The Study On Image Mosaic Using Genetic Algorithm

Remote sensing technology is playing a more and more important part in environmental monitoring, land and natural resources exploration, archeology military surveillance and many other fields. With its unique flexibility and rapid reaction capability , aerospace remote sensing is extremely important role in the real-time dynamic control to deal with a variety of emergencies and disasters of all of the real-time dynamic control. However, due to various factors such as the height of the flight restrictions and camera equipment ,when the special area is too large, we will need a number of remote sensing images to cover it.Image mosaic technology is a kind of technology that using two or number of images which have overlapping area to generate a larger panoramic image through image registration and deformation. The technology is widely used in computer vision, image processing and computer graphics, and other important aspects of the application .In the field of remote sensing applications, especially aerial remote sensing applications, the flexibility and dynamic of aerospace remote sensing often makes mosaic of aerial remote sensing images more complex than the satellite images. Therefore this paper aimed at developing a new method based on the point matching , which can perform the registration of images and the image mosaic automatically.The corner matching process often involves a lot of computation. Of many optimization algorithm , the genetic algorithm is used in different areas and often gives a good result with its unique ability to search strategies and global optimization. In this paper we have developed a new method to encode the corners detected in the images, use normalized cross-correlation algorithm to evaluate the fitness of the individuals, specify the method used in selection, crossover and mutation. Thus we used the Genetic Algorithm successfully to optimize the point matching process, and get a perfect result.