Development of A Pixel Segmentation Method for Enhanced Accuracy Assessment of Moisture Content and Evapotranspiration using Satellite Data and Field Measurements

Soil moisture (MC) and evapotranspiration (ET) are considered as the most
significant boundary conditions controlling most of the hydrological cycle's processes
especially over humid areas. However, monitoring them continuously over large areas
using the high temporal-resolution optical satellites is very demanding. Satellites such
as the Advanced Very High Resolution Radiometer (AVHRR) and the Moderate
Resolution Imaging Spectroradiometer (MODIS), has a coarse spatial resolution in
their images. Thus it not only impedes the acquisition of an accurate MC and ET but
also represents multispectral reflections from the holistic surface features. The study
aims to enhance the spatial MC accuracy by pixel segmentation method. MC for each
pixel is segmented into three (3) different classes namely urban, vegetation and multi
surface cover according to their respective MC weightage.