Bambang Trisakti, Mahdi Kartasasmita, - Kustiyo, Tatik Kartika


Terrain correction is used to minimize the shadow effect due to variation of earth`s topography. So, the process is very useful to correct the distortion of the pixel value at the mountainous area in the satellite image. The aim of this paper is to study the terrain correction process and its implementation for Landsat TM. The algoritm of the terrain correction was built by determining the pixel normal angle which is defined as an angle between the sun and surface normal directions. The calculation of the terrain correction needs the information of sun zenith angle, sun elevation angle (obtained from header data), pixel slope, and pixel aspect derived from digital elevation model (DEM). The C coefficient from each band was determined by calculating the gradient and the intercept of the correlation between the Cos pixel normal angle and the pixel reflectance in each band. Then, the Landsat TM image was corrected by the algorithm using the pixel normal angle and C coefficient. C Coefficients used in this research were obtained from our calculation and from Indonesia National Carbon Accounting System (INCAS). The result shows that without the C coefficient, pixels value increases very high when the pixel normal angle approximates 900. The C coefficient prevents that condition, so the implementation of the C coefficient obtained from INCAS in the algorithm can produce the image which has the same topography appearance. Further, each band of the corrected image has a good correlation with the corrected band from the INCAS result. The implementation of the C coefficient from our calculation still needs some evaluation, especially for the method to determine the training sample for calculating the C coefficient. Keywords: Terrain correction, Pixel normal angle, C coefficient, Landsat TM

Full Text:

Full Text PDF


  • There are currently no refbacks.