Abstract: The image mosaic technology is an important branch in computer vision and digital image processing. In this paper, firstly, research the influence on similarity measurement time and accuracy by absolute difference, difference of squares and normalized correlation coefficient .Then, eliminated the effect of splice joints by the integration of different methods of arithmetic average, weighted average and multi-resolution wavelet fusion. Finally, through analyzed the time complexity and accuracy of matching and the integration effect to find the suitable method for fabric image mosaic. The result showed that the use of gray-scale grid-based matching and application of the weighted average method is well positioned to complete the integration of image stitching fabric.