Fast and Effective Feature-Preserving Mesh Denoising

Abstract

We present a simple and fast mesh denoising method, which can remove noise effectively, while preserving mesh features such as sharp edges and corners. The method consists of two stages. Firstly, noisy face normals are filtered iteratively by weighted averaging of neighboring face normals. Secondly, vertex positions are iteratively updated to agree with the denoised face normals. The weight function used during normal filtering is much simpler than that used in previous similar approaches, being simply a trimmed quadratic. This makes the algorithm both fast and simple to implement. Vertex position updating is based on the integration of surface normals using a least-squares error criterion. Like previous algorithms, we solve the least-squares problem by gradient descent, but whereas previous methods needed user input to determine the iteration step size, we determine it automatically. In addition, we prove the convergence of the vertex position updating approach. Analysis and experiments show the advantages of our proposed method over various earlier surface denoising methods.

A geomorphology-based approach for digital elevation model fusion – case study in Danang city, Vietnam

“…The higher weight was applied for GDEM in the valley areas, because of the limitation of SRTM in those areas. The output-fused DEM was filtered using a denoising algorithm according to Sun et al (2007). Finally, the fused DEM was compared to the reference DEM to assess the efficiency of the DEM fusion method.…”

“…Combination of two data can take into account the advantages of each DEM source and provide complementary inputs to enhance the quality for the global DEMs. DEM fusion workflow combines the weighted averaging and denoising algorithm (Sun et al, 2007).…”

“…According to the landform types, the different weights were used for GDEM and SRTM. Finally, a denoising algorithm (Sun et al, 2007) was applied to filter the output-fused DEM. This fused DEM shows excellent correlation to the reference DEM, having a correlation coefficient R 2 = 0.9986, and the accuracy was also improved from a root mean square error (RMSE) of 14.9 m in GDEM and 14.8 m in SRTM to 11.6 m in the fused DEM.…”

An evaluation of EU-DEM in comparison with ASTER GDEM, SRTM and contour-based DEMs over the Eastern Mecsek Mountains

“…According to the previously stated correlation results the continuous forest areas, in regions with higher slope values, were subdivided based on aspect categories. For smoothing the noises, residual errors on the models, Sun's denoising algorithm (Sun, X. et al 2007;Stevenson, J.A. et al 2010) was applied.…”

Three-dimensional Modeling of Carbon Fiber Cloth Surface from Dense and Scattered Points Cloud

“…The diffusion factors which depend on main curvature and main direction of curvatures are sensitive to obvious characteristics on surface. And diffusion curvature on certain direction of sharp edge and corners can be decreased [10]. Fleishman's iterated bilateral filter [11] is a typical anisotropic method.…”