Share Email Print
cover

Proceedings Paper

AI-based technique for tracking chains of discontinuous symbols and its application to the analysis of topographic maps
Author(s): Alessandro Mecocci; Massimiliano Lilla
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Automatic digitization of topographic maps is a very important task nowadays. Among the different elements of a topographic map discontinuous lines represent important information. Generally they are difficult to track because they show very large gaps, and abrupt direction changes. In this paper an architecture that automates the digitalization of discontinuous lines (dot-dot lines, dash-dot-dash lines, dash-asterisk lines, etc.) is presented. The tracking process must detect the elementary symbols and then concatenate these symbols into a significant chain that represents the line. The proposed architecture is composed of a common kernel, based on a suitable modification of the A* algorithm, that starts different auxiliary processes depending on the particular line to be tracked. Three auxiliary processes are considered: search strategy generation (SSG) which is responsible for the strategy used to scan the image pixels; low level symbol detection (LSD) which decides if a certain image region around the pixel selected by the SSG is an elementary symbol; cost evaluation (CE) which gives the quality of each symbol with respect to the global course of the line. The whole system has been tested on a 1:50.000 map furnished by the Istituto Geografico Militare Italiano (IGMI). The results were very good for different types of discontinuous lines. Over the whole map (i.e. about 80 Mbytes of digitized data) 95% of the elementary symbols of the lines have been correctly chained. The operator time required to correct misclassifications is a small part of the time needed to manually digitize the discontinuous lines.

Paper Details

Date Published: 30 December 1994
PDF: 12 pages
Proc. SPIE 2315, Image and Signal Processing for Remote Sensing, (30 December 1994); doi: 10.1117/12.196768
Show Author Affiliations
Alessandro Mecocci, Univ. di Pavia (Italy)
Massimiliano Lilla, Univ. di Pavia (Italy)


Published in SPIE Proceedings Vol. 2315:
Image and Signal Processing for Remote Sensing
Jacky Desachy, Editor(s)

© SPIE. Terms of Use
Back to Top