Shoreline mapping and monitoring are crucial in heavily eroding coastal areas. Frequently monitoring shoreline changes in eroding areas can help government agencies to understand the causes and formulate plans to protect the shoreline; however, the cost of data sources and human labor involved can be prohibitive. To address this problem, this research proposes a cost-effective and semiautomatic shoreline delineation procedure that uses WorldView-2 satellite images as the data source. The advantages of the proposed procedure are the use of shadow and spectrum information, as well as the application of objectoriented classification. The classification process incorporates new bands from WorldView-2 to perform spectrum analysis on multispectral images; subsequently, a new pan-transferring process for the classification result is proposed to maximize the resolution. Finally, a classification adjustment process based on scenario analysis using shadows as information is applied. This procedure remedies the problems caused by shadow areas and converts them into elevation information to minimize errors in shoreline delineation. The analysis results indicate that the proposed procedure can derive the location of an instantaneous shoreline with an accuracy of 1.8 m (root-mean-square error) in a test area of Painesville, Ohio, USA.
"INSTANTANEOUS SHORELINE MAPPING FROM WORLDVIEW-2 SATELLITE IMAGES BY USING SHADOW ANALYSIS AND SPECTRUM MATCHING TECHNIQUES,"
Journal of Marine Science and Technology: Vol. 24:
6, Article 18.
Available at: https://jmstt.ntou.edu.tw/journal/vol24/iss6/18