An Image Mosaic Algorithm Taking into Account Speed and Robustness

In this paper, a new image mosaic algorithm is proposed by studying the advantages and disadvantages of existing image mosaic algorithm. Firstly, finding a certain column at last frame in the two images which to be mosaic, making the gray difference between this column and other ones most obvious, that is to find the column whose feature is the most obvious as the matching template. Secondly, the correlated calculation is conducted between this column and each one at the next frame in turn,then the column that is the most relevant to the image at the next column is found and it is the best match one of the two images, thus the image mosaic is completed. Experiments show that the image mosaic can be successfully carried out by means of the algorithm.