The Application Of SUSAN Algorithm And Virtual Circles In Image Registration

Image registration is an important research area of digital image processing, which is also an important subject of computer vision and pattern recognition ,and an important technique for a great variety of applications such as military affairs,pattern recognition,remote sensing and medical image processing . In brief, image registration is a process to match two or more images of the same scene taken at different times, from different viewpoints, or by different sensors.The methods of image registration can be classified into two categories: the intensity-based registration approach and the feature-based registration approach. For its simplicity,rapidity and other features ,feature-based registration approach is used in many application, edge detection is one of the common methods to get image feature. In this paper, we use SUSAN(Smallest Univalue Segment Assimilating Nucleus) algorithm to detect image edge. Compare with traditional edge detection algorithm, SUSAN edge algorithm has more advantages.There are two tasks which need to be handled during an image registration process. That is, the feature extraction and correspondence establishment. This text mainly completed the following work:First, we analyses the foundation of image registration,mathematic model and common methods of image registration.Second, aiming at the feature-based registration approach, we chose traditional edge detection as our focus, and some experiments have done to respect their performance.Third, SUSAN algorithm is the nucleus of this paper, so, this part mainly introduces the knowledge of SUSAN algorithm including its conception,principle,performance,SUSAN edge detection etc.Fourth, in this part we introduce the technique of virtual circle.Fifth, in this part we combine SUSAN edge detection with virtual circle technique to formulate our new registration approach. Traditional feature-based registration approach use line, circle or others as the feature of image. On the contrary, we use the white space of image as our feature from which we get the set of virtual circle. By comparing the virtual circle sets, we have managed to get the parameters of image registration.