Shoreline data extraction from QuickBird satellite image using semi-automatic technique

Abstract

The alternative technique of acquiring shoreline data using satellite images or from aerial photos has renowned as a fast and less time consuming process as compared to the conventional method of land surveying technique at the sea shore, especially when covering large and rugged area. This paper discussing the process of extracting shoreline data from high resolution satellite image using various image classification method. The aim of the study is to utilize remote sensing application in determining shoreline boundaries by conducting several image classification at the chosen study area. In order to obtain better accuracy of the shoreline position, image selection has been made prior to the high water during high tide of the study area. The selected image was then went through several processes like geo-referencing, ortho-rectification, image sub-setting, masking off the cloud and shadows as to ensure the image is corrected from error. Then, various semi-automatic methods like raster colour slice, band ratio, Iterative Self-Organizing Data Analysis (ISODATA) and Mahalanobis distance was applied to extract the shoreline data. From the analysis, only shoreline result which shown smooth separation between the edge of vegetation line boundary and the sea shore is considered to be best approximation to the highest water mark. Among all of the tested methods, qualitatively the shoreline data extracted using Mahalanobis distance has shown better performance in delineating the shoreline.