Based on Harig 2024, TU Dresden — Automated Delineation of Inner Development Areas Including Result Evaluation.
This document is the reference for all IB-Tool 3 processing parameters: what each one controls, its mathematical background, and practical guidance for tuning. For the algorithmic context of each parameter within the processing pipeline, see docs/how-it-works.md.
IB-Tool 3 performs the automated delineation of the Innenbereich (§ 34 BauGB) — the coherently built-up part of a municipality — based on geometric and topographic object data. The aim is a uniform, data-driven, and reproducible approximation of that legally defined area; see terminology.md for the definition, for the relationship to the term Urban Growth Boundary used in the international publication, and for which source covers which part of the method.
The parameter values documented here follow the dissertation Harig (2024).
The methodology combines:
- Building footprints (HU-DE) as an indicator of actual development
- Topographic line data (ATKIS road network) to partition the spatial structure
- Auxiliary geometries (ATKIS forests, water bodies, bogs, and swamps)
- Density-based clustering algorithms to identify developed areas
- Graph-based aggregation (Minimum Spanning Tree) to merge contiguous built-up zones
The goal is an objective and reproducible delineation that remains planimetrically interpretable. The method forms a core component for settlement monitoring, inner-development potential analyses, and urban sprawl studies (cf. Harig 2024, Chap. 1.2, 1.4, 8.4).
The processing follows the workflow described in the dissertation (see Fig. 4.1–4.10 in Harig 2024):
- Partitioning of the study area into street blocks using the road network.
- City block generation by removing non-enclosed portions (Fig. 4.6).
- Building coverage ratio (BCR) calculation per block based on building footprint area (Fig. 4.7).
- Filtering of small or irrelevant buildings (e.g. garages, outbuildings).
- Identification of dense blocks according to density thresholds (Chap. 4.5.3).
- Aggregation using a Minimum Spanning Tree algorithm (Chap. 4.4.6).
- Post-processing: filling building gaps and connecting adjacent settlements (Fig. 4.10).
The configurable parameters control different stages of this process and govern its sensitivity, spatial coherence, and level of generalisation.
- Type: Float
- Function: Controls the identification of densely built-up blocks in
identify_dense_blocks(). - Definition: Specifies the minimum share of a block that must be covered by building footprints for it to be classified as "densely developed".
- Background: This parameter operationalises the building coverage ratio (BCR) described in the dissertation, defined as the ratio of building footprint area to total block area (cf. Harig 2024, Chap. 4.4.3, p. 29 f.).
- Effect:
- High values (e.g. > 25) — only very compact, enclosed built-up blocks are recognised as settlements (e.g. town centres).
- Low values (e.g. < 15) — looser development patterns are also captured (e.g. rural areas).
- Recommended range: 18–22 (cf. Chap. 4.5.3, empirical calibration).
- Notes: This step precedes the clustering and boundary-snapping stages. Its primary benefit is computational: blocks that can safely be classified as inner development are excluded from all subsequent steps. The default of 18 is intentionally conservative so that the false-positive rate remains very low.
- Type: Float
- Function: Used as a fallback reference threshold for the full study area when no local density values are available (
calc_footprint_density()). - Definition: Mean building coverage ratio across all partitions of the study area.
- Background: Provides robustness in areas of sparse development or incomplete building data (cf. Harig 2024, Chap. 4.5.3).
- Effect:
- Low values (< 10) → buildings in relatively loose development patterns are also aggregated.
- High values (> 20) → only very dense settlement cores are aggregated.
- Recommended setting: Automatic calculation (0 = auto); set manually for highly heterogeneous study areas.
- Notes: This value specifically affects small or dispersed settlements. In those cases its influence is substantial.
- Type: Float
- Function: Filters out undersized building objects in
input_hu_filter(). - Definition: Buildings with a footprint area below this threshold are removed from all subsequent calculations.
- Goal: Exclusion of non-settlement-relevant structures (e.g. garden sheds, carports, outbuildings).
- Background: Small buildings and outbuildings do not meet the conditions required to establish a contiguous development context. They are therefore removed at the beginning of the pipeline.
- Recommended values: 50–60 m², depending on the local context and regional average.
- Effect: At the settlement fringe, boundaries are drawn directly along residential buildings. Outbuildings at the rear of plots are thus excluded.
- Type: Integer
- Function: Defines the minimum number of buildings required for a cluster to be recognised as a valid settlement or locality.
- Usage: Applies during post-processing (
patch_remove()): patches below this threshold are removed. - Background: Based on empirical analysis (Harig 2024, Chap. 4.5.6) and expert surveys (Appendix A.1.3), a locality is only considered an independent inner development area from approximately 20 buildings or 1 ha of area.
- Recommended values: 15–25 buildings.
- Effect:
- Prevents isolated groups of buildings from being classified as settlements.
- Stabilises results in rural areas.
- Type: Float (m²)
- Function: Used in
patch_remove()to discard small settlement fragments. - Definition: Patches whose area falls below this threshold are rejected as settlement areas.
