Use this URL to cite or link to this record in EThOS:
Title: Shoreline mapping using satellite sensor imagery
Author: Muslim, Aidy Mohamed Shawal M.
ISNI:       0000 0001 3614 6026
Awarding Body: University of Southampton
Current Institution: University of Southampton
Date of Award: 2004
Availability of Full Text:
Access from EThOS:
Full text unavailable from EThOS. Please try the link below.
Access from Institution:
Remote sensing has been used widely to map the shoreline and offers the potential to update maps frequently. The shoreline could be mapped accurately from fine spatial resolution satellite sensor imagery. Utilizing fine spatial resolution satellite sensor imagery a shoreline prediction with an RMSE of 1.80 m was achieved. But this is an impractical approach for use over large areas. A pilot study was conducted to examine the potential of these methods on a linear stretch of shoreline. Using a simulated 20 m spatial resolution imagery, a conventional hard classification yielded a shoreline prediction with an RMSE of 6.48 m. To increase the positional accuracy, methods of fitting a shoreline boundary at a sub-pixel scale were examined. Initially a soft classification was applied to predict the class composition of image pixels which were located geographically using sub-pixel mapping techniques. Several sup-pixel mapping methods were applied; contouring, wavelet interpolation and two-point histogram. In the pilot study, the two-point histogram method obtained the most accurate prediction with an RMSE of 2.25 m followed by wavelet interpolation and contouring with an RMSE of 2.82 m and 3.20 m, respectively. This work was extended by analysing effects of shoreline orientation on the prediction. Using a 16 m spatial resolution imagery as a basis for analysis the accuracy of the shoreline prediction varied with orientation. For example, result from the two-point histogram method varied from the RMSE from 1.20 m to 2.08 depending on the shoreline orientation. To further increase the accuracy of the shoreline prediction, the method was revised by using localised training statistics in the derivation of the soft classification. Using the two-point histogram method, the use of the revised approach yielded shoreline prediction with RMSE ranging from 0.97 to 1.10 m. The result indicates that the accuracy of the shoreline prediction was positively related to the accuracy of the soft classification. This approach of shoreline mapping satisfied the requirement for mapping at a 1: 1,500 scale.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available