A Scalable Approach for Generalization of Land Cover Data

Thiemann, Frank; Warneke, Hendrik; Sester, Monika; Lipeck, Udo W.

doi:10.1007/978-3-642-19789-5_20

“…Regular partitioning includes two methods, the use of a regular (often rectangular) grid, or the use of a quad tree. Regular grids have been used for land use generalization by (Thiemann et al 2011), and for most generic processes in (Briat et al 2011) and (Thiemann et al 2013). The size of the grid is adjusted based on the amount of data that can be found in one cell of the grid.…”

Section: Partitioningmentioning

confidence: 99%

“…Once a partitioning technique has been chosen, it is necessary to add a mechanism to restore some kind of geo-graphical context, at least for the features that are located near the boundary of a partition cell. The main proposition in the literature is to provide a buffer area around the partition where data is added for context, but not generalized by the process that handles the partition cell (Briat et al 2011), (Thiemann et al 2011). One difficulty here is to find the buffer size that gives enough context without making the partition cell too large to be processed.…”

Section: Distribution Of Geographical Contextmentioning

confidence: 99%

Experiments to Distribute Map Generalization Processes

Berli¹,

Touya²,

Lokhat³

et al. 2018

Proc. Int. Cartogr. Assoc.

View full text Add to dashboard Cite

Automatic map generalization requires the use of computationally intensive processes often unable to deal with large datasets. Distributing the generalization process is the only way to make them scalable and usable in practice. But map generalization is a highly contextual process, and the surroundings of a generalized map feature needs to be known to generalize the feature, which is a problem as distribution might partition the dataset and parallelize the processing of each part. This paper proposes experiments to evaluate the past propositions to distribute map generalization, and to identify the main remaining issues. The past propositions to distribute map generalization are first discussed, and then the experiment hypotheses and apparatus are described. The experiments confirmed that regular partitioning was the quickest strategy, but also the less effective in taking context into account. The geographical partitioning, though less effective for now, is quite promising regarding the quality of the results as it better integrates the geographical context.

show abstract

“…Regular partitioning includes two methods, the use of a regular (often rectangular) grid, or the use of a quad tree. Regular grids have been used for land use generalization by Thiemann et al (2011), and for most generic processes in Briat, Monnot and Punt (2011) and Thiemann et al (2013). The size of the grid is adjusted based on the amount of data that can be found in one cell of the grid.…”

Section: Partitioningmentioning

confidence: 99%

“…Once a partitioning technique has been chosen, it is necessary to add a mechanism to restore some kind of geographical context, at least for the features that are located near the boundary of a partition cell. The main proposition in the literature is to provide a buffer area around the partition where data is added for context, but not generalized by the process that handles the partition cell (Briat, Monnot and Punt 2011), (Thiemann et al 2011). One difficulty here is to find the buffer size that gives enough context without making the partition cell too large to be processed.…”

Section: Distribution Of Geographical Contextmentioning

confidence: 99%

“…Briat, Monnot and Punt (2011) use an attribute field on features that says if it has been processed, and only the first processed partition is able to write the result on the feature at the reconciliation step. Thiemann et al (2011) cut the features at the limits of the partition and reconciliate by merging the cut parcels once generalized. However, cutting features is dangerous as shown in (Chaudhry and Mackaness 2008) with the example of Douglas & Peucker algorithm (Douglas and Peucker 1973) that fixes the initial/final vertices of the simplified lines, so more line cuts means more fixed points and less quality in generalization.…”

Section: Reconciliationmentioning

confidence: 99%

See 1 more Smart Citation

Experiments to Distribute and Parallelize Map Generalization Processes

Touya

¹

,

Berli

²

,

Lokhat

³

et al. 2017

The Cartographic Journal

View full text Add to dashboard Cite

Automatic map generalization requires the use of computationally intensive processes often unable to deal with large datasets. Distributing the generalization process is the only way to make them scalable and usable in practice. But map generalization is a highly contextual process, and the surroundings of a generalized map feature needs to be known to generalize the feature, which is a problem as distribution might partition the dataset and parallelize the processing of each part. This paper proposes experiments to evaluate the past propositions to distribute map generalization, and to identify the main remaining issues. The past propositions to distribute map generalization are first discussed, and then the experiment hypotheses and apparatus are described. The experiments confirmed that regular partitioning was the quickest strategy, but also the less effective in taking context into account. The geographical partitioning, though less effective for now, is quite promising regarding the quality of the results as it better integrates the geographical context.

show abstract