Morphological methods for detection and classification of edges in range images

Abstract

Range images provide an explicit encoding of the shape and geometric structure of the objects in the image from the point of view of the sensor. Since morphological methods are inherently geometric in nature, they are ideally suited for the analysis of range images. However, morphological edge operators meant for intensity images cannot be used to detect edges in range images because roof and crease edges do not correspond to depth discontinuity. In this paper two schemes for detection and classification of edges in range images based on morphological operations are presented.The first method uses the residues of openings and closings to detect roof and crease edges. Directional sensitivity to edges is incorporated by using structuring elements oriented in different directions. The second method employs the residues of dilation and erosion at multiple scales and provides a richer description of the surface structure at each point in the image by classifying each pixel as belonging to one of the eight possible structure types: positive roof, negative roof, positive crease, negative crease, top of step, base of step, ramp, and constant surface. Several examples, including some involving the fusion of edge information from registered range/intensity images, are presented.

Theoretical aspects of vertically invariant gray-level morphological operators and their application on adaptive signal and image filtering

“…In principle, the types of features extracted in the PUF procedure are the highly varying parts (or the roof edges [16]) of a signal. An important characteristic of the features extracted by a PUF procedure is that no additional features will be introduced when the scales varying from larger to smaller.…”

“…A shift-invariant opening is an MV opening whose SE's are the same with respect to every spatial positions. In fact, shift-invariant openings are widely used in image and signal processing [7], [16], [18], [19], [30] and are introduced by many textbooks [10], [11], [13]. 3 Hence, the PUF procedure can also be equivalently implemented using shift-invariant openings by considering that the scale function (of the MV opening) assigned in each iteration is piecewise constant.…”

Extraction, Quantification and Visualization of Protein Pockets

“…In 2D range images, Krishnapuram and Gupta [30] uses dilation and erosion operations to detect and classify edges; Gil and Kimmel [24] discussed algorithms for computing onedimensional dilation and erosion operators. In addition to the extraction of polygonal surfaces from volumetric functions, 3D polygonal models are also converted into volumetric representations and then modified, repaired and simplified using morphological operations [21,36].…”

A Stochastic Algorithm for 3D Scene Segmentation and Reconstruction

“…The space structure is typical, and will be valid even when other classes of models are used in future. In the literature of range segmentation, many methods are applied, like edge detection [9], region growing methods [5,1], clustering [6,4], and some energy minimization method, generalized Hough transforms. But none of these methods can search in such complex spaces.…”

“…Therefore these algorithms only deal with polyhedra scenes. In the literature, algorithms for segmenting intensity images have been introduced or extended to rang image segmentation, e.g., edge detection [9], region growing methods [5,1] and clustering [6,4]. An empirical comparison study was reported jointly by a few groups in [7].…”