| Sumario: | This article presents information on the procedures and research that are designed to evaluate the accuracy of aerial photographic interpretation data on urban residential structures. The study presented in this article was conducted on 17 residential subareas in Birmingham, Alabama in which housing data obtained by photographic interpretation were correlated with data collected by ground observation. In both the photographic and ground data collection, as well as in the organization and analysis of the data, attention was focused on two housing items of primary interest: first, the percentage of all dwelling-units classified as single-units, detached, and second, the average number of dwelling-units per block. The analysis of the data, correlating photographic and ground observations, was concerned with the "absolute" accuracy and the "net" accuracy of the photographic information. Absolute accuracy pertains to the validity of the photo keys. It is measured in terms of the percentage of correctly classified residential structures in each subarea. This type of analysis, revealing each individual discrepancy in photo interpretation, provides a possible basis for constructing systematic correction factors through statistical study of error direction and amount.
|
|---|