Sharp feature detection in point clouds

Abstract

This paper presents a new technique for detecting sharp features on point-sampled geometry. Sharp features of different nature and possessing angles varying from obtuse to acute can be identified without any user interaction. The algorithm works directly on the point cloud, no surface reconstruction is needed. Given an unstructured point cloud, our method first computes a Gauss map clustering on local neighborhoods in order to discard all points which are unlikely to belong to a sharp feature. As usual, a global sensitivity parameter is used in this stage. In a second stage, the remaining feature candidates undergo a more precise iterative selection process. Central to our method is the automatic computation of an adaptive sensitivity parameter, increasing significantly the reliability and making the identification more robust in the presence of obtuse and acute angles. The algorithm is fast and does not depend on the sampling resolution, since it is based on a local neighbor graph computation.

Sharp feature preserving MLS surface reconstruction based on local feature line approximations

“…All feature point computations including detection and the list of nearest connected neighbors can be pre-processed. Timings are given in [30]. Without these preprocessing steps, timings close to the classical MLS can be reached, since only a few number of the points (8% in the cube example of Fig.11) undergo a neighborhood modification with O(k) operations for feature curve computation, clipping and up-sampling (k = |N p |).…”

“…Since we only need the points belonging to a sharp feature to be marked, any of the methods [30,6] seems appropriate. Several other existent techniques for point clouds are dedicated to detect points on any kind of characteristical feature, e.g.…”

“…In 3D the clustering behavior is similar, except that more clusters may appear for a corner-type feature. [30] presents finally an augmented adaptive version of this basic idea. It is adaptive in the sense that optimal local parameters are computed automatically so that very flat as well as very acute features can be recognized robustly even in the presence of noise.…”

“…The Gauss-map clustering method [30] has been proven to be reliably enough. It can be observed, for example that perfect (exact) input data without noise leads to perfect recognition of feature points and our resulting MLS surface is perfect as well, see first row of Figure 11.…”

Voronoi-based feature curves extraction for sampled singular surfaces

“…Recently, Weber et al [6] apply Gauss map clustering [7] to find feature points and locally fit feature curves as cubic Beizer splines. Dey et al [1] use Gaussianweighted graph Laplacian to identify singular points and Reeb graph to connect them into feature curves.…”