The principal objectives of this SBIR proposal are to design, test and market an automated system for producing current population estimates from remotely sensed data. A procedure to transform both aerial photography and low cost, high resolution satellite imagery by combining image processing technology and demographic estimation techniques may make it feasible to provide monthly population estimates for any user- defined area in the United States or Canada. The project's key innovation involves the unique combination of image processing technology, demographic estimation techniques and remotely sensed data resulting in a radically new and more robust, empirically-based method for obtaining current population estimates. Specific goals of this Phase I proposal include: 1) installing and testing image processing software, 2) calibrating software to 1-meter aerial photographic test data, 3) evaluation of pattern recognition algorithms in identifying housing units, and 4) validation and calibration of housing estimates at selected Ohio and Nevada test sites. Since the majority of commercially available population estimates are based on either often outdated administrative records and symptomatic indicators or presumptive mathematical procedures, any process which can generate estimates from empirica!ly-based, current housing unit counts will have enormous commercial potential.
The creation of a rapid, low cost, computerized procedure for deriving population estimates of user defined areas based on current, remotely sensed data will be of strategic importance to a wide range of commercial interests seeking """"""""up-to-the-minute"""""""" counts and measures of population change for wide ranging purposes including site location and precise estimates of market size.