- Background: Corresponds to the empirically determined minimum size of a locality (Chap. 4.5.6).
- Recommended value: 10,000 m².
- Effect:
- Small fragments are removed → cleaner inner-development boundaries.
- Reduces over-segmentation in sparsely populated areas.
- Note: Shape is not taken into account; only area is evaluated.
- Type: Float (m²)
- Function: Used in
gap_close()to fill undeveloped areas enclosed within contiguous settlement geometries. - Definition: Specifies the maximum area up to which open spaces inside a settlement polygon are filled (treated as inner development rather than open land).
- Background: Reflects planning definitions of the inner development area.
- Recommended value: 10,000 m².
- Effect:
- Smaller open spaces (gardens, courtyards, playgrounds, meadows, car parks) are incorporated.
- Prevents excessive fragmentation of settlement areas.
- Type: Float (m²)
- Function: Used in
gap_close()to bridge gaps at the settlement fringe. - Definition: Specifies the area threshold below which gaps at the settlement boundary are closed.
- Background: Building gaps count as inner development. Which undeveloped area can still be considered a gap is highly contested and difficult to define precisely — local conditions must always be considered. As a general rule, the looser the surrounding development, the larger a gap may be before it falls outside the inner development area.
- Recommended value: 40–100 m (equivalent to 4,900 m² at 70 m radius).
- Effect: The algorithm identifies areas located between existing settlement cluster polygons that border those polygons along at least 75 % of their perimeter. Only when the share of contact with settlement areas is substantially greater than that with open land is the area classified as a gap. Such areas are then incorporated into the inner development polygon (cf. Harig 2024, Chap. 4.5.4, Fig. 4.10).
- Type: Float
- Function: Passed to
identify_dense_blocks()duringpatch_remove()to determine which blocks are considered densely developed before the final merge step. - Definition: Minimum building coverage ratio (BCR) required for a block to be classified as a dense block and retained regardless of patch size.
- Background: Conceptually identical to
min_overlap_blocks, but applied specifically within thepatch_removepost-processing step to preserve high-density areas even if they would otherwise fall below themin_patch_sizethreshold. - Default: 18 (fixed internal threshold; not currently exposed in the UI).
- Effect: Dense blocks with BCR ≥ 18 % and a combined footprint area ≥
footprint_area_sumare always retained, even if their polygon area is belowmin_patch_size.
- Type: Float (m²)
- Function: Used in
patch_remove()alongsidefootprint_density_thresholdto decide which dense blocks are retained. - Definition: Minimum total building footprint area within a dense block for it to be kept in the final output.
- Background: Prevents very small but technically "dense" blocks (e.g. a single large building surrounded by a tiny polygon) from artificially inflating the settlement area.
- Default: 6,000 m² (fixed internal threshold; not currently exposed in the UI).
- Effect: A dense block is retained only if its polygon area is ≥
min_patch_sizeor its total footprint area is ≥footprint_area_sum. This ensures that compact, fully developed micro-blocks are not discarded.
| Parameter | Default value | Primary effect | Sensitivity |
|---|---|---|---|
min_overlap_blocks |
18 | Reduces processing time; pre-classifies dense blocks | high |
global_footprint_density |
auto (0) | Delineation of small / sparse settlements | high |
min_area |
56.8 m² | Excludes small buildings at the settlement edge | low |
min_bdg_count |
20 | Minimum building count per locality | medium |
min_patch_size |
10,000 m² | Noise suppression; removes splinter areas | medium |
max_hole_size |
10,000 m² | Fills enclosed open spaces within settlements | medium |
max_gap_size |
4,900 m² | Bridges building gaps at the settlement edge | high |
footprint_density_threshold* |
18 | Dense-block retention threshold in PatchRemove | low |
footprint_area_sum* |
6,000 m² | Minimum footprint area for dense-block retention | low |
* Internal threshold — currently set as a fixed default in ibtool.py, not exposed in the UI.
The full reference list, with the role of each source, is in terminology.md → References. The parameter values on this page are taken from:
- Harig, O. (2024): Automatisierte Abgrenzung von Innenbereichen einschließlich Ergebnisevaluierung – Grundlage für ein Siedlungsflächenmonitoring. TU Dresden, Faculty of Environmental Sciences.
- Chapter 4.4 — Methodology
- Chapter 4.5 — Parameterisation of the Method
- Appendix A.1.3 — Expert Survey
- Bukies, M.; Meyer, G.; Rabe, H. (2009): Abgrenzung des Innenbereichs im unbeplanten Siedlungsgebiet.
- Spannowsky et al. (2020, 2022): Baugesetzbuch — Commentary on § 34 BauGB.
| File | Content |
|---|---|
docs/how-it-works.md |
Full algorithmic pipeline with pseudocode for each step |
docs/input-data.md |
Input layer specifications, filter file format |
docs/error-handling.md |
Logging system and debug mode |