Creating an Intuitive and Effective User Interface for Map Processing in a Geographic Information System
ISBN 978-85-88783-11-9
Authors
1Fernandes, R.; 2Chiang, Y.
1SPATIAL SCIENCES INSTITUTE,UNIVERSITY OF SOUTHERN CALIFORNIA Email: rvfernan@usc.edu
2SPATIAL SCIENCES INSTITUTE,UNIVERSITY OF SOUTHERN CALIFORNIA Email: yaoyic@usc.edu
Abstract
Map labels communicate key information about geographical phenomena, but can be difficult to manage in image form. However, converting map text into digital form is difficult because most Optical Character Recognition (OCR) techniques are not designed for the sporadically placed, non-uniform text in map images. Furthermore, existing text recognition tools for maps are not coupled with a Geographical Information System (GIS), a tool through which map data formatting and evaluation is primarily done. This limits users who want to work with the digital form of map text for some map processing work because it forces them to switch between a text recognition tool and a GIS. In a previous work, we established an approach for recognizing and extracting text from map images within GIS software. This research culminated in a GIS plug-in ArcStrabo. In this paper, we tackle the challenges of creating an effective User Interface (UI) for ArcStrabo. We present a customized UI, which includes a toolbar that gives users multiple options (for map layer, language, text examples, etc.). This workflow does not require GIS users to have technical knowledge of text recognition algorithms. Rather, ArcStrabo’s workflow is structured so that users perform routine GIS tasks (such as editing a vector layer) to feed data into ArcStrabo’s underlying OCR algorithms. ArcStrabo keeps the algorithm tuning and training hidden from the user. We also extend ArcStrabo’s run-time response to avoid crashing or outputting false results due to common user mistakes, allow users to run multiple tasks, and display results in a GIS-style format. This user interaction strategy is essential to the successful use of ArcStrabo and fosters correct and easy use of complex text recognition algorithms.
Keywords
User Interface; Text Recognition; Geographic Information System