Thinning Algorithm to Generate k-Connected Skeletons

León, Juan Luis Díaz de; Yáñez, Cornelio; Guzmán, Giovanni

doi:10.1007/978-3-540-30463-0_76

Cited by 1 publication

(1 citation statement)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such a small change, again, does not introduce too much overhead. In our prototype, we adapt a mask-based thinning algorithm (Diaz de Leon et al, 2004) for biased thinning. The gain in running time is insignificant.…”

Section: Biased Thinningmentioning

confidence: 99%

Completing Fragmentary River Networks via Induced terrain

Lau

Franklin²

2011

Cartography and Geographic Information Science

View full text Add to dashboard Cite

Fragmentary river segments have to be joined appropriately before becoming useful to addressing transportation problems like route planning and pollutant tracking. We suggest the induced terrain approach to complete the job with high accuracy and efficiency. By first approximating a terrain compatible with the partial river segment and height information, and then deriving a river network from that induced terrain, we effectively enforce the restrictions imposed by the height observations. Natural neighbor interpolation with stream burning is capable of generating induced terrains that predict river locations accurately without adjustment for any parameters. Considerable time is saved from executing the global river derivation again and again to figure out the optimal parameter values, especially when we are confronted with increasingly massive terrain dataset. In the subsequent river derivation, we propose biasing in favor of the known river locations. Their water amounts are set to the critical value just sufficient for the locations to be regarded as river locations. In the final thinning process, these locations are simply set not to be trimmed. All known river locations are thus fully recovered at the end. We expect the same techniques can be applied to the recovery of some other 2D and 3D networks, like road networks and dendrite networks.

show abstract

Section: Biased Thinningmentioning

confidence: 99%