항공 원격탐사 자료의 융합을 통한 연안 해저 암반 탐지 방안 연구

Abstract

The distribution of seabed rock in the coastal area is connected to the navigation safety and development of ocean where it is a very important measurement target. Currently, the distribution of seabed rock is surveyed with interpretation of water depth data or point based bottom sampling, which shows low efficiency. This study aims to suggest an efficient and accurate detection method of seabed rock using airborne remote sensing data. Airborne bathymetric LiDAR data and hyperspectral image were used to detect seabed rock, respectively. Then a method was suggested for seabed rock detection using fused data which could overcome limitations from single data. The seabed rock could be detected using texture information of airborne bathymetric LiDAR data. The result shown 88% detection accuracy and24% commission error due to blur effect of moving window operator for texture calculation. The spectral information was used to detect seabed rock using airborne hyperspectral image. The spectral angle mapper (SAM) algorithm was used to compare spectral similarity of each pixel to training samples. The result shown 79% detection accuracy with 16% commission error and 26% ommission error. The reasons of error might be limitation of image correction method and environmental variation at local area. The two data was fused to enhance detection accuracy that the texture data was added to hyperspectral image as a band. The maximum likelihood (MLH) method was applied to detect seabed rock using fused data. The result shown 96% detection accuracy, 6% commission error and 1% ommission error which is 8%p and 17%p enhanced accuracy compare to results with bathymetric LiDAR data or hyperspectral image, respectively. Therefore, the suggested method should increase accuracy on seabed rock detection with fusion of two remote sensing dataset. Additionally, it may improve efficiency to survey seabed rock distribution. Distribution of seabed rock can be a useful data on application fields such as navigation safety, military operation and sustainable development of coastal area. For this, further studies are needed aspect on accuracy enhancement, subdivision of class and expansion of area with various coastal characteristics of water and environments around Korean Peninsula.