Application of water column corretion in benthic habitat mapping

Abstract

QuickBird satellite can capture image of large area without contact with it. The image needs to be process to have better accuracy. The main objective of this project is focused on the differences between image that had undergone water column correction and without water column correction process. The images used are from QuickBird satellite. The software I used to process the image is IDRIS! Andes. Image processing of Lang Tengah island image included geometric correction, atmospheric correction, masking, water column correction, supervised correction and lastly accuracy assessment. The algorithm used for atmospheric correction is ATCOR 2 that used on flat terrain. Masking in this project is to remove the unwanted area that is the land and consider the sea as wanted area. Water column correction used in this project is Lyzenga method to remove the effect of absorption and scattering caused by water. Method used for supervised classification is Minimum Distance which was applied to the images that the water column had been corrected and uncorrected. Accuracy assessment was done to the images to find out which image has better accuracy. To complete the supervised classification and accuracy assessment, in-situ data that is bottom types are required. Sampling activity was done to get the in-situ data. The result of the project show that images with water column correction is more accurate than images without water column correction. Band 1 (blue band) image has better accuracy compared to band 2 (green band) and band 3 (red band) image.