The aim of this work was to research and implement automatic paper precision counting using digital image technology. By analyzing the image texture of overlaid wafers, a counting algorithm was put forward targeting at the situation of large-format paper counting. First, first-order differential operator and morphological gradient were used for localization and extraction of texture region, then the oriented morphological grayscale was used for texture repair, finally, statistical counting of refined texture was conducted, and the advantages and disadvantages of each method as well as their influences on counting were compared. This algorithm strengthened the weak points of default information, and was suitable for repairing the damaged texture by trimmings and wear under the actual production condition. This counting method was experimentally validated. In conclusion, this counting method, with the characteristics of highly efficient and stable, can effectively and accurately count large-format paper under actual condition.