Three dimensional sensor network surface skeleton extraction method

Abstract

The invention discloses a three dimensional sensor network surface skeleton extraction method. The three dimensional sensor network surface skeleton extraction method includes the steps: every internal node calculates the feature point thereof; a restrictive broadcast mode is used to enable the feature points of the nodes to form a plurality of feature connected components; surface skeleton points are identified; a maximum independent set of the surface skeleton points is constructed; the surface skeleton nodes in the maximum independent set are used to obtain a Voronoi graph; and based on the Voronoi graph, triangularization of the dual graph Delaunay can be obtained so as to obtain a three dimensional sensor network surface skeleton. The surface skeleton point identifying method has no use for depending on a special grid structure and cannot be influenced by network noise, distance round-off error and sparse network, and can obtain a steady surface skeleton.

Description

technical field [0001] The invention belongs to the technical field of wireless sensor networks, and more specifically relates to a method for extracting a surface skeleton of a three-dimensional sensor network. Background technique [0002] Surface skeleton is an important structure of 3D sensor network, which can fully reflect its geometric topological features; existing research shows that using the geometric topological features of sensor networks is helpful to design high-performance network protocols. In the field of computer vision research, there are endless researches on 3D object surface skeletons, but most of them focus on continuous domains and use centralized algorithms, so they cannot be directly applied to discrete sensor networks with resource constraints. The existing literature on skeleton extraction of sensor networks mainly focuses on two-dimensional sensor networks, while the surface skeleton extraction method proposed by XiaSu et al. is the only researc...

AI Technical Summary

Problems solved by technology

However, this method based on morphological refinement, which is essentially similar to the burnt grass model proposed by Blum, is very sensitive to the distance metric used and usually cannot accurately locate the surface skeleton points.

In addition, it is very susceptible to boundary noise, resulting in a bandage skeleton, which must be processed post-hoc

More importantly, this method relies on the established UTC network structure, and the establishment of a UTC structure usually requires a relatively high network density, where the network is relatively sparse, there are small-sized holes, and only inter-node connection information (rather than node coordinate information) When available, it is difficult to get accurate UTC results

Finally, since the UTC structure needs to be constructed in advance, this method cannot be applied in dynamic networks

When the network topology changes due to node failure or new node joining, rebuilding UTC will bring huge communication overhead

Images