An input apparatus for a computer-aided picture data base management system, which has an image input device that optically reads an original picture image, such as a map having objects drawn by closed loop lines, and produces its digital image. An object image pattern extracting unit executes a digital image processing to extract map objects from the image and produces vector data of a series of coordinate data streams that define the geometric shape of each extracted object. The vector data of the extracted objects are automatically affixed with identification labels. An image presentation device outputs a reference image including the extracted objects affixed with the identification labels and presents it to an operator. When the operator manually prepares a document describing the list of plural sets of attibute data to be input, which are also affixed with identification labels, by referring to the reference image and sets the document to an optical character reader, which produces a character data signal that represents plural sets of attribute data corresponding to the extracted objects. This signal is supplied to an automatic map data generator, which automatically combines plural sets of vector data of the extracted objects with plural sets of attribute data through the identification labels in such a manner that one set of attribute data is merged with vector data having the same identification label as this attribute data.