Package documentation: terrain

The terrain package contains functions necessary for analyzing map data to identify choke points, split the map into logical zones, create potential fields etc.

Modules:

• potential.py
• terrain_analysis.py
• zone.py

potential.py

Contains implementation of potential functions.

view_potential(potential, plot_type='colorgraded'):

Note: This function will be ported to mapworks.visualizer shortly.

view_potential(potential) displays the potential field indicated in potential as a color graded map with yellow for highest potential and blue for lowest. Here, potential is a 2D array of potential height values.

view_potential(potential, plot_type) displays the potential in a selected plot type. Plot type can be one of the following:

• 'colorgraded': Default type. Color graded map with yellow for highest potential and blue for lowest.
• '3d_surface': 3D surface plot for potential.
• 'contour': Contour plot.

coulomb(map_data, d0=2, nu=800, scale=100):

coulomb(map_data) generates a potential field (2D array of potential heights) based on Coulomb repulsion or inverse r^2 rule. Requires scipy.

Shaping Parameters:

• d0: Fall off parameter.
• nu: Potential height parameter.
• scale: Scale. Similar to fall off.

goalpost(map_size, goal, xi=1 / 700):

coulomb(map_data, goal) generates an attractive paraboloidal potential field centred at the (x, y) tuple goal.

Shaping Parameters:

• xi: Attractive force parameter.

manhattan(map_data, max_depth=5, return_depth_vector=False):

manhattan(map_data) creates a potential field that decays linearly with distance from the nearest map obstacle.

Shaping Parameters:

• max_depth: Maximum depth beyond which potential flattens out.
• return_depth_vector: If set to True, a list of depth vectors is also returned. The nth depth vector is a list of tuples(x, y) at depth n.

distance_transform(map_data, max_depth=6):

distance_transform(map_data) achieves exactly the same thing as manhattan at less than 1% of the time. Requires scipy however.

terrain_analysis.py

Implements the TerrainAnalyzer class that identifies choke points and splits the map into logical zones.

TerrainAnalyzer.view_terrain(self, grid=True, gates=True, graph=True):

Note: This function will be ported to mapworks.visualizer shortly.

TerrainAnalyzer.view_terrain() displays the map split into logical zones.

• grid=False can be used to turn off the grid.
• gates=False if you don't want to see gate points marked in green.
• graph=False to turn off the connected graph of zones.

Properties:

• TerrainAnalyzer.map_data: The passable map.
• TerrainAnalyzer.map_size: [m, n] as in map_data.shape.
• TerrainAnalyzer.height_map: manhattan potential field.
• TerrainAnalyzer.height_vector: list of lists. The nth list is a list of points at height n.
• TerrainAnalyzer.max_height: Maximum height of potential field.
• TerrainAnalyzer.zone_map: 2D array of same size as map_data. Contains zone id of each tile.
• TerrainAnalyzer.zone_no: Number of logical zones.
• TerrainAnalyzer.zones: List of zone type objects. See zone.py below.
• TerrainAnalyzer.gate_points: List of points at the boundaries to zones.

zone.py

Implements Zone type objects for the TerrainAnalyzer.

Zone.update_center():

Zone.update_center() is used by TerrainAnalyzer to calculate mean of all points in the zone.

Zone.add_gate_point(self, adj_zone_id, gate_point):

Zone.add_gate_point(adj_zone_id, gate_point) is used by TerrainAnalyzer to add a gate point to an adjacent zone.

Properties:

• Zone.zone_id: Identification number. Starts from 0.
• Zone.points: List of points in the zone.
• Zone.size: No. of points in the zone.
• Zone.center: Center point (mean) of zone.
• Zone.entry_points: Dictionary with keys being the adjacent zone ids. Values are corresponding gate_points to these zones.