University of Twente Student Theses


Detecting slow, gradual changes with remote sensing: fuzzification of random forest classification for coastal management

Patil, Bakul (2021) Detecting slow, gradual changes with remote sensing: fuzzification of random forest classification for coastal management.

Link to full-text:
(only accessible for UT students and staff)
Abstract:The city of Semarang faces the risk of frequent coastal and tidal flooding in combination with coastal erosion and land subsidence. Therefore, monitoring shorelines has a great significance in providing information on its dynamic nature, resource management, and evaluation of potential risks for sustainable coastal development and management. Changes in the shoreline are gradual and slow, and often, they are overlooked for a long time until it becomes a big problem. However, shorelines do not have clear boundaries as it is a transition zone between land and water. Therefore, it is a vague object. Hence, the soft classification will provide better information for the fuzzy shoreline boundary rather than hard classification. Being dynamic in nature, shorelines are challenging to identify, and therefore, their position contains a degree of uncertainty. This study focuses on understanding the problem at hand and aims at developing an approach to study the slow and gradual change in the shoreline using remote sensing and machine learning techniques. The probabilistic results of random forest give membership degrees for classes in the context of fuzzy logic. Moreover, these fuzzified results are used to determine three measures of uncertainty, namely, Confusion index, Ambiguity index, and Fuzziness for the study area. The implementation of the results of this study is further explored in contributing to the indicators of Integrated Coastal Zone Management. The uncertainty maps can also show the vulnerable areas to while disaster risk management and the areas that can be possibly reclaimed. The uncertainty in the classification shoreline over five years will result in uncertainty maps that coastal planners can use for sustainable spatial planning or disaster risk management. This study applies an approach that can be used as a tool for Coastal monitoring and management and disaster risk management.
Item Type:Essay (Master)
Faculty:ITC: Faculty of Geo-information Science and Earth Observation
Programme:Spatial Engineering MSc (60962)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page