Design And Implementation Of Image Stitching APP Based On Feature Point

Image stitching is an important subject in the research fields of digital image processing, computer graphics and computer vision.Traditional image stitching technology is mainly used in medical diagnosis, military reconnaissance, remote sensing and other fields and these splicing work need to use professional image processing software and large computer to participate in operation. With the rapid development of mobile terminals, image mosaic technology can also develop towards the application direction with relatively cheap, fast and convenient.In view of the unnatural way of panoramic camera on mobile platforms so far and the current situation that image stitching experience is poor and its function is relatively simple, this paper designs and implements the image stitching software on Android platform. The main work is as follows:Firstly, the advantages and disadvantages of operators with different feature points are verified on the basis of a large number of test experiments. And this paper designs a feature detection strategy for the mobile platform according to the experimental results. The local stitching algorithm is proposed in this paper, which can effectively reduce the operation data of image detection and reduce the computation time,comparing with the free mosaic.Also, this paper designs the overall framework of the software according to the requirement analysis of the system and divides the system into four modules: panoramic shooting, free stitching, image editing and panoramic viewing. Then, the design and implementation of each module of the system are described in detail, and the key codes are described at the same time.Finally, we test the software function and terminal adaptation of image stitching APP on different Android mobile phones. Test results indicate that: The image stitching App can be used normally in the low-end Android phones and the effect of image stitching is good.