Real-Time Registration in Image Stitching Under the Microscope

Image registration and fusion has always been an important part of computer vision, but most of the existing algorithms aim at the registration between two images, paying attention to accuracy, while sacrificing the requirements of speed The application scenario of this paper is real-time registration in image stitching under the microscope. That is, while moving the operating platform to observe the cell structure under the microscope, the stitching is performed in real time to obtain an entire spliced image. In order to speed up the image matching process and improve the accuracy, this paper creatively combines the phase correlation algorithm and the feature point matching algorithmfor image registration, and improves them, completing the registration of multiple images. This paper focuses on the real time ability and accuracy of the stitching program.