An automatic image registration methods for multi-domain images by using geometric relationships

Abstract

In the internet technology world, all the information is considered as an image format only. Because one image is equal to a thousand words. Automatically increasing the image database size, high resolution and including multi-domain concepts. The human perception of getting the required image is very poor during the image retrieval process. In this case, the image retrieval system is consists of many problems like more time consuming, irrelevant results of the user. Registration is necessary to be able to compare or integrate the data obtained from different measurements of the query image. Image registration is that the method of reworking totally different sets of knowledge into one organization and orienting 2 or additional pictures of constant scene. This method involves designating one image because the reference image, conjointly known as the mounted image, and applying geometric transformations or native displacements to the opposite pictures so they align with the reference. To solve the above problems by using an automatic image registration methods. In this paper, we are mainly focusing the automatic image registration methods for multi-domain images for better human understating of required results. Finally, evaluate the performance of image registration methods with length, counts of matched points and uniformly distributed points of multi-domain images for better understanding.