Automated Extraction of Data Display Configuration From Telemetry Applications

摘要

This paper presents a methodology for a knowledge-based approach to effectively extract configuration data from raster images and source code of display systems. Our approach uses three different means of recognizing display objects from data display images and source code: extracting information from multiple raster images; vectorizing the images and using a rule- based object identification; and parsing the source code to extract graphics, dynamics, and data variables of the display system. No single technique would be able to completely extract display information. For example, it is impossible to completely extract display dynamics even if numerous raster images were provided. To solve this, we apply data fusion techniques in order to decrease the uncertainty inherent in each of the above-mentioned techniques.