Research On The Application Of Fast Panorama Stitching On Mobile Phone

With the increasing widespread popularity of smart mobile terminals, today mobileterminal is not just an effective communication tool, but also a device withhigh-performance computing power. This smart phone meets many sides of people’s lives,such as high-pixel camera and video recording, high-quality color display, accurate mobilelocation and convenient access to do shopping and browse the web coupled withhigh-performance GPU hardware processor. Thus those applications that only found indesktop running on the computer can also be implemented in the smart phone, such as facerecognition and image effects processing. In recent years, mobile panorama also gains thefavor of many domestic and foreign developers.but because the phone panorama in imageprocessing requires more computation and memory space, so a mobile smart phone withlimited resources can not meet the panorama processing requirements well. Now thedevelopment of effective and rapid panorama stitching algorithm is imperative. Forpanorama the smart phones now there are two main processing problems. Firstly due touneven lighting and photographing scenes of different perspectives, the captured imagescan easily lead exposure difference. Secondly, the user needs require sport scene, so it isprone to produce “ghosting.” If these two problems can be resolved well，it not onlyensures the quality of the panorama, but also can accelerate the stitching speed of thepanorama.In this paper, the above-mentioned problems have been studied. Firstly analyze thegenerational process of mobile panorama and point out the process differences of themobile phone and the usual panorama. Secondly, the key technologies of the traditionalpanorama are described. Thirdly, to resolve large exposure for the original images thispaper puts forward to a new color correction algorithm. Finally, to achieve a dynamic scenepanorama, this paper provides the best way that finding optimal seams in overlapping areas,followed by image fusion technology to eliminate ghosting. Experimentations verify thatthe color correction algorithm not only can effectively reduce color differences, but alsoaccelerate the splicing speed of the image fusion, and that the blending method provided fordynamic scenes not only can avoid the objects of movement, but also can guarantee thequality of the image.