Objectives: The water contour extraction models using satellite images could be highly useful in monitoring the long term changes in river tributaries, lakes and other coastal areas. The lakes are termed for a localized basin of varied size where the water from river tributaries got reserved. These lakes are one among the prime source of water consumption for human needs. Hence, change monitoring in water levels of lakes is a highly needed measure for sustainability. Methods/Statistical Analysis: Though use of bathymetry and elevation data can aid the coastal monitoring models, a simple automated closed contour extraction model can support shape based change monitoring or satellite image based water volume assessment models. In general, the contour extraction of lakes or any water bodies is relatively difficult when compared with urban structures from spatial data. It is due to the indefinite shape of water bodies on varying levels of water. The use of image filters to improve the power of boundary discrimination is inevitable. However, it may also deteriorate the image quality leading to loss of information. Hence, the proposed model has examined the influence of widely used image filters both for smoothening and sharpening with respect to satellite image on applying suitable image quality assessment metrics. Findings: On the empirical analysis of the chosen image filters, the Adaptive wiener and Gaussian filters are found to be the more effective sharpening and smoothening filters respectively for satellite images. Further, the outstanding image filters found on evaluation is combined with local thresholding; shape based filtering and morphological operations in a sequence to extract an effective closed contour of water bodies. Applications: The proposed model can be applied for satellite image preprocessing for effective noise suppression and edge preservation. The closed contour extraction can be applied with change detection based applications for satellite images.
展开▼
