地形地貌
地质学
人工智能
启发式
地貌学
计算机科学
作者
R.A. MacMillan,P. A. Shary
出处
期刊:Developments in psychiatry
日期:2009-01-01
卷期号:: 227-254
被引量:91
标识
DOI:10.1016/s0166-2481(08)00009-3
摘要
Landform types are conceptualize as mainly consisting of waveform features that exhibit entire repeating cycles of variation in morphological properties, such as slope gradient, slope lengths, relief, curvatures, and moisture regime. These cyclic patterns can be identified and characterized by analyzing the distribution of variation in morphological attributes within neighborhoods defined by windows of appropriate dimensions and shape, as morphological variables computed for any given cell describe only a small portion of the total cyclic variation that characterizes a repeating landform type. Landform types provide information on the size and scale of landform features and how this size and scale might affect the amounts of energy available for geomorphic, pedogenic, and hydrological processes. Landform types provide context that can be used to inform and improve the further sub-division of the landscape into landform elements. Landform elements have been classified based solely on consideration of their local surface shape, on consideration of a combination of surface shape and slope gradient, and on consideration of a combination of surface shape, slope gradient, and contextual measures of relative landform position. Procedures for automatically extracting and classifying landform types and landform elements differ in terms of the kinds of classification methods applied to extract the entities. Repeating landform types have mainly been classified using Boolean rules based on expert knowledge and Heuristic beliefs. Classification of landform elements has been achieved using a wide variety of classification methods including knowledge-based heuristic approaches, supervised classification, and unsupervised classification.
科研通智能强力驱动
Strongly Powered by AbleSci AI