information of the previous region. Through the test images demonstrate that the algorithm is effective.