A structured approach to object segmentation in aeronautical charts.

摘要

Object segmentation is an important step in the conversion of a paper document image into an electronic document. While a digital image is simply composed of a sequence of pixel values, an electronic document represents the elements of an image in a content-specific manner. Objects that have been extracted from an image can be used to support a process that interprets the image at a higher level.;This work presents a methodology for automatically segmenting objects in scanned images of aeronautical navigation charts. Both colour information and domain specific knowledge are employed to accomplish this task. The process is described in the context of developing a system to aid in the recognition of aeronautical charts by pilots. In the high workload environment often encountered by pilots, there are many opportunities for error when performing common tasks such as map reading. If the flight parameters obtained from the manual inspection of a chart can be automatically verified, then the level of flight safety is consequently increased.