A Study of Applying Genetic Algorithm Combining with Operation Tree (GAOT) for Remotely Sensing Chlorophyll-a in Alpine Lakes in Taiwan - A Study Case of Lake Cuifeng

• Main Authors: HUANG , HONGG-REN, 黃泓仁
• Other Authors: Li Chen
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/5dzzaw

碩士 === 中華大學 === 土木工程學系碩士班 === 101 === This paper examines the influence that satellite images have on different stages of chlorophyll-a (Chl-a). Using Cuifeng Lake in Yilan County, Taiwan, as the study region, the research subject was Chl-a, which exhibited significant concentration differences in low-water seasons and high-water seasons. SPOT-4 and FORMOSAT-2 satellite image information was collocated with local Chl-a water quality sampling data and the spectral band values of local sampling sites. To compare the 2 models, multiple linear regression (MLR) and a genetic algorithm operation tree (GAOT) were employed to establish an equation for the relationship between the spectral band value of satellite images and Chl-a. The spectral band values in the study region were extracted and substituted into the equation to estimate the overall Chl-a value. The conventional method for determining water eutrophication was adopted. This method is employed by the United Nations Organization for Economic Co-operation and Development to indicate the degree of eutrophication of lakes. The results indicate that Chl-a concentrations are higher during low-water seasons and lower during high-water seasons. Furthermore, the estimation accuracy for low-water seasons is higher than that for high-water seasons, and GAOT is a superior model for estimating Chl-a. Satellite images can be used to estimate Chl-a, conserve the resources required for conventional artificial sampling, and also monitor Chl-a variations in Cuifeng lake.