| This research addresses two topics that provide support towards automating the cartographic generalization of streams. The study was conducted on USGS topographic maps that were manually generalized over a sequence of four scales ranging from 1:24,000 to 1:250,000. The first topic quantifies the amount of change occurring for eight geometric measures. The second topic addresses whether streams of similar complexity (the amount of undulations and crenullations) are generalized similarly. A comparative analysis of Divider-Walk Fractal Dimensions and the Inflection-Point Sinuosity as methods for indexing the complexity of the line resulted in a high correlation between the two methods, but technical considerations favored using Inflection-Point Sinuosity. The rates of geometric changes to a cartographic stream line do not usually change in direct proportion to the rate of scale change. The pattern of change displayed over the range of scales provides a unique signature for each line. Each geometric measure has a unique signature for each line segment and for each scale increment. The stream lines can be clustered into four to six groups based on common signatures. The initial complexity of a line shows little correlation to the signatures, but it may prove useful in determining the rage of scales to which generalization rules can be applied. |
