Sampling Strategy and Accuracy Assessment for Digital Terrain Modelling

Remote sensing, Geographic information science and geodesy

Li, Zhilin

University:University of Glasgow

关键词: Remote sensing, Geographic information science and geodesy;   

Others  :  http://theses.gla.ac.uk/78082/1/11007361.pdf

来源: University of Glasgow

【 摘 要 】

In this thesis, investigations into some of the problems related to three of the main concerns (i. e. accuracy, cost and efficiency) of digital terrain modelling have been carried out. Special attention has been given to two main issues - the establishment of a family of mathematical models which is comprehensive in theory and reliable in practice, and the development of a procedure for the determination of an optimum sampling interval for a DTM project with a specified accuracy requirement. Concretely, the following discussions or investigations have been carried out:- i). First of all, a discussion of the theoretical background to digital terrain modelling has been conducted and an insight into the complex matter of digital terrain surface modelling has been obtained. ii). Some investigations into the improvement of the quality of DTM source data have been carried out. In this respect, algorithms for gross error detection have been developed and a procedure for random noise filtering implemented. iii). Experimental tests of the accuracy of DTMs derived from various data sources (i. e. aerial photography, space photography and existing contour maps) have been carried out. In the case of the DTMs derived from photogrammetrically measured data, the tests were designed deliberately to investigate the relationship between DTM accuracy and sampling interval, terrain slope and data pattern. In the case of DTMs derived from digital contour data, the tests were designed to investigate the relationship between DTM accuracy and contour interval, terrain slope and the characteristics of the data set. iv). The problems related to the reliability of the DTM accuracy figures obtained from the results of the experimental tests have also been investigated. Some criteria have also been set for the accuracy, number and distribution of check points. v). A family of mathematical models has been developed for the prediction of DTM accuracy. These models have been validated by experimental test data and evaluated from a theoretical standpoint. Some of the existing accuracy models have also been evaluated for comparison purposes. vi). A procedure for the determination of the optimum sampling interval for a DTM project with a specified accuracy requirement has also been proposed. Based on this procedure, a potential sampling strategy has also been investigated.