Author: Barry Bitters, Ph.D., GISP
This paper describes a semi-automated process developed to extract and merge raw, geospatial feature data from multiple sites on the Internet. This process was tested on varied set of feature classes (specifically airfields, golf courses, maritime lights, police stations, and post offices of South Africa) to evaluate the effectiveness of the procedures. As an illustration, the results for airfield processing are presented. During processing of raw disparate data captured on the Web, attributes were aligned, duplicate records were eliminated, record-level metadata was preserved, and unit of measure conversions were performed. The processed raw data is then merged into a single dataset and attributes are integrated and duplicates are again eliminated. The final quality control review includes a final record-by-record imagery review of all candidate features to insure they actually exist on the ground, their geo-coordinates are accurate and that all duplicate instances have been eliminated. It has been found that this process is faster than the traditional wide-area search technique used to populate and maintain feature databases. The process also provides significant increase in location accuracy and increased feature counts relative to heritage data sources used in the study.
You are reviewing the
Download.
Your machine is locate at 3.138.135.4.