The use of edges of photographic images as specifiers of image quality
Loading...
Date
1972-11
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Ohio State University. Division of Geodetic Science
Abstract
The primary objective of this study was to search for some measurable quantity which would specify the quality of an image with reference to a photogrammetrist's ability to measure to an edge. The image quality criterion was the closeness together of the edge determinations of four experienced observers who were asked to set a measuring mark tangent to a series of edges generated by rapidly rotating various patterns. An edge of high measuring quality was defined as any edge which rendered a high degree of agreement in edge determination and was measured by computing the standard deviation in edge location of the four observers. The secondary objective of studying where an observer determines the edge to be located was easily studied as a result of the many edge determinations required in ranking the edges as to image quality. A new term called decutance was derived which correlates very highly with the measured image quality. Additionally, acutance and the slope of the density edge trace also displayed a high correlation with image quality. Other possible specifiers were examined and found to be of lesser value. It was found that an observer subjectively locates an edge slightly into the less dense region as measured from the inflection point of the density edge trace. Based on this study and assuming that one micrometer on the photograph subtends an angle of 10 seconds of arc at the observer's eye, the ambiguity in subjective edge location can cause an error as great as ±3μm in edge location at the photo scale. It would seem that the subjective error connected with edge location could be greater than the otherwise obtainable accuracy using present day photogrammetric techniques.
Description
Prepared for U.S. Army Engineer Topographic Laboratories, Ft. Belvoir, Virginia: Contract No. DAAK02-71-C-0122, OSURF Project No. 3147-A1