Method for capturing a document image, a scanner using the method and a

Abstract

A document image capture method and scanner, and an image processing apparatus incorporating such a scanner, in which a document is scanned two or more times. The first scan preferably provides bi-level image data, which is analyzed to identify blocks of uniform image type (for example, text, line drawing, grayscale image, or full-color image) within the document. The second scan, preferably performed at lower resolution than the first, provides grayscale or color information, which is substituted in the grayscale or color blocks, respectively, for the bi-level information obtained in the first scan. A third scan, to provide information of the third type, may also be performed. An operator preferably views an image of the document, based on the scanned information, to be sure that the identification and typing of the various blocks has been done correctly, and may instruct that the document be rescanned to provide new data for a designated portion of the document image, if it appears that an error has occurred. The information representing the document image obtained in this way is preferably stored using a set of linked bit maps, one bit map for each block. The memory capacity needed to store the information can be reduced further by treating the page and its margins as a frame, and by storing information about the frame, and any horizontal or vertical lines in the document, in simple vector form. Any portion of the document which is just background is not stored.