The effect of Gaussian blurring on the extraction of peaks and pits from digital elevation models

Abstract

Gaussian blurring is an isotropic smoothing operator that is used to remove noise and detail from images. In this paper, the effect of Gaussian blurring on the extraction of peaks and pits from digital elevation models (DEMs) is studied. First, a mathematical morphological-based algorithm to extract peaks and pits from DEMs is developed. Gaussian blurring is then implemented on the global digital elevation model (GTOPO30) of Great Basin using Gaussian kernels of different sizes and standard deviation values. The number of peaks and pits extracted from the resultant DEMs iscomputed using connected component labeling and the results are compared. The application of Gaussian blurring to perform the treatment of spurious peaks and pits in DEMs is also discussed. This work is aimed at studying the capabilities of Gaussian blurring in the modeling of objects and processes operating within an environment.